WO2001027801A2 - Electronic shopping management: user interface - Google Patents

Abstract

A system and method that provide an easy-to-understand presentation of information about customer movements, moods, and demographics for a website. Such a presentation makes it easier for the owner or designer of a website to understand how customers are interacting with his site. In the described embodiments, a high level view of the website is preferably represented as a "campus" having buildings corresponding to predetermined tasks performed by customers on the website. In the described embodiments, a midlevel floorplan view of each task represents the steps of the task as rooms within the building. Certain tasks can also have sub-tasks, which are represented as multiple floors of the building. Customers of the website are represented graphically in both the high level campus view and the midlevel floorplan view. The location, mood, and actions of the animated characters help the user understand how customers are interacting with his website. As customers interact with the website, the corresponding animated characters move through the high level campus view or the midlevel floorplan view.

Description

ELECTRONIC SHOPPING MANAGEMENT: USER INTERFACE

Inventors

Elizabeth B Charnock, Loki Der Quaeler, Mark Yu-hung Lm, Curtis Thompson

RELATED APPLICATIONS

This application claims the benefit, under 35 U S C § 119(e), of Provisional Application Number 60/159,226, entitled "Method and Apparatus for Electronic Shopping Management" filed on October 13, 1999, by Elizabeth B Charnock, Loki Der Quaeler, Philip Chang, Jesse Burns, Mishkin Berteig, and Curtis Thompson, and incorporates that application by reference in its entirety

This application also claims the benefit, under 35 U S C §119(e), of Provisional Application Number 60/201,183, entitled "Electronic Shopping Management" filed on May 2, 2000, by Elizabeth B Charnock, Loki Der Quaeler, Philip Chang, Jesse Burns, Mishkm Berteig, Curtis Thompson, and Mark Yu-Hung Lin, and incorporates that application by reference in its entirety

This application is related to copending U S application seπal number , entitled "Electronic Shopping Management Task Models" filed on October

13, 2000, by Elizabeth Charnock, Jesse Burns, Philip Chang, and Fabien Gerard Norbert Hertschuh, which is incorporated by reference herein in its entirety

This application is related to copending TJ S application seπal number , entitled "Electronic Shopping Management User States" filed on October 13, 2000, by Elizabeth Charnock, Loki Der Quaeler. Philip Chang, Jesse Burns, Curtis

Thompson, Michel Henπ Guzy, Fabien Gerard Norbert Hertschuh, and Wenxin Mao, which is incorporated by reference herein m its entirety

This application is related to copending U S application serial number , entitled "Electronic Shopping Management Intervention" filed on October 13, 2000, by Elizabeth Charnock and Philip Chang, which is incorporated by reference herein its entirety

Background

Technical Field

The present invention relates generally to website management, and more particularly, to a way of facilitating analysis of customer behavior at websites

Background of the Invention

The e-commerce industry is growing at an astounding rate More and more companies are creating web-based stores in which customers can shop One of the advantages of e- commerce, which the e-commerce industry counts on, is the ease with which customers can shop on-line, instead of physically having to go to conventional stores

However, despite the numerous advantages that e-commerce businesses have over conventional stores, e-commerce businesses also face some unique problems Floor managers m conventional stores can physically observe potential customers, and assess their moods For instance, m conventional stores, floor managers or other staff can assess when a customer is confused It is of tremendous value to a business to be able to assess the moods of customers, since such mood-assessment can help the floor managers to organize the conventional store in a more convenient and easily navigable manner For instance, if several potential customers appear to be confused while finding some specific product(s), the floor managers might implement a change in the layout of the store to alleviate such confusion Similarly, if several customers appear to be pleased by some particular feature/promotion offered by the store, such features/promotions can be increased or repeated

This kind of information, however, has not been available for e-commerce businesses

E-commerce executives have very little understanding, if any, of the moods of their current customers, and of how the design of their website affects the intents, goals, and moods of their potential customers

In the intensely competitive field of e-commerce, only companies that can observe customer behavior, and accordingly refme the design of their websites, will have the most user-friendly websites. Only then will their websites keep customers happy and satisfied, and make it easy for them to navigate through the website, to locate what they need, and to perform desired tasks. Such companies increase their chances of surviving and succeeding in the world of e-commerce. Website customers may abandon other companies that do not make such changes, especially when the customers have other, better options, available to them.

Accordingly, it is desirable to provide a system and method for observing customers, and for assessing their satisfaction levels at different points in their traversal of a website. Furthermore, there exists a need for a system and method to deduce the "moods" of the customer as he or she progresses through the site, given a set of customer actions on a website. Furthermore, there exists a need for a clear method of presenting this data to the owner or designer of the website.

Summary of the Invention

The described embodiments of the present invention provide an easy-to-understand presentation of information about customer movements, moods, and demographics for a website. Such a presentation makes it easier for the owner or designer of a website to understand how customers are interacting with his site. The owner or designer is hereinafter called a "user" when he is using the described customer analysis tool. Persons accessing the website are refeπed to as "customers." In the described embodiments, a high level view of the website is preferably represented as a "campus" having buildings corresponding to predetermined tasks performed by customers on the website. In the described embodiments, a midlevel floorplan view of each task represents the steps of the task as rooms within the building. Certain tasks can also have sub-tasks, which are represented as multiple floors of the building. In a first embodiment, customers of the website are represented graphically in both the high level campus view and the midlevel floorplan view. At least one described embodiment uses animated characters to represent customers or groups of customers. The location, mood, and actions of the animated characters help the user understand how customers are interacting with his website. As customers interact with the website, the corresponding animated characters move through the high level campus view or the midlevel floorplan view.

In the described embodiments, the user can choose to view graphical representations representing demographic information about the customers interacting with the website. This demographic information can be, for example, indicated by shirt colors or clothing styles on the animated characters representing the customers Alternatively, or in addition, the user can decide to view mood information about the customers The mood information can include, for example, whether a customer is continuing to interact with the website or has stopped interacting (1 e , has "died") Furthermore, customer mood during interaction with the website can be indicated by causing the character to jump up and down in anger, look at his watch, curse, etc

In the described embodiments, the location of the animated character on the display within a room can represent how long the customer has been involved in the task step represented by the room Customers who move swiftly though the task steps of the website are displayed as characters that do not stray from the path connecting the rooms represented by the task steps The longer a customer remains in a task step, the further his character wanders from the path

In certain embodiments, the actions performed by the animated characters change depending on the type of task step that the customers are cuπently performing Thus, for example, the characters of customers who are in a shopping task step and are shopping are represented as pushing shopping carts, while the characters of customers who are in a registration task step are represented as filling out forms

At least one embodiment of the present invention runs m "real time " That is, the animated characters on the display coπespond to actual customers who are currently accessing the website In another embodiment, information about customer movements within a website are recorded in memory or on a storage medium ("taped") and analyzed for display at a later time A display in accordance with the present invention can be based on the recorded data In this case, the display uses a VCR or DVD metaphor, asking the user to load a tape to be viewed The described embodiments also allow the user to view statistical information about customer actions (for example, how many customers have completed each task, how many customers have been satisfied, angry, etc ) This displayed statistical information can be updated as the customers take further actions, resulting m real-time running statistical totals, or the stattstical information can be displayed as a non-moving "snapshot" of statistical information that is not updated regularly

The described embodiments also allow the user to search the data for the various tasks for particular customer actions Past actions can be searched when the display is being generated in real-time Both past and future actions can be searched when the data is being generated from a tape Thus, the user can search through a large amount of data for customer actions that are of particular interest or concern

The descπbed embodiments are also capable of generating printed reports concerning customer actions The user can also save snippets of the display for later playback and analysis The user can also take a "snapshot" of the display for future review and analysis

The descπbed embodiments are also capable of "filteπng" the incoming customer data so that the display only reflects the demographics and/or moods of certain customers For example, data might be displayed only for customers having a domain of " edu" or only for customers who have been identified as "bored " The descπbed embodiments can also follow a designated customer through the website and can present information about the customer, such as his mood history and the history of the rooms/task steps he has visited

BRIEF DESCRIPTION OF THE DRAWINGS

Fig 1 shows an example of a high-level view of tasks in a website

Fig 2(a) is a block diagram representing a "tape" mode of an embodiment of the invention

Fig 2(b) is a block diagram representing a "real-time" mode of an embodiment of the invention

Fig 2(c) is a block diagram showing an example m which multiple client side agents communicate with a central collection point, which passes data to the cam software

Fig 2(d) is a flow chart showing steps performed to update the display of the descπbed embodiment

Fig 3 shows an example in which the user chooses "taped" data to view

Fig 4(a) shows an example of task steps in a task, where demographic data is being displayed

Fig 4(b) shows an example of task steps in a task, where mood data is being displayed

Fig 4(c) shows example characters in various moods

Fig 5 shows an example of room enlargement Fig. 6 shows an example of allowing the user to record a snippet of the display.

Fig. 7 shows an example of allowing the user to follow a particular customer and/or display data about a particular customer.

Fig. 8 shows an example of allowing the user to search for certain customers.

Fig. 9 shows an example of allowing the user to look at statistics about the customers.

Fig. 10 shows an example of a problem indicator on the display.

Fig. 11 shows an example of traffic monitoring on the display.

Figs. 12(a)- 12(f) show examples of various preferences that can be set by the user in the described embodiment.

Figs. 13(a)-13(c) show examples of parameters that the user can set for a search.

Fig. 14(a) shows an example format for customer tracking information.

Fig. 14(b) shows an example format for data about a particular customer.

Fig. 15(a)-15(m) show an example format for properties used to specify the campus and floorplan displays.

Fig. 16(a)- 16(d) show example formats for data used to display the campus and floorplan displays.

DETAILED DESCRIPTION OF THE INVENTION

Embodiments of the present invention are now described with reference to figures where like reference numbers indicate identical or functionally similar elements.

A first embodiment of the present invention uses a display that resembles the appearance of a VCR or DVD player. The described tool embodiment is called a "Cam" (short for camera) and a user of the tool is called a "Cam user." The described embodiment allows the Cam user to monitor the interactions of customers within a website by displaying graphical representations of the tasks and task steps of the website as buildings and rooms. The customers are represented as graphical characters positioned and/or moving within the buildings/tasks and rooms/task steps. A. User Interface

Fig. 1 shows an example of a high-level view of tasks in a website. Each task is represented as a building 102, 104, 106 on a scrollable campus view 100. In the described embodiment, a website can contain an arbitrary number of predefined tasks. Some tasks may be logically broken into subtasks, which are represented visually as multiple floors in a single building. A building may be distinguished separately from other buildings by its visual appearance. For example, the larger a building, the more sub-tasks (floors) it may contain. As another example, a shopping task building may look like a store. In Fig. 1, a sign with the name of the task (personal health insurance, find a corporate quote, find a personal quote) sits just outside the building.

Different buildings are logically connected by paths 108, which represent the flow of customer traffic between different tasks. Paths are represented visually as sidewalks between different buildings. Thus, in Fig. 1, customer traffic is shown as moving between a personal Health Insurance task 102 and the other tasks 104 and 106. Any other visual area, not a building or a path, is just lawn.

In Fig. 1, website customers are represented as small colored shapes 109 that navigate from task to task. The shapes 109 representing customers moving on the sidewalks from building to building. On a website, the transition from building to building (task to task) can be accomplished with literally just one click. The shapes 109, however, are often shown as walking or moving so that it is visually clear where different customers, and types of customers, are traveling to and from. In Fig. 1, customer shapes 109 have a color associated with them, corresponding to a customer demographic type, such as ".com", "Gold Star member", etc, or corresponding to cuπent mood state, depending on whether a demographic mode or a mood mode was currently selected by the user. The type of shape 109 used to represent customers changes accordingly. For example, colored triangles represent mood data, while colored circles (as shown) represent demographic data. Fig. 1 shows an example of a user-settable filter 110, as described below in connection with Fig. 4.

Paths between buildings in the campus view can be used to indicate the amount of traffic that typically travels along it. Paths can be single paths, two-way paths with a divider line, or 4 lane highways, two lanes in each direction. If customer traffic is too heavy to show with shapes 109 because the path literally has insufficient area to draw the desired number of circles or triangles, buses will appear to represent an aggregate of customers. The number of individual customers that the bus represents can be written on the bus.

In the described embodiment, the shapes 109 are moving in real-time as the customers interact with the website. Other embodiments may display only sequential "snapshots" of customer shapes or snapshots that are updated periodically. In Fig.1 , when a customer has dropped out of a website, that customer's shape will appear traveling randomly on the grass area outside of the building for a couple of seconds before disappearing.

In one embodiment, the buildings and the paths that connect them are laid out according to explicitly defined relationships between the tasks. In certain embodiments, if two tasks have no explicit relationship, but are often performed in sequence, a normal path won't be drawn, but a dirt path of overlapping footprints will appear. For example, there may be no link between a registration task and a sales task on the website, but customers frequently register and then buy things.

Another embodiment (not shown) takes more of a spider eye or "wide widget" view. In this embodiment, the most important task buildings form a downtown section, while fewer and fewer important tasks/buildings are rendered in increasingly distant outer circles. The scheme for path connections is the same as described above. "Importance" means business importance, as described in copending application number entitled "Electronic

Shopping Management: Task Models" which is hereby incorporated by reference herein. . This campus layout makes it easy to spot patterns of customer traffic that depart from the way the site was planned.

Another embodiment includes a "zoom" feature that allows the user to display a campus view that is zoomed larger or smaller than the view shown in Fig. 1. Thus, an arbitrary number of tasks and buildings can be shown in the campus view.

Fig. 2(a) is a block diagram representing a "tape" mode (retroactive mode) of an embodiment of the invention. In this embodiment, customer tracking information reflecting the interaction of customers with the website is collected and stored in a database of "tapes" 202 and sent to Cam software 206 in batches or periodically. The customer tracking information can be transmitted between programs 202, 206 running on a single data processing system or programs 202, 206 running on multiple data processing systems. The customer tracking information can be transmitted via a bus, wire, line, network (such as the Internet), wireless network, cellular telephone, and any other appropriate method of communication and/or information transmission. As another example, in retroactive mode, the information could be transmitted via removable hard disk media. The data may be sent using the TCP/IP protocol (possibly using the http protocol) or any other appropriate protocol In certain embodiments, user defined display appearance data 204 are accessed and used by Cam software 206 Creation of appearance data 204 is descπbed below in connection with floorplan and campus editors

Fig 2(b) is a block diagram representing a "real-time" mode of an embodiment of the invention In this embodiment, customer tracking information reflecting the interaction of customers with the website is collected in real-time and 252 and sent to Cam software 256 In the described embodiment, the peπodicity of transmission is a definable parameter which each web site owner can configure to his liking The transmission peπodicity needs to be fast enough to maintain a desire display quality The customer tracking information can be transmitted between programs 252, 256 running on a single data processing system or programs 252 256 running on multiple data processing systems The customer tracking information can be transmitted via a computer bus, wire, line, network, wireless network, cellular telephone, or any other appropriate method of data transmission T e functionality of Cam software 256 and agent software 252 can be distributed across one or more processors, systems, or devices In a prefeπed embodiment, agent 252 is implemented as a Java applet In certain embodiments, user defined display appearances of the Cam 254 are accessed and used by Cam software 256

Fig 2(c) is a block diagram showing an example in which multiple client side agents communicate with a central collection point, which passes data to the cam software In this embodiment the agents 262, 263 are implemented as one or more Java applets on the web pages being viewed by the customer, although any appropriate client-side agent could be used The client-side agents 262, 263 communicate with a central collection point 265 which collects customer information data and sends it to the cam software 266 The descπbed embodiment optionally includes user-defined appearance data 264 as descπbed above

Fig 2(d) is a flow chart showing steps performed to update the display of the descπbed embodiment As described below, the Cam user enters input 262 indicating that he wishes to switch the display mode (I e , to switch between the high level campus view the floorplan view or the enlarged room view as described below) In this case, the cam software displays 264 a campus floorplan, or room view as appropriate

Cam software 206 receives 266 customer tracking information that indicates actions taken by the customers within the website This customer tracking information is received peπodically or continuously Fig 14 shows an example format of customer tracking information although any appropriate format could be used If the Cam user has cuπently chosen to display customer mood information 268, Cam software 206 updates 270 the position and mood of the appropπate customer characters in the display m accordance with the customer tracking information If the Cam user has cuπently chosen to display customer demographic information 272, Cam software 206 updates 274 the position of the appropπate customer characters in the display in accordance with the customer tracking information It should be understood that in the descπbed embodiment, the Cam user is able to view customers as animated and/or profiled characters with no mood state information displayed It should also be understood that in the descπbed embodiment, the characters are shown as animated characters In addition, certain embodiments, such as certain animated embodiments, may update the display more frequently than shown in Fig 2(d)

Fig 3 shows an example in which the Cam user chooses "taped" data to view This example allows the Cam user to indicate which prestored customer tracking information to use to generate the display The user clicks on a picture of a VCR tape slot 302 and is prompted 304 to enter the name of prestored customer tracking information

Fig 3 also shows the user interface of the descπbed embodiment Icon 310 corresponds to a snapshot function Icon 312 corresponds to a print function Icon 314 coπesponds to a search function Icon 316 coπesponds to a statistics function Icon 318 coπesponds to a report generation function Icon 320 coπesponds to a filter function Icon 322 coπesponds to a help function Icon 324 coπesponds to an exit function

In addition, in the described embodiment, buttons on the user interface exist as follows play button 330, rewind button 331 (double backward aπows), fast forward button 333 (double forward aπows), and stop button 334 (dot) Activation of each of these buttons causes a coπesponding playback function to occur on the display Thus, the play button 330 causes the Cam software to begin displaying characters in accordance with the interaction of users on a web site The playback speed can be faster than realtime for taped or delayed data, as discussed below in connection with Fig 12(a) The rewind button 331 causes the display to move backward to an earlier display time (this function can display while rewinding or not in vanous embodiments) To implement this function, the Cam software must of course, maintain a record of the data used to display the characters thus far and what display times coπespond to these characters The fast forward button 333 causes the display to move forward to a later time in retroactive mode (this can display while fast forwarding or not in various embodiments). The stop button 334 causes the movement of the display to stop. Incoming realtime data can be queued in the pause situation.

In addition, other buttons used in the described embodiment include: record button 332 (red dot); go to beginning button 335 (backward double aπow with line); go to end button 336 (forward double aπow with line); and stop button 337 (two parallel lines). Activation of each of these buttons causes a coπesponding function to occur on the display. The record button 332 is discussed below. The go to beginning button 335 causes the tape (or realtime playback) to jump to the beginning (and to begin playing again in certain embodiments). To implement the go to beginning function, the Cam software must, of course, maintain a record of the data used to display the characters thus far and what display times coπespond to these characters. The go to end button 337 causes the tape to jump to the end in retroactive mode. The pause button 337 causes the tape to be paused. Incoming realtime data can be queued in the pause situation.

Fig. 4(a) shows an example of a midlevel floorplan view 400 of task steps in a task, where customer demographic data is being displayed. In the described embodiment, the Cam user clicks on a building on the campus view to see the floorplan view 400 of that building/task. The rooms 402, 403, 404, 405 (and other rooms scrolled off the display) in the floorplan coπespond to the task steps in a task ("What to expect," "Specify Preference," Compare chosen," and "Apply online"). It will be understood that any appropriate tasks can be used. In a floorplan view, characters, such as characters 420, ideally move down a path/hallway 406 in a single direction, from the entrance to the exit. Hallways 406 represent the flow of customer traffic between task steps. Larger rooms represent steps that historically have more customer traffic. If an individual customer starts to linger noticeably longer than the statistical average, his character starts to get pushed off the path and further into the room. This allows other characters to pass them by without interfering with them, and also provides an immediate visual cue as to how long different customers have been in the task.

An ordered task is a task in which the task steps have some required order (for example, step c has to be performed after step b). This ordering of task steps need not be linear. The rooms in an ordered task may have multiple entrances and exits. If an ordered task happens to be strictly linear, then there exists only a single entrance. There may be more than one exit, depending on how many blessed egresses were defined for the task. A blessed egress is an exit that represents a path leaving the website, but going to another "approved" website (such as a partner's suite). A room that isn't part of an ordered task has only a single entrance, which leads to a corridor. There may be more than one exit from a task, depending on how many blessed egresses were defined for the task. An entrance represents a visual cue to when customers enter the room/task step. An exit represents a visual cue to when customers exit the room/task step. When a customer moves from step A to step B in a task for example, the customer' s character should leave Room A from its appropriate exit, and then enter Room B from its appropriate entrance. In a non-ordered task (not shown), hallways connect the entrances and exits of the different rooms and the rooms are not contiguous. If there exists a way to move from Step A to Step B, then there must be a hallway connected from the exit of Room A to the entrance of Room B. In an ordered task, there are no hallways, because the rooms in an ordered task are directly connected to each other.

In order to visually distinguish between entrances and exits, aπows are placed along the entrances and exits depicting the flow of customer traffic. An aπow is drawn along the entrance pointing 'inwards' towards the room, while another aπow is drawn along the exit pointed 'outwards' out of the room.

There are several different kinds of representations of doors. These are as follows: Single or primary entrance: circular or revolving doors, as displayed in traditional architectural drawings. - Multiple entrances: single width sliding doors

- Single or primary blessed egress: double width sliding doors

Multiple blessed egresses: single width sliding doors with the word "exit" above it.

- Subtask M -> Subtask N egress: an elevator with sliding doors.

In real-time mode, the doors remain as static images for efficiency reasons, much as if they were in an architectural floorplan. But in retroactive mode, the doors animate. The sliding doors slide open when customers come or go. The revolving doors turn. This is a visual cue that draws attention to both successful completions, and to new customers arriving in the task. Whether or not to animate the doors in retroactive mode is configurable as a user preference, as is whether a bell should ring upon entry to or exit from a task step.

Much as in a "brick & mortar" department store, there are paths or corridors through the store of varying importance to the merchant, and therefore various visibility and attractiveness to customers. In the Cam visualization of a floorplan, there are two types of corridors. Both types have aπows indicating a prefeπed direction of travel; however the same path might have aπows pointing in opposite directions if the path is direction-neutral The main coπidor through the floorplan has a distinctive visual appearance Other paths or coπidors are of a different color, and are thinner In some configurations of task steps, one or more of these secondary paths may physically overlap the ma path For example, if a task has several entrances and blessed egresses, secondary paths running between various pairs of entrances and egresses might cross a main path that runs straight through the middle of the floorplan This visual artifact can be used to indicate commonly trodden paths through vanous steps and even tasks A similar approach can be taken to look at the path taken by an individual customer, or customer type In order to be able to limit the number of characters on screen to something manageable even for high traffic sites, both filtering and sampling is provided Filtenng is provided by customer profile, cuπent mood state, and eventual outcome (i e , did complete task successfully, did not complete task successfully ) One, two, or all three of these dimensions may be specified simultaneously Note that "eventual outcome" is only available in the retroactive/tape version of the system Sampling is random, but will occur within the bounds of the filtenng

Fig 4(a) shows a filter panel 440 that allows the characters to be filtered on the basis of any or all of a) customer mood (satisfied, confused, bored, or angry), b) outcome of visit (I e , dead, purchased, shopping), or c) demographic information (for example, domain of customer or customer IP address) In this example, the user has indicated that he only wishes to see characters coπesponding to customers who have a " com" domain The user has also asked to view a window 442 showing the key to characters that represent different demographic information In this example, the different characters have differently colored shirts As showed in Fig 12(c), the user can indicate as a preference whether he wants to filter by demographic information or by mood state In this example, the user has indicated that he wants to filter by demographic information Thus, only characters shown representing the ' com ' color are displayed and all have the " com" shirt color

If detailed information is available for a particular room from observer agent 252 in Fig 2(b) the customer information can be made more precise, since the system would have available to it information such as (roughly) where on a page the customer was (Specifically, the obser\ er agent 252 can detect scrolling events, mouseovers, and other events that yield a fair idea ) When this is the case, certain embodiments display a video camera m the room to so indicate In this event, motion within the room is relative to the position of the customer within the coπesponding web page. (It should be noted that, in this embodiment, a room coπesponds to a task step, not a web page. There can be arbitrarily many pages associated with a single task step.)

Fig. 4(b) shows an example of a midlevel floorplan view 401 of task steps in a task, where mood data is being displayed. The task steps coπespond to rooms 462, 464, and 466 (rooms 464 and 466 are labeled as their coπesponding tasks: "enter search criteria" and "display results"). Fig. 4(b) shows a filter panel 450 that allows the characters 420' to be filtered on the basis of any or all of: a) customer mood (e.g., mood is satisfied, confused, bored, or angry), b) outcome of visit (e.g., dead, purchased, shopping), or c) demographic information (e.g., domain of customer or IP address). In this example, the user has indicated that he only wishes to see characters coπesponding to customers who have a mood of confused, bored, or angry. As showed in Fig. 12(c), the user can indicate as a preference whether he wants to display demographic information or mood state. In this example, the user has indicated that he wants to display mood state. Thus, all characters shown in Fig. 4(b) have one of four colors of shirts - the shirt colors representing the satisfied, confused, bored, and angry moods.

Fig. 4(c) shows example characters in various moods. In the Figure, moods include angry, bored, confused, dead (two types), satisfied, purchased, and shopping. Certain embodiments animate the mood characters, as described below, although other embodiments merely display a non-moving representation of mood and merely flip the character to a new non-moving mood representation when the customer's mood changes. An angry mood state 482 is represented here by "curse words" above a character shaking his fist. A bored mood state 484 is represented by a character looking at his watch. A confused mood state 486 is represented by a character scratching his head with question marks. A blessed (normal) death state 488 is represented by the character fading away. In contrast, an angry death state 490 is represented by the character disappearing in a puff of smoke. A satisfied mood state 492 is represented by the character standing in a neutral position (or by the character clicking is heels if he is extremely satisfied). When the customer completes a purchase task, his character carries a shopping bag 494. When the task type is "shopping", the character pushes a shopping cart 496. In certain embodiments, the amount of items in the cart indicates whether the customer is actually shopping or is merely browsing with an empty cart.

In addition, in certain embodiments, if a task type is 'registration' (not shown) the characters do not walk around, but rather walk promptly over to a desk or table, take a seat, and start filling out a form The hand the character is using to fill out the form will move back and forth to indicate that the form is being filled out At some point, either they finish the form and leave it at a customer service counter, or they ball up the form and throw it in a nearby trashcan (if they abandoned the site or task,) or just leave the form on the table if they saved it for later Thus, m certain embodiments, what motion is appropπate is dependent on the function of the room

The definition of mood states varies m different embodiments and any appropriate definition can be used The following paragraphs provide examples of mood states used in the embodiment of Fig 4(c) Happ\ (Satisfied) 492

In one embodiment, the fundamental definition of this state is that the customer appears to be making good progress towards achieving their desired goal Specifically, there is no evidence that they are encountering any difficulty in getting done what they came to the site to do In one embodiment, customers can enter the happy/satisfied 492 mood state in several ways First, the default in an embodiment is that a customer is considered to be in the happy 492 mood state when she first aπives at the web site, since there is at that stage nothing to observe which could be indicative of a problem Second, should a customer enter any other mood state, she can pass back into the happy 492 mood state in several ways, including (1) by performing a productive search on the web site, or (2) by completing the task that she is cuπently engaged in, without exhibiting any further of the below descπbed behavior which would otherwise place her into any other mood state (such as falling out of the task before returning to it again, perhaps by happenstance) or (3) by showing successful sustained intent to complete another task after having abandoned the first task

It is to be noted that if a customer's searches are productive (that is, they yield more than zero and less than an unreasonably large number of results), then this facilitates a happy 492 mood state for the customer This is due to the fact that this action is highly suggestive of the successful initiation of a fresh task Further, if a sequence of searches is conducted, it is determined, m one embodiment, whether each subsequent search is a successful refinement of the previous search If so, this encourages a customer's mood state to change into a happy 492 state since it denotes progress in the customer's attempt to accomplish her own task regardless of whether that task is a site specified one or not In one embodiment, for purposes of the report generator 318, a customer is accounted as being m the happy mood 492 if and only if he never entered any other mood during his visit This definition is necessary because otherwise all of the customers would be accounted as happy 492, since all customers start off in this state With respect to a particular task, a customer will be accounted as being in the happy 492 state if he left the task in a happy 492 state This is to ensure that a customer who was confused as a result of a pπor task, but who became unconfused while in the new task can be properly accounted as "happy" with that task

Confused 486

In one embodiment, the state of confusion is evidenced by repeated, partially but not completely successful attempts to make forward progress towards completing a goal In essence, these are cases of 'two steps forward, one step backward ' Mathematically, this can be expressed as functions describing the paths of confused customers across a task graph having both a high proportion of duplicate values (that is, values which occur more than once) and a high re-occuπence of certain particular values of y (e g f(N) = y), where y is the unique number of the node in the task graph, and f() takes the number N, where N is the Nth step of that particular user dunng a continuous session on the cuπent site It is to be noted that this includes visits to any related or partner sites on which the customer may be tracked as they were on the initial site

More generally, in one embodiment of the invention, the function must contain at least one cycle outside the permissible context of a set of browsing steps in order to connote a confused 486 state In another embodiment at least two such cycles are needed to connote this state The greater the number of cycles the number of repeat cycles, and the number of nodes contained in each cycle, the more severe the problem It is to be noted that scrolling up and down within a page is considered a cycle for this purpose In one embodiment, cycles that include the successful completion of a task are not counted in this analysis In one embodiment, a small number of simple or atomic cycles indicate mild confusion, while repeated cycling through the same steps indicates severe confusion, as does straying far from the task at hand (that is, the case of non atomic cycles)

In one embodiment, if successive searches by a customer oscillate between producing zero results and a very large number of results, this leads to a confused 486 and eventually an angry 482 mood state, if the unsuccessful or 'head banging' behavior continues For instance, if a customer enters the search term "1-ιron," then "1 iron," and then "one iron," followed by "one-iron," and so on, and all of these searches are non-productive, the customer would initially enter a confused 486 state Once the same search(es) start to repeat, an angry state is flagged in one embodiment (It is to be noted that in one embodiment, when creating reports, failed searches that occur with any non-trivial frequency should be checked against the appropriate fields and meta-fields of the site s inventory database So, for example, if the site was out of 1 -irons, and therefore was not cuπently listing them, or if this was a commonly occurπng request even if the site never sold 1 -irons, such information can be provided to the web-site owner or merchant by the report generator 318) Large oscillations in returned result size is indicative of confusion For example, a search that retrieved 50,000 results followed by another related search that returned 0 results would be indicative of a customer stumbling around the web site It is to be noted that the number of results returned can be evaluated by looking at the size of the returned page While any single search may be mood-neutral or even mood positive, a single search coupled with an immediate back track from the results returned page indicates confusion

In one embodiment, customers are labeled as having entered the confused 386 mood state when they circle between different steps repeatedly (In one embodiment, there is an exception to this, which are so-called 'repeated browsing' tasks in which this behavior is more contextually appropriate For example, browsing different hotels In this case, such behavior is considered neutral or acceptable until the number of cycles performed exceeds the number that is statistically associated with a successful completion to the task, specifically at the point at which the probability of a successful completion noticeably decreases with the next trip ) In another embodiment, customers are labeled as having entered the confused 386 mood state when they traverse a spiral pattern (or a "flower" pattern where each petal is a loop from a central node) on a task graph, for example, labeling a series of steps in a task as A,B,C,D, a spirahng path would be A->B->A->C->A->D (The spiral or flower is a special case of the cyclic behavior discussed earlier) In >et another embodiment, a customer will be deemed confused 386 if he falls in and out of several tasks and then revisits them withm the same session Repeated visits to pages that have been designated to the sighted constructor as being help pages are also indicative of confusion This will occur with only one cycle

In one embodiment, other indicators of confusion are • A statistically significant number of mouse-overs of specific images This will appear in the reports with an automated suggestion to enlarge the image if it is below a certain size, to include text in the image, or to simply reconsider the choice of imagery used

• A statistically significant number of customers starting to fill in a form (as determined by a field editor being invoked) but then abandoning it This will appear in the reports as a likely form or page name problem

While neither of these actions by themselves can flag a confused state, if one or the other of these is detected m conjunction with two scrolling events in opposite directions and the customer leaves the site or falls out of the cuπent task to an unrelated one, a state of confusion will be flagged retroactively

In one embodiment, with respect to site-wide mood state reporting, any customer who entered a confused state dunng their visit will be accounted as confused With respect to task- level reporting, a customer will only be accounted as confused if she fell out of a task while in a confused state, the sequence of traversals that initiated the flagging of a confused state occuπed with that task, or the same step in that task has been visited more than once in the same series of "confused" actions The step that is visited most in this scenano is said to be the locus of confusion, and the report generator 318 will log warnings, and account the cost of the lost users, to this step

In one embodiment, for purposes of real-time predictive analysis, the probability of a state of confusion for a particular user is as follows The first time a customer commits a traversal that results in her ending up in a lower numbered node than the one they started on (where the numbers progress monotonically upwards from the root node,) the probability of a confused state being signaled is the same as the probability that this transition or edge traversal belongs to a set of confused actions as opposed to a set of normal actions (Specifically, a set of contused actions is the full, continuous sequence of traversals from the time that the initial divergence from the task graph began to either the point where the user left the site, or the start of a change to another mood state is detected)

Assume a customer w's traveling path at ι'^h step is denoted by N = {n_u n_ul, n_w } , here n s denote the task graph nodes that the user has traversed Define j = mm{r d(n_u ) - d(n_u ) = -1} , where d( ) is the "depth" function from the root, which can only be determined from the task graph The "set of confused actions" is determined from histoπcal data for a particular task Y, specifically, for each confused customer h for task Fin the histoπcal data, the traversal path is denoted by N^h = {N_hl,N_h2 } The "set of confused actions" T₍ is defined to be

{N_hl N_hl c N^h ι = mιn{ k N_hk g Y}, j = mιn{ N_hlc g Ω or mood change at N_hk }, h e H}

where Ω denotes the complete site and H denotes all the confused users for task Y in the histoπcal database

With the above notations, the formula for the probability of the user being confused is

P(confused | N ) = P(confused \ T_C) .

Here P(confused \ T_c) is to be computed from historical data For further details, please refer to copending application number ₁ entitled "Electronic Shopping

Management Intervention" which is hereby incorporated by reference herein

Some of the types of site design problems that tend to contribute to a statistically significant number of customers becoming confused are as follows an extremely large numbers of links (e g. greater than 80), inconsistencies m displayed navigational bars, inconsistencies in look and feel (for example, change in style sheet), alternating or banner style navigational bars, appearance of badly formed task phrases, large numbers of circular references, missing 'next step' or alternately too many next steps, lack of a lexical match between link anchor and destination page (l e. non-stop-list words in link anchor should appear in destination page), lots of inline help links

Bored (Distracted) 484

In one embodiment of the present invention, customers can enter the bored 484 state when for a predetermined number of consecutive page visits that do not e along a monotomcally increasing path on the task graph (that is, the index of the nodes is continuously increasing), the length of time spent on each page by the customers is substantially lower than expected For example, if a customer only sequentially completes a small number (such as 2- 3) of steps of a task that is 7 or more steps, and then jumps to another similar task, and then abandons that one after a small number of steps as well, and stays for a very short period of time on each, this would tπgger the change in state to boredom In one embodiment, this can be measured by assuming that for any given page, the average time taken by customers to read the page would form a noπnal Gaussian distribution, and that customers labeled as bored 484 e on the left tail of this Gaussian distribution The reasoning behind this algoπthm is that bored 484 or disinterested customers tend to very quickly tour through several, generally unrelated pages, often m an attempt to find some interesting material to latch onto

In one embodiment, boredom is measured by seπes of actions that do not a) coπespond to any existing or logically "missing" or virtual path in the task graph, b) do not result within X clicks in arnval at a blessed egress or final step node, regardless of path taken to get there The value of X can be calculated by doubling the length of the average path length from the home page to a blessed egress It is to be noted that, in one embodiment, if, in this scenano, there are indications of scrolling behavior or image rollovers then a confused state will result instead of a bored state However, these cases will rarely fall mside the statistical test for boredom on the time side A state of boredom 484 is ended by an attempt at a search, or any evidence of a sustained attempt to initiate and complete a new task

The pnncipal difference between boredom 484 and confusion 486 is that confusion implies that the customer has a clear goal, but is having difficulty figuring out how to attain that goal m an efficient manner Boredom, on the other hand, is a lack of any detectable goal The difference between confusion 486 and anger 482 is that confusion can be defined as a senes of different trial and eπor attempts to achieve a goal, while anger, as descπbed in further detail below, is a seπes of repetitive actions that are clear evidence of sheer frustration, rather than ongoing forward process towards achieving a goal

In one embodiment, with respect to mood state reporting on a site-wide basis, a customer will be accounted as bored 484 if she ever duπng her visit entered that state With respect to task-level reporting, a customer will only be accounted as bored if she fell out of the task while in a bored state or if the start of the chain of traversals that initiated the bored behavior started in that task This is to retroactively catch the case before the boredom counter was incremented the first time

In one embodiment of the present invention, for real-time predictive analysis, the customer will be flagged as bored at the point at which the behavior indicating a possible bored 484 state has repeated itself enough times so that there is an histoπcal probability 'p' or higher that a fallout from the site will be the immediately subsequent event The value of p is determined by business knowledgeable users, in order to give them fine-grained control over how proactive to be about flagging potentially bored customers

A system in accordance with an embodiment of the present invention considers the following types of site design problem as likely to be causing boredom too many steps in tasks, task inteπuption with links leading away from the task at hand, large numbers of links

Angry (Unhappy) 482

In one embodiment, customers are labeled as entering the angry 482 mood state when they are already in a confused state, and subsequently meet the criteria to again be flagged as confused Specifically, the angry 482 state in this case is flagging repeated, unproductive circles It is to be noted that scrolling up and down within a page is considered a cycle for this purpose Any attempt to start filling out a form will be accounted similarly, since this is an exacerbating factor if it was done m enor and/or will have to be repeated In one embodiment, repeated non-productive searches indicate the angry 482 mood state Repeated access of pages designated as eπor pages are indicative of an angry 482 state Excessive repetition of the same sequence of actions within the same session outside the context of a browsing task will flag an angry 482 state if it exceeds the norms of the populations Note that usages of the browser's 'back' and 'forward' are considered as repetitive actions and qualify to flag an angry state Note further that while a user will transition from happy -> confused -> angry, it is possible to go straight from happy -> angry

In one embodiment, with respect to site-wide mood state reporting, any customer who entered an angry 482 state during their visit will be accounted as angry With respect to task- level reporting, a customer will only be reported as angry if she fell out of the task while in an angry 482 state, entered the angry state while in the task, or the chain of traversals that initiated the flagging of an angry state was initiated with that task It is to be noted that an angry 482 mood state is different from the confused 486 or bored 484, because customers are anecdotally most likely to remember and complain about such a state afterwards

In one embodiment of the present invention, for real-time predictive analysis, the customer will be flagged as angry at the point at which the cycle of behavior that initiated a possible angry 482 state has repeated itself enough times so that there is a histoπcal probability 'p^' or higher that a fall out from the task will be the immediately subsequent event The value of p is determined by a business knowledgeable user, in order to give them finegrained control over how proactive to be about flagging potentially angry customers.

A system in accordance with one embodiment of the present invention will consider site design problems that led to the original bad state - in most cases confusion - as being likely responsible for causing anger. The attribution is done on the basis of the severity of the initial confusion, using both the duration of the confused state, and the number of triggers for the confused state, as well as the percentage of customers attempting the task who are impacted, as indicators of severity.

A confused 486, bored 484, or angry 482 mood state is cleared, in one embodiment, by the initiation of a search by the customer, if the previous action was not a search, and if the new search is not a repeat of a search previously entered during this session, as this signals an attempt to initiate a new task. It is to be noted that in one embodiment, if the attempt is not successful, the customer will not become "happy" merely as a result of changing from one task to another. In one embodiment, a confused 486, bored 484, or angry 482 mood state is also cleared by the customer traveling three nodes along a known (existing or virtual) path in the task graph of the web-site. Such bad mood states can also be cleared, in one embodiment, by any other set of actions that is clearly indicative of a sustained effort to complete a new task.

The following discussion illustrates the various general states that a customer can be in:

Shopping 496

Customers are labeled, in one embodiment, as being in a shopping 496 state when it is detected from their URL page clicks that they have begun using a shopping cart. In one embodiment, customers are labeled as being in a shopping state when they have added an item to their shopping cart. This data is obtained by integration with the site's shopping cart solution. They transition from this state if they remove all items from their shopping cart, or if they pass successfully through the checkout line.

Purchased 494 In one embodiment, customers are labeled as being in a purchased 494 state if they've successfully completed a shopping transaction; that is, that they have had items in their shopping cart and that they have subsequently successfully completed a check-out process. This data is obtained by integration with the site's shopping cart solution.

Dead 488, 490

In one embodiment of the present invention, customers who have had no activity within a predetermined, configured amount of time (time-out duration) are considered timed- out and thereby 'dead.' In that embodiment, a customer is actually marked as 'dead' after a certain amount of time following his last recorded activity. This amount of time is predetermined and configurable and can be different from the time-out duration. In one embodiment, the time-out duration is weighted by certain characteristics of the step, such as size and page content complexity (i.e. a large number of images).

There are several different ways in which a customer can die. These include neutral death, "blessed" death, and "bad" death. Neutral death is when the customer simply leaves the web-site while in a happy/satisfied 492 mood. That is to say that while the customer left the site without successfully completing a task, there is no evidence that they experienced any difficulty in using the site. The customer just simply, for whatever the reasons, failed to make a purchase on that particular day. A "blessed" death is when a customer leaves the web-site through a "blessed egress." A blessed egress is a step in a task through which it is permissible from the merchant's point of view for the end user to leave the site. This is typified by a checkout step. A "bad death" is when the customer leaves the web-site while he is in one of the bored 484, angry 482, or confused 486 mood states. As mentioned above, the bored 484, angry, or confused 486 mood states are often collectively called the "bad" mood. states. Simply put, it is undesirable or bad for the web-site owner to have customers in one of these mood states, and it is a bad death when a customer leaves the web-site while in one of these mood states.

Certain embodiments also include a reincarnation mode in which a character representing a customer who is revisiting the web-site initially appears with a flash of lightning and a puff of smoke (possibly with sound effects) to indicate that the character has been reincarnated. Any appropriate technique can be used to determine whether a character has visited the website previously (during some defined time period). For example, the system may track a user via cookies, via his IP address, via an optional or required sign-in procedure, or tracked by via http parameters such as GET and POST. In addition to the mood representations descπbed above, other mood states can be used singly, or in conjunction with one another to connote progressive states of customer mood For example, a character representing a customer who has just become angry might place his hands on his hips If he is still angry two minutes later, he might start jumping up and down The following is a non-mclusive list of other possible mood states that might be used to represent a customer's mood and the character actions that indicate the mood states Angry stomping feet face turns red • jumping up and down hands on hips throwing hands in air scowhng/frownmg holding head m hands • crossing arms cartoon angry symbols above head steam off of head rolling eyes sobbing • clenched fists shaking if pushing a shopping cart, then getting angry and, while angry, removing a the sole item from the cart, shoving the cart into a wall and abandoning it there (if sound is enabled, an appropriate crashing noise should be made) • if caπying a folder while becoming angry, throwing the folder in the air and scattenng little pieces of paper on the floor (that eventually blow away)

Satisfied/Happy smiling nodding whistling while walking around • tipping imaginary cap skipping around doing a little jig gn ing the "thumbs up" sign as they transition successfully from step to step • if pushing a shopping cart, using the cart as a scooter. Even doing wheehes with it

• if carrying a bag, swinging it back and forth

Bored looking around (mainly up and around) tapping foot arms crossed across chest talking on cell phone, one hand to their ear playing a gameboy or similar device • looking at watch sighing, via a callout whistling - while standing still

Confused • fuπowed brow

• question marks above head

• looking around, broader head motion

• shrugging

• scratching head • hands on hips

• walking in small circles

• looking at a store directory (where the store directory is a static image placed in an "entrance" room)

Blowing up in a bad state (bad death)

• disappeanng in cloud of smoke, for example, one with a reddish hue

• melting into a puddle

• ground opening up and swallowing the character in a dark hole

• dollar signs melting, going up m smoke, or disappeanng underground • if caπymg a folder, small pieces of paper debris float around until they burn, liquify, or get swept underground by the cuπent when the character gets sucked underground

• if pushing cart, the cart rattles around for a few more frames after the character has expired, then it blows up, melts, etc, independently Death after exit from a blessed egress - after walking out the door (good death)

• fading out - no smoke or steam

• waving and walking off into the distance • stopping, waving and getting taken away via whirlpool effect

• skipping or jigging out the door with bag swigging if a purchase was made (if never unhappy at any point in the session)

Neutral Death - no bad state, but before completing a task • fading away

• beaming away (science fiction motif)

• disappearing amidst a puff of white smoke, vapoπzmg

• if caπymg a folder, pieces of paper (and folder) drop on the ground, eventually blowing out of sight • if pushing a shopping cart, cart rattles around for a few frames after the character vanishes, and then disappears m the same manner as the character

Sound Effects

• A "poof noise when characters vaporizes • A small explosion noise when an angry character bites the dust

• A bell or door opening noise when customers enter or exit a task

• A crash when a shopping cart hits a wall

• A beaming noise

• A trap door and screaming noise if character falls into it

Fig 5 shows an example of room enlargement If the user clicks on an empty patch of floor in a room 502, a new window 504 is brought up that is an enlargement of the room whose area was clicked on This feature helps the Cam user to more easily select a particular character It is also useful to visually disambiguate slight visual differences between different types of characters, for example a male and female character of otherwise the same customer profile The enlarged room 504 appears substantially the same as the original room 502

Fig 6 shows an example of allowing the user to record a snippet of the display A "record" button 332 allows snippets to be created The Cam user hits the "record" button 332 at the point in time he wishes to start the recording process After he has recorded what he wishes, he hits the "stop" button 333 to stop recording (or hits the "record" button to toggle recording) He can then select a name for the new tape snippet (not shown), and hit the "save" button on the dialog box that comes up for this purpose The system saves the information required to re-generate the display generated during recording as a separate file or document Alternately, the user can open a recording panel 602 from a drop down menu "Tape" that allows the user to stop and record

A samphng panel (not shown) that allows the user to specify either a fixed sampling percentage to be used for the entirety of the tape or else select "autosampling" which will automatically adjust the sampling rate based on cuπent traffic To aid the user in making the right decision with respect to sampling, this panel contains a thumbnail image of the most trafficked floorplan at a time of average traffic In the case of tapes than span several hours or more, a button is provided that displays one such thumbnail per hour to give the user a better picture of the impact of his sampling across the entire length of tape These actions allow the user to preview the impact of his sampling decisions before setting them m motion

Moreover, in certain embodiments, a character can represent more than one customer if the customers are performing substantially the same tasks

Fig 7 shows an example of allowing the user to follow a particular customer and/or display data about a particular customer Fig 7 shows an individual customer's property panel 702 In the example, the user has clicked on an individual customer 704 in the floorplan view This action causes panel 702 to be displayed This configurable panel can display all known information about a particular customer that is selected in the Cam screen Fig 12(f) shows a panel to allow the user to configure what information will be shown in panel 702 All characters except the selected character are shown as ghosts in this embodiment

Fig 14(b) shows an example format for data about a particular customer This data can include information that is known from a customer information database or cookie database as well as data such as the customer's ID 1402, the customer's IP address 1404, and machine name 1406, (which is obtained by doing a reverse DNS lookup) Scrolling history lists 1412 of each known step and task the customer has visited during this session can be displayed in this panel, as is the associated history of moods 1410 This information can be presented in either textual or graphic form The data can also include registration data 1408 If the latter, a titled frame from the animation of that mood state is displayed instead of just the name of the mood state, and, where there are relevant static images associated with particular rooms, a titled version of this gif can be substituted for the text name Which kind of display is used is left to a user preference The panel also optionally contains a record button (not shown), to dump its history contents to an HTML file for subsequent review

Customers can also be spotlighted (not shown) The character of a spotlighted customer is shown as "Marked" for easy recognition For example, a user may wish to spotlight a customer that is frequently bored When this customer's character appears in the future, it will be visually marked by, for example, showing a spotlight graphic on that character More than one customer may be spotlighted simultaneously, but only one may be tracked If a customer is tracked, the Cam view will follow him as he moves from task to task Specifically, if he leaves a task or task step, the Cam view will follow him down the path to another task/task step. If he leaves the site, an appropπate message appears m the status bar, and tracking is discontinued Other embodiments will not follow a tracked customer out of a room

Fig 8 shows an example of allowing the user to search for certain customers This function is initiated by, for example, the user selecting search icon 314 For example, the user has decided to search in the cuπent task 802 where there are many people 804 who have completed this task 806. The system will fetch a temporal portion of the display that complies with these search parameters In the example, there are no occuπences in the cuπent task where many people have completed the task The search capability uses markers for anomalous events that were placed in the tape. It allows such events to be searched in both directions of time in retroactive mode. Specifically it allows fast forward between statistically outlier events on either side of the curve

Figs. 13(a)- 13(c) show the possible search parameters in the described embodiment. Searching can be performed in the cuπent task, or in one of the other named task steps in the cuπent task Searching can be for a time when there are many or few customers who have completed the task (or task step) or fallen out of the task (or task step) In other embodiments, event types include unusually many or few people a) completing a task, b) getting in a particular mood state, c) leaving the site in a particular mood state, d) attempting a task, or e) falling out of a task. Note that each of these event types can be naπowed to apply to only a subset of customer profiles. They can also be broadened from the scope of an individual task to apply to the whole site. Certain embodiments allow a search on individual customer level, such as by IP address.

Certain embodiments include the ability to combine the results of searches as a new display. Using a panel similar to the one used for the search function, a tape can be made which, instead of reflecting a span of chronological time, contains a configured number of occuπences of a particular type of event. For example, a Cam user might want to see 50 premium customers all trying to complete a certain task at the same time. If in fact there were 50 premium customers trying to complete the particular task at any given time, this can be accomplished by filtering out all other types of customers. However, if it were never actually the case that this many premium customers attempted the task during the time span in question, a different approach must be taken.

In this "time shifting" approach, the Cam user may specify one or more customer types she is interested in observing, and one or more tasks she is interested in observing. She must also specify a number of customers that she would like to see. She may also specify other dimensions, such as mood state. The cuπent tape (if there is one) is searched for the desired number and kind of occuπences. If the cuπent tape does not contain a sufficient number of the desired events, other tapes will be searched through if they are known to the search function. A new tape will be constructed that is of the length of time needed for the longest task attempt to complete, of all of the tasks specified by the Cam user to examine. The regular clock in the TV frame will be replaced with a stopwatch that does not include a date when such a "virtual" tape is loaded. The new tape will include only the requested kind(s) of customers who fit the criteria specified by the Cam user, for example, who attempted a particular task. All customers will start the task at the same moment. This makes it easy to, for example, detect differences in task performance between two kinds of customers. The automatically generated description of the tape will include information about the regular tapes it was distilled from.

Fig. 9 shows an example of an interface that allows the user to look at statistics about the customers while the display is being generated (e.g., while the characters are moving). This function is initiated by, for example, the user selecting statistics icon 316. A configurable statistics panel 902, like an odometer, has a reset button 904 to reset all the statistics to 0.

Statistics can, by user preference, be displayed as raw numbers, in percentages, or with smiley faces, frowning faces, and a straight face to indicate whether performance meets predefined standards. Which statistics to display, and which tasks to display data for is user configurable through user preferences. Such statistics include, for each task, how many customers got into each mood state how many fell out, and how many completed, how many purchased, how many left the site, and how much money was lost to site design problems. Further metrics of interest, such as whether a certain usability goal is being achieved, can be added programmatically. Anomalous data appears in green or red font, depending on whether it is positively or negatively anomalous. The panel tracks the movement of the tape; if, for example, the tape is rewound 2/3 of the way, the statistics panel will roll back to that marker.

In the retroactive version that uses taped data, the system can display snapshots 310 representing different times of day in some templates of the Task Manager reports. Thus, in certain embodiments, the cuπent UI allows snapshots to be taken for time periods that hadn't been played from the tape yet, but were already recorded on it.

It should be noted that the described embodiment also includes a generate report function 318. An example report is shown in Table 1, which forms a part of this disclosure and is herein incorporated by reference.

Fig. 10 shows an example of a problem indicator 1002 on the display. In the described embodiment, a problem indicator can be shown as a sawhorse (in the example) or a traffic cone, although other danger indicators can be used as appropriate. Orange striped sawhorses are erected in a room when site design problems were discovered in that room by analysis of the site (see patent #1), and there were noticeable problems in user behavior that were presumably caused as a result of the flaw. There are three levels of sawhorse to indicate the severity of the problem observed. In the most severe case, the sawhorse has a blinking orange light on top of it. In less severe cases, the light doesn't blink, or isn't present.

If the Cam user clicks on the sawhorse, it will launch a browser window 1004 that contain information about the problems that were detected as well as statistical and cost to merchant information during the time span of that tape. Note that this information, including infoπnation about design eπors detected in the site at that time, is stored and versioned so that the Cam user will always see information about design problems that existed at the time the tape was cut, and not after or before.

If there is detailed information available from observer agents 252 for this room, the position of the sawhorse within the room is roughly similar in position to the position in the page where the customers seemed to start experiencing difficulty. Otherwise the position of the sawhorse within the room is random. Only one sawhorse will be drawn, regardless of how many distinct problems were found. If customer problems were found throughout the page, the sawhorse will be rendered in the center of the room. A variety of different visual artifacts can be placed in individual rooms to indicate that the room received a relatively high level of traffic during the time span covered by the cuπently loaded tape. These include the type of orange cone commonly used in drugstores and supermarkets to indicate that they are cleaning, or various debris such as debris or bottles lying around.

Fig. 11 shows an example of traffic monitoring on the display. In the described embodiment, a sleeping security guard animation 1102 will appear in any room in which either the traffic is less than one half the average traffic for that room, or in which the traffic is close to zero. These two cases are differentiated from one another by the fact that the sleeping guard has three states. In the first state he is squarely sitting on his chair, with small "zzz'"s coming out of his mouth. In the second state, the "zzz'"s get larger and the guard is very noticeably slumping on his chair. In the third state, the guard falls off his chair and is lying on his back on the floor with large "zzz'"s coming from his mouth. If there is a non-neglible amount of traffic in the room even if it is a below average amount, the guard will never progress past the first state. The guard progresses from the second state to the third state if the duration of time that there is negligible traffic persists for a long period of time.

If the Cam user clicks on the sleeping guard, a panel 1104 will appear with a traffic history that provides information on the maximum, minimum, median, mode, and average traffic during this tape, as well as offering a comparison to tapes that covered other analogous time periods (e.g. a tape that covered last Monday instead of this Monday,) assuming that this information is available.

Cobwebs, dust, or similar appropriate indications of disuse can be used indicators of longstanding disuse of a room instead of a sleeping guard, which implicitly presumes that customers will show up in the room at some point, even if not right at the cuπent moment. Certain embodiments also contain static images in the room, that are used to cue the

Cam users as to the type of task step represented by that room. Static images suggestive of the function of a room can help the Cam user initially identify the room's purpose. Such static images include but are not limited to: a) a help desk in a help or general information room, b) sets of checkout lines with cashiers and cash registers in a checkout step, c) sets of stacked up shopping carts in a browsing task, d) a blinking promotion sign with word "special" on it for promotions, e) Store directory in entrance or home page, f) escalators to go view another floor of merchandize, g) a thin table as exists at banks for filling out deposit slips (for registration rooms.) Figs 12(a)- 12(f) show examples of various preferences that can be set by the Cam designer in the descπbed embodiment Fig 12(a) shows a "Playback" tab 1202 in a preference window 1204 The playback speed determines a proportion between the speed at which the customers interact with the website and the speed that the characters move through the rooms on the display In the described embodiment, the user can chose a proportion of between 1 2 and 1 10

Fig 12(b) shows a "Startup" tab 1212 in the preference window 1204 The startup tab allows the user to specify the view that is first displayed when the Cam software is first executed The Cam software can initially display a campus view, a floorplan view, or a last view shown during the previous execution of the Cam software

Fig 12(c) shows a "Campus Map" tab 1222 in the preference window 1204 The Campus Map tab allows the user to specify whether the characters represent demographic information (see Fig 4(a)) or mood information (see Fig 4(b))

Fig 12(d) shows a "Statistics" tab 1232 in the preference window 1204 The Statistics tab allows the user to specify the whether the statistics are shown as percentages of a total (see Fig 9) or as straight numbers (not shown) The statistics tab also allows the user to specify whether to hold the statistics panel in a fixed position relative to the TV frame (I e , to "snap" to the TV frame)

Fig 12(e) shows a "Room Display" tab 1242 m the preference window 1204 The Room Display tab allows the user to specify whether to display a traffic monitoπng indicator (such as a sleeping guard), and/or whether to display problem indicators (such as parking cones)

Fig 12(f) shows a preference window 1204 This window allows the user to adjust the appearance of the statistics (see Fig 9) by specifying whether completions, fallouts, etc should be included In the statistics of Fig 9, the user has indicated that he wants all possible statistics displayed This window allows the user to adjust the appearance of the customer properties (see Fig 7) by specifying whether the IP address, present room, etc should be included In the individual customer properties of Fig 7,the user has indicated that he wants all possible customer information

B Setting up the User Interface and Data Structures

Customer data Fig 14(a) shows an example format for customer tracking data As shown m Figs 2(a) and 2(b), the customer tracking information is passed to the Cam software so that the Cam software can use the customer tracking information to generate the display of characters in the campus view of floorplan view The format shown is only one example of customer tracking data Any appropπate data format can be used that reflects user interaction with a website The customer tracking data includes at least a customer ID 1403, a room/task step that the customer is presently in 1405, a mood that the customer is presently in 1407, and a cuπent time for the customer 1409 Alternately, the Cam software can determine the mood and other pertinent data for the customer based on low level data passed to the Cam software

Fig 14(b) shows an example format for data about a particular customer This data can include information that is known from a customer information database or cookie database as well as the customer's IP address 1404 and machine name 1406, which is obtained by doing a reverse DNS lookup Scrolling history lists 1412 of each known step and task the customer has visited duπng this session can be displayed in this panel, as is the associated history of moods 1410 This information can be presented in either textual or graphic form The data can also include registration data 1408 If the latter, a titled frame from the animation of that mood state is displayed instead of just the name of the mood state, and where there are relevant static images associated with particular rooms, a titled version of this gif can be substituted for the text name Which kind of display is used is left as a designer preference The panel also optionally contains a record button, to dump its history contents to an HTML file for subsequent review

In a prefeπed embodiment, customer data is queued as it is received by the Cam software In this embodiment, each record is structured as follows Note that with the exception of possibly customer ID, customer profile type, and Intervention History Value, in one embodiment all of this information is on a per session basis only In another embodiment, the information could span an amount of time specified by the administrator, assuming that there is a mechanism in place to reliably access unique customers across sessions Data retained for each customer includes

Customer Information

Customer ID Customer Profile Type(s) Total time spent on the site this visit Last checked time stamp Intervention History Value

Cuπent Mood State

State code

Number of flaggings Total time in this mood during this visit

Probability of diagnosis

Cuπent Task

Task ID

Task importance for this type of customer Time customer has spent m this task

Cuπent step

Previous Task(s)

(same as structure for cuπent task, but with one additional field, a BOOL, completed⁷)

Previous Mood State(s) State code

Number of flaggings From these, two additional fields are denved

Age = the total amount of time they have spent collectively in the bad states duπng this visit

EDITOR INTERACTION The appearance of a campus view and a floorplan view are determined by a Cam designer Shapes can be drawn onto the editor window in the same way as in any paint program In one embodiment, a gπd, useful for resizing and moving graphic objects, always has ' snap" mode turned on (although there will be a preference choice to visually turn the grid off) The rationale for this is because placement of room graphics next to one another is not a trivial task when talking about pixel-by-pixel placement of graphic objects Having the grid m snap mode somewhat alleviates this problem

There preferably are various tools in the toolbar window They include

• Add Building (shown in Campus Editor) • Add Building Walkway (shown in Campus Editor)

• Add Campus Entrance / Exit (shown in Campus Editor)

• Add Room (shown in Floorplan Editor)

• Add Room Hallway (shown in Floorplan Editor) • Add Escalator (shown in Floorplan Editor)

• Add Floor Exit (shown in Floorplan Editor)

• Add Line (shown in both Campus and Floorplan Editor)

• Add Shape (shown in both Campus and Floorplan Editor)

GRAPHIC OBJECTS

Graphic objects are created by the Cam designer using the graphics tools listed above All graphic objects are shapes except line graphics, which are treated differently Shape graphics can be resized and moved by dragging the mouse, just like a Cam designer would in any paint program Line graphics can be moved by clicking on the line graphic and dragging the line graphic to it's desired position The individual vertices of line graphics can be repositioned as well, simply by clicking on a vertex and dragging it Mouse clicking on the canvas and dragging will result in a bounding box that can be used in selecting graphics within the bounding box area Multiple graphics may be selected in this fashion

Room Layout

In order to lay out the rooms coπectly, it is necessary to be able to calculate the rectangular bounds (a x, y position, plus a width and height) of each room, plus the direction of the entrance and exits, if the task is an ordered task

Room properties configurable by a display designer include

• Floor Size = expressed in square units, the entire size of the floor

• Threshold Room Height = expressed as a percentage, this percentage represents the height of the floor To get the exact floor height measurements, we take this percentage and multiply it by the 'Floor Size'

For each step in a task, each of which is represented as a room, the editor software determines how large (expressed as square units) to make each room using the following steps la) Based on previous customer traffic data stored in a database, record the total number of people that have visited any single room lb) For each room, take the total number of unique visitors that have visited a room and divide that by the total number of customers that have visited this floor, thus giving you the percentage of: customers to a room / total number of all customers in all rooms. lc) Multiple the percentage from step lb with 'Floor Size', thus giving you the total room size for each of the rooms in a floorplan, expressed in square units.

Next, the editor software determines the bounds (x, y position, plus a width and height) of each individual room:

2a) Determine how many rooms to place in a 'column' of rooms. A 'column' is defined as a single vertical aπangement of rooms ordered from top to bottom or bottom to top. Given the 'Threshold Room Height' percentage and the percentage sizes of each room, determine how many rooms will fit in a single column. So if the 'Threshold Room Height' percentage was 30%, and the room sizes were [ 10% 15% 20% 10% 15%] consecutively, then only the first and second rooms will fit in the column. 2b) For each room in the column, normalize their percentage sizes. So if room sizes in the column were [10% 15% 20%], normalized they would become [22.2222222% 33.3333333% 44.4444444%]. For each normalized room size percentage, multiply this percentage by the 'Floor Height' (the 'Floor Height' is calculated by multiplying the 'Threshold Room Height' with the 'Floor Size') to get the 'Room Height' for that particular room in the column.

2c) To get the 'Room Width', divide the 'Room Size' (from step lc) by the 'Room Height'. The room widths should be the same for all the rooms in a particular column. 2d) In order to calculate the position of each room in the column, we must first determine which direction to stack the rooms vertically (top to bottom or bottom to top). The stacking direction alternates direction between columns. We start from top to bottom for the first column and with position (0, 0) for the first room in the column, and increment the 'y' position by the 'Room Height' of the cuπent room, so that the rooms are all stacked one after another in the column. Note the position of the room is expressed relative to the position of the column. 2e) If the rooms in a column are stacked top to bottom, the room at the top of the column has an entrance to the 'West' of the room. The exit of this room is to the 'South'. If the room is at the bottom of the column, the entrance of the room is to the 'North' of the room, and the exit to the 'East' of the room. For the rooms in between, the entrances are to the 'North', while the exits are to the 'South'. 2f) If the rooms in a column are stacked bottom to top, the top most room will have a 'South' entrance and 'East' exit. The bottom most room's entrance will be 'West' and its exit 'North'. Any rooms in between will have a 'South' entrance and 'North' exit.

2g) Now create another column, and repeat step 2a until the bounds and entrance / exit directions of all the rooms have been calculated.

Floors within a building represent individual tasks. It is important to remember here that a floor can exist in two or more different buildings. Here is the sequence of events for the creation of a semantic graphic that represents a task: 1. The Cam designer creates a building in the Campus Editor (see below)

2. The first floor of the building automatically gets created by the application. One of the purposes of the automated layouts is to have tasks that are overly long or circuitous visually appear so to the user.

3. The Cam designer specifies a task name to associate with the first floor of the building 4. The Cam designer has now created a building with a single floor, which represents a task

Then the Cam designer will want to create steps for the new task:

1. Cam designer drills down by double clicking on a building in the Campus Editor (see below). 2. The application is now in the Floorplan Editor for the first floor of the building

3. The Cam designer can create individual rooms for this floor of the building

4. For each room the Cam designer creates, the Cam designer specifies a task step name. If the room is specified as representing an optional task step it will be rendered to appear differently from non-optional task steps in both the designer/editor software and in the display of the Cam software.

5. If the floor represents a task with ordering constraints, the Cam designer may aπange the rooms next to one another in a manner consistent with how the task steps are ordered

6. The Cam designer has now created rooms on the floor of a building that represent task steps within a task

Now suppose the Cam designer wants to create a subtask in the cuπent task:

1. The Cam designer creates an escalator graphic object by dragging it from the palette to the drawing canvas. 2 When an escalator graphic is created by the Cam designer, the application also automatically adds a floor to the existing building

3 The Cam designer then assigns a task name to this new floor, which implies that the escalator graphic links to the new floor in this building 4 If the floor represents a task with ordering constraints, the Cam designer can choose to aπange the escalator graphic next to the room graphics in a manner consistent with how the subtasks and task steps are aπanged within the task

5 The Cam designer has now created an escalator m a floor of a building which represents a subtask withm a task The descπbed Cam embodiment has three windows The primary mam window is a frame that is used by the Cam designer for creating diagram-like shapes as shown in the Figures herein Another toolbar window (not shown) appears onscreen which is used in the selection of various Canvas manipulation tools Finally, the last window (not shown) represents a properties editor which is used by the Cam designer in configuring canvas, campus, building, floor, and room properties

View Properties

There are vaπous types of graphics as described below Semantic graphics carry meaning in the sense that they represent task and task step model objects whereas cosmetic graphics exist purely for cosmetic reasons Another category of semantic graphics are background images that are used to help visually claπfy the function of the room, such as cash registers in a checkout room These were listed above The other vanous graphic types are

Semantic Graphics Building Shape Graphic

Building Walkway Shape Graphic

Campus Entrance / Exit Shape Graphic

Room Shape Graphic

Room Hallway Shape Graphic Escalator Shape Graphic

Floor Exit Shape Graphic Cosmetic Graphics

Rectangle Shape Graphic

Diamond Shape Graphic Text Shape Graphic Circle Shape Graphic Triangle Shape Graphic Image Shape Graphic Line Graphic

All shape graphics are polymorphic in the sense that a rectangle shape can be turned into a circle shape, can be turned into a room graphic shape, etc. thru the properties editor pane. However, depending on whether the Cam designer is in the Campus Editor or Floorplan Editor, different shapes may not be selectable (for example, the Cam designer is not permitted to turn a building graphic into a room graphic, but is allowed to turn a building graphic into a building walkway graphic).

GRAPHIC PROPERTIES

Fig. 15(a)-15(m) show an example format for properties used to specify the campus and floorplan displays.

These properties are specified in the properties editor window when no graphic is cuπently selected:

Name of tabbed pane: Canvas

Show Grid (boolean value)

Grid Size (integer value set by a slider) Canvas

Size (a width and height value entered by the designer)

SHAPE GRAPHIC PROPERTIES

These properties are specified in the properties editor window only when a shape graphic is selected (all graphic types other than line graphics): Name of tabbed pane: Shape Fill color Color (color value)

Tiled Image (image resource) Stroke

Color (color value) Stroke Width (integer value set by a slider) Shape Graphic Type (pull down menu) Possible options are

Available only in Campus Editor Building Shape

Building Walkway Shape Campus Entrance / Exit Shape Available only in Floorplan Editor Room Shape Room Hallway Shape

Escalator Shape Floor Exit Shape Available in both Editors Rectangle Shape Diamond Shape

Text Shape Circle Shape Triangle Shape Image Shape

LINE GRAPHIC PROPERTIES

These properties are specified in the properties editor window only when a line graphic is selected

Name of tabbed pane Line Stroke

Color (color value)

Stroke Width (integer value set by a slider)

IMAGE SHAPE GRAPHIC PROPERTIES

These properties are specified in the properties editor window only when an image shape graphic is selected

Name of tabbed pane Image Image (image resource) TEXT SHAPE GRAPHIC PROPERTIES

These properties are specified in the properties editor window only when a text shape graphic is selected

Name of tabbed pane Text Text (text field)

Campus Editor

The Campus Editor software is where the designer creates building graphics and associates task names with them The designer may also create building walkways by using the building walkway tool A task name cannot be assigned to more than one building The campus entrance / exit graphic is responsible for depicting where customers come from when they enter the site, as well as where they will go when they leave the site It is the designer's responsibility to connect up the building walkways with the buildings m a manner such that all the buildings are connected to each other by some series of walkways, and all the buildings are connected to the campus entrance / exit graphic by some seπes of walkways If this is not the case, the campus layout will not be viewable in the Cam and the designer will get a senes of warnings when he or she tries to save the city layout (although the city layout is still savable and restorable from the underlying database)

CAMPUS PROPERTIES

These properties are specified in the properties editor window only when in campus view Name of tabbed pane Campus Walkway

Colored (color value) Image tiled (image resource)

BUILDING PROPERTIES

These properties are specified m the properties editor window only when a building graphic is selected Name of tabbed pane Building

1st Floor Name (pull down menu)

The pull down menu lists all the available tasks on the site Building image (image resource) Edit First Floor button (clicking on this takes you to the Floorplan Editor and lets you edit the first floor of the building)

BUILDING WALKWAY PROPERTIES These properties are specified in the properties editor window only when a building walkway graphic is selected

Name of tabbed pane Walkway Walkway

Walkway traffic direction (pull down menu) West-East

North-South Colored (color value) Image tiled (image resource)

CAMPUS EXIT PROPERΉES

These properties are specified in the properties editor window only when the campus exit is selected (note only one campus exit can be created) Name of tabbed pane Campus Exit

Customer's exit direction (pull down menu) Possible options are North, South, East West

Floorplan Editor

In the editor, double clicking on a building takes the designer to the first floor of the building, which in essence, is the same as selecting a task Once the designer does this, the Campus Editor screen is replaced with the Floorplan Editor screen This screen is where the designer creates room, hallway, exit, and escalator graphics A task step cannot be assigned to more than one room graphic withm the same task A task, however, can be assigned to more than one escalator graphic In a task that lacks ordenng constraints, the room graphics may be aπanged in any order, provided they do not physically overlap each other (a warning will result when the designer tries to save the city layout) In a task with ordenng constraints, the room graphics must be aπanged in an order which stays consistent with the ordering of the task steps If this is not the case, the designer will receive eπors from the application when he or she tries to save the campus city layout (although the campus city layout will still be savable and restorable from the database). The entrances and exits of a room graphic are calculated automatically by the application as the rooms are placed next to each other. The floor exit graphic specifies visually where customers will exit after leaving the task. When specifying a room task step name, the designer may also choose to specify a "Complaints Room". The Complaints Room is where customers go after falling out of a task. This room does not need to be connected to any of the other rooms which represent task steps. There can be only one Complaints Room per floor.

ROOM PROPERΉES These properties are specified in the properties editor window only when a room shape graphic is selected:

Name of tabbed pane: Room

Task step name (pull down menu)

The pull down menu lists the logical task steps in this task, plus "Complaints Room" (hardcoded)

ROOM HALLWAY PROPERTIES

These properties are specified in the properties editor window only when a room hallway graphic is selected: Name of tabbed pane: Hallway

Hallway

Colored (color value) Image tiled (image resource)

ESCALATOR PROPERTIES

These properties are specified in the properties editor window only when a room escalator graphic is selected:

Name of tabbed pane: Escalator Task name (pull down menu) The pull down menu lists all the available tasks on the site FLOOR EXIT PROPERTIES

These properties are specified in the properties editor window only when the floor exit graphic is selected (note only one floor exit can be created for each floor) Name of tabbed pane Floor Exit Customer's exit direction (pull down menu)

Possible options are North, South, East West

Fig 16(a)- 16(d) show example formats for data used to display the campus and floorplan displays These data include a city payout, a canvas, a graphic, and a graphic attribute

C Other Embodiments The frame of the view that contains the characters is not limited to a VCR, TV, or DVD metaphor Other embodiments of the invention include the following different visual themes space ship console radio console kiosk touch-screen style futuπstic stereo/video device sports scoreboard military vehicle control panel (making it easy to target individual customers) video editing console (for retroactive system) commercial jet plane console a single apartment building looking straight on, apartment windows with lights on are tasks with traffic in them

The characters can be displayed in either 2D or 3D, as can the scenes in which the characters are placed Example scenes might include

• bluepπnt (for use with pre-production sites)

• island region in the ocean (instead of campus view)

• shopping mall style floorplan • building style floorplan

• library floorplan (with the appropriate static images such as shelves of books would be used for research-oriented sites )

• Museum style floorplan In certain embodiments, animations of both men and women characters wearing assorted t- shirt colors are provided to represent customers of different customer profiles In other embodiments, a set of unisex characters is provided, for those cases in which gender data of the customer is unavailable to the system In another embodiment, domain-specific characters are provided For example, a pharmacist customer profile might be represented by a unisex character in a white lab coat or an educational domain may be represented by a student character In any of these forms, the Cam designer can select which animated character will coπespond to which customer profile In yet other embodiments, the customers are represented by animal characters or by geometric shapes or by shapes chosen by the customers of the websites themselves (thus providing information about how the customers view themselves)

In certain embodiments, the campus view, the color scheme of the suπoundmgs can be made to shift gradually from a bnght state to a pitch black state order to help visually cue to the Cam designer as to the time of day or night that was or is being recorded In the nighttime state, the buildings and paths would be lit with spotlights so that they remain visible, and so that no information is lost Specifically, everything other than a small area around the paths and the buildings and the buildings' signs should begin to progressively dim according to the approximate hour of sunset Factors such as daylight savings time and time of year are considered In the floorplan view, there is a similar behavior only the inside of the building, the path, and any signs external to the floorplan will remain at full brightness at all hours

In certain embodiments the base of TV frame display is expanded to include buttons for viewing different "channels ' Each channel coπesponds to a website whose activity may be viewed with an offset in time of several minutes or longer The buttons are configurable, the Cam designer can select whether to name or number them The name or number of the cuπently selected channel will appear supenmposed at the top of the TV screen In order for the Cam designer to be able to see the activity on a site, the Cam application will have to be being served data from a streamer resident inside the firewall of the sites in question When the TV is turned on in this mode it will start up with the channel it was last on unless there is a designer preference specified to start with a particular channel In an alternate embodiment, tapes of specific hours of the day are created automatically and are placed in a location where they can be loaded into the Cam software when the Cam user hits the appropπate button This version might be used by venture capitalists or any other type of Cam user who has both a vested interest in, and the abιht> to demand monitoring pπvileges from such companies In certain embodiments, the TV frame is removed, and the TV screen is used as a screen saver.

In certain embodiments, the Cam user/designer can decide to have several incarnations of the TV set on their screen at once. How many TV screens to have, and what sizes they should be are configurable with a simple drag & drop interface. In order to ensure the usability of the configured interface. Cam designers must select from a variety of different fixed sizes of the TV. In this variation, an additional master control panel is provided that allows, if desired, the different TV sets to be tied together, or driven by the same master directive. Specifically, the master control panel provides a means of causing all of the TV's to a) set themselves to a particular time, b) set themselves to a particular site, or a particular task on that site, c) cause a certain filtering or sampling to be enabled. This is an important convenience for trying to do simultaneous comparisons between different times of day, different parts of the site, or even different language versions of the same site.

In certain embodiment, busy buildings - buildings that cuπently contain a higher than one standard deviation from average amount of traffic at the cuπent time will appear with their windows lit up in the campus view. In order to achieve this, each building must have two images associated with it: one with the windows lit, and one with the windows dark. Further gradations are possible, but require more images.

In certain embodiments, buildings also have parking lots outside them in the campus view. The size of the lot indicates the maximum amount of concuπent customer traffic that is common for that task. The cuπently parked number of cars indicates what fraction of that peak traffic is presently there. If the traffic is higher than expected, individual cars can be replaced with buses, using the same approach as is used in the campus view to indicate heavy path traffic. The described Cam can be used in traditional usability tests. Specifically, while a usability test is being conducted, all customer events associated with the usability test could be directed to a special webserver, from which a Cam tape could be made.

In certain embodiments, in order to facilitate tracking a particular customer throughout his stay on the site, a spotlighting capability is provided. Specifically, this feature causes the selected character to appear under a brightly colored spotlight that travels with him until he leaves the site, or until the Cam user turns the spotlighting off. This can be enabled from the Customer Properties panel. If more than one customer is being spotlighted, different color spotlights will be used. Certain embodiments include a "morgue" for characters that have died The morgue is one way of visually presenting information about customers who have already left the site, and did so while in one of the disgruntled states This is one particular way of slicing the available information that is of great interest to marketers, since it can assist them m tailoring their email promotions For example, an email might be sent to all repeat customers who left the site disgruntled dunng their previous trip, asking them if they were able to find what they were looking for, telling them how much they are valued, and perhaps offeπng a discount to prove it

There are at least two vaπations on the visual theme ghosts and a traditional morgue A ghost morgue is similar to the display of a regular room, except that all of the characters m are represented ensconced in a pale white film, and are also partially transparent The characters can be filtered and sampled as in a normal room, with the additional filtenng property of "time " (This is because a lot of dead bodies can stack up, especially duπng a long tape ) Thus, the ghost morgue can be filtered to show only characters that died during a period of time specified by the Cam user

In the traditional morgue, the Cam user must specify a set of attributes for the type(s) of bodies they want to inspect If the Cam user specifies several attributes, and no matches are found as a result, matches that have at least some of the specified attributes will be sought

Once they do this, a list of dead customers who meet the Cam user's criteria will be brought up The list offers the following information a) the name of the customer (if available), b) in what task and step the customer died, c) in what mood state he died, d) which tasks he attempted during the session and did complete successfully, e) which tasks he attempted but did not complete successfully, and f) prior purchasing summary (if available ) By selecting a customer in this list, the Cam user has the option of spotlighting and tracking them in the tape

When the tape is over, snow will fill the screen There is however a loop mode that allows the same tape to be shown over and over again

In the real-time system, a character selecting cursor is available to make it easier to grab onto a particular customer in a crowded room In this mode, a bold selection rectangle will be drawn around the customer that is cuπently selected This customer will remain selected even if he moves, until the user moves her mouse away from him

It should be understood that certain embodiments of the present invention display only characters representing mood states or only characters representing demographic information Not all embodiments of the present invention display characters representing both types of data Moreover, certain other embodiments may display characters representing other useful web-site interaction data besides the mood state and demographic data discussed herein without departtng from the spint and scope of the present invention

From the above descriptions, it will be apparent that the present invention disclosed herein provides a novel and advantageous method and system for assessing states of customers of web-sites The foregoing discussion discloses and descnbes merely exemplary methods and embodiments of the present invention As will be understood by those familiar with the art, the invention may be embodied in other specific forms without departing from the spirit or essential characteπsttcs thereof Accordingly, the disclosure of the present invention is intended to be illustrative, but not limiting, of the scope of the invention, which is set forth m the following claims and equivalents

Table 1

Example of Generated Report (note that "customers" are called ' users' in this report)

Report for site ABCcompany com

For Wed, Sep 13, 2000 03:23 PM through Wed, Sep 13, 2000 04:25 PM

[]User visits 458 [] Unique user visits 458

++++ For mood Angry

Users in this mood 24 Users completing a task 0 Percentage completing 0 0%

Users falling out of a task 17

Percentage falling out 70 83% Users leaving site from this mood 9 Percentage leaving 37 5% Average user life span 01 17

Percentage in demographic Com 0 0% Percentage in demographic Edu 0 0% Percentage in demographic Unknown / Non-US 100 0% Percentage in demographic Gov 0 0% Percentage demographic Net 0 0%

Percentage in demographic Org 0 0% For mood Bored

Users in this mood 451 Users completing a task 32 Percentage completing 7 1%

Users falling out of a task 402 Percentage falling out 89 14% Users leaving site from this mood 116 Percentage leaving 25 72% Average user life span 01 10

Percentage in demographic Com 0 67% Percentage in demographic Edu 0 22% Percentage in demographic Unknown / Non-US 99 11 % Percentage in demographic Gov 0 0% Percentage in demographic Net 0 0%

Percentage in demographic Org 0 0% For mood Confused

Users in this mood 41 Users completing a task 0 Percentage completing 0 0% Users falling out of a task 34

Percentage falling out 82 93% Users leaving site from this mood 5

Percentage leaving 12 2% Average user life span 01 13 Percentage in demographic Com 0 0%

Percentage in demographic Edu 0 0% Percentage in demographic Unknown / Non-US 100 0% Percentage in demographic Gov 00% Percentage in demographic Net 0 0% Percentage in demographic Org 0 0%

For mood Satisfied

Users in this mood 2 Users completing a task 0 Percentage completing 0 0% Users falling out of a task 0

Percentage falling out 0 0% Users leaving site from this mood 2 Percentage leaving 100 0% User life spans insufficient 5 Average user life span 00 00

Percentage in demographic Com 0 0% Percentage in demographic Edu 0 0% Percentage in demographic Unknown / Non-US 100 0% Percentage in demographic Gov 0 0% Percentage in demographic Net 0 0%

Percentage in demographic Org 00% For mood Purchased

Users in this mood 0 For mood Shopping Users in this mood 456

Users completing a task 36

Percentage completing 7 89% Users falling out of a task 392 Percentage falling out 85 96% Users leaving site from this mood 197

Percentage leaving 43 2% Average user life span 01 10 Percentage in demographic Com 0 66% Percentage in demographic Edu 022% Percentage in demographic Unknown / Non-US 99 12%

Percentage m demographic Gov 0 0% Percentage in demographic Net 0 0% Percentage in demographic Org 0 0% ++++ For demographic Com

Users in this demographic 3 Users completing a task 0 Percentage completing 0 0% Users falling out of a task 3 Percentage falling out 100 0%

Average user life span 01 16 Percentage in mood Angry 0 0% Percentage in mood Bored 100 0% Percentage in mood Confused 0 0% Percentage in mood Satisfied 0 0%

Percentage in mood Purchased 0 0% Percentage in mood Shopping 100 0% For demographic Edu

Users in this demographic 1 Users completing a task 0 Percentage completing 0 0%

Users falling out of a task 1 Percentage falling out 1000% Average user life span 01 1 1 Percentage in mood Angry 0 0% Percentage in mood Bored 100 0%

Percentage in mood Confused 0 0% Percentage in mood Satisfied 0 0% Percentage in mood Purchased 0 0% Percentage in mood Shopping 100 0% For demographic Unknown / Non-US

Users in this demographic 454 Users completing a task 44

Percentage completing 9 69% Users falling out of a task 454 Percentage falling out 100 0%

Average user life span 01 09 Percentage in mood Angry 5 29% Percentage in mood Bored 98 46% Percentage in mood Confused 9 03% Percentage in mood Satisfied 0 44%

Percentage in mood Purchased 0 0% Percentage in mood Shopping 99 56% For demographic Gov

Users in this demographic 0 For demographic Net

Users in this demographic 0 For demographic Org

Users in this demographic 0 ++++ []Profihng task Maintain Profile

User visits 392 Users that fell out of task 359

Percentage that fell out of task 91 58% Users that completed task 26 Percentage that completed task 6 63%

[JProfiling step Input Profile Users in task that av oided this room 59 Users in task that entered this room 333 Users that fell out of task from here 305 Percentage of users in this step that fell out from here 91 59%

Percentage of users in this task that fell out from here 77 81% []Profilιng step Update Profile Users in task that avoided this room 173 Users in task that entered this room 219 Users that fell out of task from here 165

Percentage of users in this step that fell out from here 75 34% Percentage of users in this task that fell out from here 42 09% +++++++ For mood Angry Users m this mood 14

Users completing a task 0

Percentage completing 0 0% Users falling out of a task 10 Percentage falling out 71 43% Users leaving site from this mood 3

Percentage leaving 21 43% Average user life span 01 18 Percentage in demographic Com 0 0% Percentage in demographic Edu 0 0% Percentage in demographic Unknown / Non-US 100 0% Percentage in demographic Gov 0 0%

Percentage in demographic Net 0 0% Percentage in demographic Org 0 0% For mood Bored

Users in this mood 337 Users completing a task 18

Percentage completing 5 34% Users falling out of a task 310

Percentage falling out 91 99% Users leaving site from this mood 12 Percentage leaving 3 56%

Average user life span 01 1 1 Percentage in demographic Com 0 59% Percentage in demographic Edu 0 0% Percentage in demographic Unknown / Non-US 99 41% Percentage in demographic Gov 0 0%

Percentage in demographic Net 0 0% Percentage in demographic Org 0 0% For mood Confused

Users in this mood 21 Users completing a task 0

Percentage completing 0 0% Users falling out of a task 18

Percentage falling out 85 71 % Users leaving site from this mood 1 Percentage leaving 4 76%

Average user life span 01 14 Percentage in demographic Com 0 0% Percentage in demographic Edu 0 0% Percentage in demographic Unknown / Non-US 100 0% Percentage in demographic Gov 0 0%

Percentage in demographic Net 0 0% Percentage in demographic Org 0 0% For mood Satisfied

Users in this mood 0 For mood Purchased

Users in this mood 0 For mood Shopping

Users in this mood 32 Users completing a task 0 Percentage completing 0 0%

Users falling out of a task 0

Percentage falling out 0 0% Users leaving site from this mood 32 Percentage leaving 100 0% Average user life span 01 09

Percentage in demographic Com 0 0% Percentage in demographic Edu 0 0% Percentage in demographic Unknown / Non-US 100 0% Percentage in demographic Gov 0 0% Percentage in demographic Net 0 0%

Percentage in demographic Org 0 0% ++++ For demographic Com

Users in this demographic 3 Users completing a task 0

Percentage completing 0 0% Users falling out of a task 3

Percentage falling out 100 0% Average user life span 01 16 Percentage in mood Angry 0 0% Percentage in mood Bored 66 67%

Percentage in mood Confused 0 0% Percentage in mood Satisfied 0 0% Percentage in mood Purchased 0 0% Percentage in mood Shopping 0 0% For demographic Edu

Users in this demographic 0 For demographic Unknown / Non-US Users in this demographic 389 Users completing a task 26 Percentage completing 6 68%

Users falling out of a task 356

Percentage falling out 91 52% Average user life span 01 10 Percentage in mood Angry 3 6% Percentage m mood Bored 86 12%

Percentage in mood Confused 5 4% Percentage in mood Satisfied 0 0% Percentage in mood Purchased 0 0% Percentage in mood Shopping 8 23% For demographic Gov

Users in this demographic 0 For demographic Net

Users in this demographic 0 For demographic Org Users in this demographic 0

[]Profilιng task Pharmacy Quick-Find User visits 388 Users that fell out of task 362 Percentage that fell out of task 93 3%

Users that completed task 18 Percentage that completed task 4 64% []Profilιng step Enter Store Info Users in task that avoided this room 160 Users in task that entered this room 228

Users that fell out of task from here 202 Percentage of users in this step that fell out from here 88 6% Percentage of users in this task that fell out from here 52 06% []Profihng step Display Results Users in task that avoided this room 69

Users in task that entered this room 319 Users that fell out of task from here 280 Percentage of users in this step that fell out from here 87 77% Percentage of users in this task that fell out from here 72 16% +++++++

For mood Angry

Users in this mood 9 Users completing a task 0 Percentage completing 0 0% Users falling out of a task 9

Percentage falling out 100 0% Users leaving site from this mood 0

Percentage leaving 0 0% Average user life span 01 17 Percentage in demographic Com 0 0%

Percentage in demographic Edu 0 0% Percentage in demographic Unknown / Non-US 100 0% Percentage in demographic Gov 0 0% Percentage in demographic Net 0 0% Percentage in demographic Org 0 0% od Bored

Users in this mood 338 Users completing a task 13

Percentage completing 3 85% Users falling out of a task 316 Percentage falling out 93 49%

Users leaving site from this mood 8

Percentage leaving 2 37% Average user life span 01 11 Percentage in demographic Com 0 59% Percentage in demographic Edu 0 3%

Percentage in demographic Unknown / Non-US 99 11 % Percentage in demographic Gov 0 0% Percentage in demographic Net 0 0% Percentage in demographic Org 0 0% od Confused

Users in this mood 19 Users completing a task 0 Percentage completing 0 0% Users falling out of a task 14 Percentage falling out 73 68%

Users leaving site from this mood 1

Percentage leaving 5 26% Average user life span 01 12 Percentage in demographic Com 0 0% Percentage in demographic Edu 0 0%

Percentage in demographic Unknown / Non-US 100 0% Percentage m demographic Gov 0 0% Percentage m demographic Net 0 0% Percentage in demographic Org 0 0% d Satisfied

Users in this mood 0 d Purchased

Users in this mood 0 d Shopping Users in this mood 12

Users completing a task 0 Percentage completing 0 0% Users falling out of a task 0 Percentage falling out 0 0% Users leaving site from this mood 12

Percentage leaving 100 0% Average user life span 01 12 Percentage in demographic Com 0 0% Percentage in demographic Edu 8 33% Percentage in demographic Unknown / Non-US 91 67%

Percentage in demographic Gov 0 0% Percentage in demographic Net 0 0% Percentage in demographic Org 0 0% ographic Com

Users in this demographic 2 Users completing a task 0

Percentage completing 0 0% Users falling out of a task 2 Percentage falling out 100 0%

Average user life span 01 19 Percentage in mood Angry: 0.0% Percentage mood Bored- 100 0% Percentage in mood Confused: 0.0% Percentage in mood Satisfied- 0 0% Percentage in mood Purchased. 0 0%

Percentage in mood Shopping 0.0% For demographic .Edu.

Users in this demographic: 1 Users completing a task: 0 Percentage completing: 0.0%

Users falling out of a task: 1 Percentage falling out- 100.0% Average user life span: 01: 11 Percentage in mood Angry: 0.0% Percentage in mood Bored: 100.0%

Percentage in mood Confused: 0.0% Percentage in mood SaUsfied. 0.0% Percentage in mood Purchased: 0.0% Percentage in mood Shopping. 100.0% For demographic Unknown / Non-US.

Users in this demographic: 385 Users completing a task: 18

Percentage completing: 4.68% Users falling out of a task: 359 Percentage falling out 93.25%

Average user life span- 01: 10 Percentage in mood Angry: 2.34% Percentage in mood Bored: 87.01% Percentage in mood Confused. 4.94% Percentage in mood Satisfied: 0.0%

Percentage in mood Purchased: 0.0% Percentage in mood Shopping: 2.86% For demographic .Gov

Users in this demographic: 0 For demographic .Net

Users m this demographic: 0 For demographic .Org

Users in this demographic: 0 ++++++ []Profιlιng task: Shop Valu-Rite

User visits: 458 Users that fell out of task 457 Percentage that fell out of task- 99.78% Users that completed task- 0 Percentage that completed task 0.0%

[]Profihng step Catalog Categoπes Users in task that avoided this room: 109 Users in task that entered this room: 349 Users that fell out of task from here: 175 Percentage of users in this step that fell out from here 50 14%

Percentage of users in this task that fell out from here 38 21 % []Profilιng step Catalog Details Users in task that avoided this room: 117 Users in task that entered this room: 341 Users that fell out of task from here: 188

Percentage of users in this step that fell out from here 55 13% Percentage of users in this task that fell out from here 41 05% [JProfiling step Shopping Basket Users in task that avoided this room 126 Users in task that entered this room: 332

Users that fell out of task from here: 179 Percentage of users in this step that fell out from here 53.92% Percentage of users in this task that fell out from here 39.08% []Profilιng step Login, Payment & Delivery Options Users in task that avoided this room: 134 Users in task that entered this room- 324

Users that fell out of task from here: 159 Percentage of users in this step that fell out from here 49.07% Percentage of users in this task that fell out from here 34.72% []Profιlιng step Payment Details & Personal Info Users in task that avoided this room: 135

Users in task that entered this room: 323 Users that fell out of task from here: 169 Percentage of users in this step that fell out from here- 52 32% Percentage of users in this task that fell out from here- 36.9% []Profihng step Veπfy

Users in task that avoided this room: 124 Users in task that entered this room: 334 Users that fell out of task from here: 170 Percentage of users in this step that fell out from here: 50.9% Percentage of users in this task that fell out from here. 37.12%

[]Profϊhng step Transaction Complete Users in task that avoided this room: 121 Users m task that entered this room: 337 Users that fell out of task from here: 139 Percentage of users in this step that fell out from here. 41.25%

Percentage of users in this task that fell out from here: 30.35% +++++++ For mood Angry

Users in this mood 16 Users completing a task 0

Percentage completing 0.0% Users falling out of a task. 6

Percentage falling out 37 5% Users leaving site from this mood: 6 Percentage leaving 37 5%

Average user life span 01 - 17 Percentage in demographic .Com: 0.0% Percentage in demographic .Edu: 0.0% Percentage in demographic Unknown / Non-US 100.0% Percentage in demographic .Gov: 0.0%

Percentage in demographic Net: 0.0% Percentage in demographic .Org: 0.0% For mood Bored

Users in this mood 402 Users completing a task 0

Percentage completing 0.0% Users falling out of a task: 192

Percentage falling out 47.76% Users leaving site from this mood: 96 Percentage leaving 23 88%

Average user li fe span 01 : 11 Percentage in demographic .Com: 0.75% Percentage in demographic .Edu: 0.25% Percentage in demographic Unknown / Non-US- 99.0% Percentage in demographic .Gov: 0.0%

Percentage in demographic .Net: 0.0% Percentage in demographic .Org- 0.0% For mood Confused

Users m this mood 5 Users completing a task 0

Percentage completing 0.0% Users falhng out of a task 2

Percentage falling out 40 0% Users leaving site from this mood 3 Percentage leaving 60 0% Average user life span 01 06

Percentage in demographic Com 0 0% Percentage in demographic Edu 0 0% Percentage in demographic Unknown / Non-US 100 0% Percentage in demographic Gov 0 0% Percentage in demographic Net 0 0%

Percentage in demographic Org 0 0% For mood Satisfied

Users in this mood 2 Users completing a task 0 Percentage completing 0 0%

Users falling out of a task 0 Percentage falling out 0 0% Users leaving site from this mood 2 Percentage leaving 100 0% User life spans insufficient 3

Average user life span 00 00 Percentage in demographic Com 0 0% Percentage in demographic Edu 0 0% Percentage in demographic Unknown / Non-US 100 0% Percentage in demographic Gov 0 0%

Percentage in demographic Net 0 0% Percentage in demographic Org 0 0% For mood Purchased

Users in this mood 0 For mood Shopping

Users m this mood 456 Users completing a task 0

Percentage completing 0 0% Users falling out of a task 409 Percentage falling out 89 69%

Users leaving site from this mood 153

Percentage leaving 33 55% Average user life span 01 10 Percentage in demographic Com 0 66% Percentage in demographic Edu 0 22%

Percentage in demographic Unknown / Non-US 99 12% Percentage in demographic Gov 0 0% Percentage in demographic Net 00% Percentage in demographic Org 0 0% ++++

For demographic Com

Users in this demographic 3 Users completing a task 0 Percentage completing 0 0% Users falling out of a task 3

Percentage falling out 100 0% Average user life span 01 16 Percentage in mood Angry 0 0% Percentage in mood Bored 1000% Percentage in mood Confused 0 0%

Percentage in mood Satisfied 0 0% Percentage mood Purchased 0 0% Percentage in mood Shopping 100 0% For demographic Edu Users m this demographic 1

Users completing a task 0 Percentage completing 0 0% Users falling out of a task 1

Percentage falling out 100 0% Average user life span 01 1 1 Percentage in mood Angry 0 0%

Percentage in mood Bored 100 0% Percentage m mood Confused 0 0% Percentage in mood Satisfied 0 0% Percentage in mood Purchased 0 0% Percentage in mood Shopping 100 0%

For demographic Unknown / Non-US Users in this demographic 454 Users completing a task 0 Percentage completing 0 0% Users falling out of a task 453

Percentage falling out 99 78% Average user life span 01 09 Percentage in mood Angry 3 52% Percentage in mood Bored 87 67% Percentage m mood Confused 1 1%

Percentage in mood Satisfied 0 44% Percentage in mood Purchased 0 0% Percentage in mood Shopping 99 56% For demographic Gov Users in this demographic 0

For demographic Net

Users m this demographic 0 For demographic Org

Users in this demographic 0 ++++++

Claims

We claim

1 A method of displaying customer states connected with a website, the method compπsmg displaying a graphical representations of at least two tasks performed by customers on the website, and displaying graphical representations of customers moving between the graphical representations of the tasks, the display reflecting movement of the customers between the tasks of the website

2 The method of claim 1, further comprising displaying a graphical representation having one or more areas, each area in the representation coπesponding to a respective step of a task, and displaying, in at least one area coπesponding to one of the task steps, a graphical representation of at least one customer performing that task step

3 The method of claim 1 , wherein displaying a graphical representation of at least one customer includes displaying a moving graphical representation of at least one customer

4 The method of claim 1, wherein the graphical representation of at least one customer includes graphically representing demographic information for the customer

5 The method of claim 4, wherein the demographic information includes a domain of the customer

6 The method of claim 1, wherein displaying a graphical representation of at least one customer includes graphically representing mood information for the customer

7 The method of claim 6, wherein the mood information represents a mood state of the customer, each mood state being represented by an action performed by a moving graphical figure

8 A method of displaying customer states of a website the system compnsmg displaying a graphical representation having one or more areas, each area in the representation coπesponding to a respective step of a task of the website, and displaying, in at least one area coπesponding to one of the task steps, a graphical representation of at least one customer performing that task step

9 The method of claim 8, wherein displaying a graphical representation of task steps includes displaying the task as a floorplan having one or more rooms, each room coπesponding to a task step

10 The method of claim 8, wherein displaying a graphical representation of at least one customer performing steps of the task includes displaying a moving graphical representation of the customers

11 The method of claim 8, wherein displaying a graphical representation of at least one customer performing each task step includes graphically representing demographic information for each customer

12 The method of claim 11, wherein the demographic information includes a domain of the customer

13 The method of claim 8, wherein displaying a graphical representation of at least one customer includes graphically representing mood information for the customer

14 The method of claim 13, wherein the mood information represents a mood state of the customer, each mood state being represented by an action performed by a moving graphical figure

15 The method of claim 8, wherein the speed at which the representation of customers move is faster than the customers completed the task steps in real realtime

16 The method of claim 8, wherein the display operates in a realtime mode

17 The method of claim 8, wherein the display operates in a retroactive mode

18 The method of claim 8, wherein a representation of a customer wandenng from a path m a room indicates the time that the customer has spent on the coπesponding task

19 The method of claim 8, wherein the areas include static images indicating the function of the task step coπesponding to the area.

20. A system that displays customer states of a website, the system comprising- a software portion configured to display a graphical representations of at least two tasks performed by customers on the website; and a software portion configured to display graphical representations of customers moving between the graphical representations of the tasks, the display reflecting movement of the customers between the tasks of the website.