IE84815B1 - A survey management system - Google Patents

Abstract

ABSTRACT A survey management system comprises a data processor and a user interface. It generates a plurality of questions in a database, each question including text for a survey question and at least one associated potential answer. The system dynamically generates a machine-readable identifier for each answer in the dataset, the identifier being generated according to both the answer and the campaign header data. The system captures survey campaign results by machine reading the identifier of each answer selected by a respondent, automatically analysing each identifier, automatically determining the associated campaign for the identifiers, and automatically writing data representing the selected answers to the associated dataset. The data processor instantiates an object for each question, for each answer, and for each header and manages associations of said objects to populate the campaign dataset. The objects are managed in a first level of a hierarchical object structure, a second level of the structure comprises objects each for a question and answer combination, a third level of objects containing a plurality of second level objects, and a fourth level comprising campaign objects. The data processor dynamically generates a machine readable identifier for each combination of answer object and fourth level campaign object.

Description

A Survey Management System Introduction The invention relates to management of surveys at the technical level.

It is well known that surveys can be of enormous benefit to a business. for example, for providing customer feedback for assisting with changing a product or service.

However. surveys are generally under-utilised because they are costly, they take a considerable amount of time to develop. and many businesses believe they have insufficient in-house skills to manage them.

In one example. users are provided with handheld computers to use in the field for capturing responses. This approach is expensive and requires excessive user training to be practical for many organisations. Another existing method used is for interviewers to use paper questionnaires to collect the data and then have operators manually enter the resulting data into the computers. This is slow and prone to errors. for instance from inis-keying the data.

US4954699 describes a self-administered survey.

The invention is therefore directed towards providing a more automated survey management system, in which automatic functions reduce the extent of human input required. Another objective is to provide such automation so that data integrity is improved and/or human time input is reduced.

Statements of Invention According to the invention, there is provided a survey management system comprising a data processor, and a user interface having a reader device, wherein the processor is adapted to: generate a plurality of questions in a database, each question including text for a survey question and at least one associated potential answer, generate a dataset for a survey campaign; populate the datasct with questions and with campaign header data, dynamically generate a machine-readable identifier for each answer in the dataset, the identifier being generated according to both the answer and the campaign header data, direct printing a questionnaire including text for a plurality of questions, text for the associated answers, and machine—readable identifiers alongside the answer text, capture survey campaign results from the reader device when machine reading the printed identifier of each answer selected by a respondent; automatically analysing each identifier, automatically determining the associated campaign for the identifiers, and automatically writing data representing the selected answers to the associated dataset, and analysing the selected answers to generate a report; wherein the data processor associates a component of a machine-readable identifier with rules, and automatically executes said rules upon machine reading of said identifier.

In one embodiment, the data processor is adapted to assign each question to a classification, and to automatically select questions for a campaign survey by receiving criteria from a user and automatically retrieving question records which satisfy said criteria.

In one embodiment, the data processor is adapted to populate a campaign dataset with question records by receiving identifiers from the reader device, determining the question and answer text from the database and writing said text to the dataset.

In one embodiment. the reader device is a bar code scanner.

In one embodiment, the data processor instantiates an object for each question. for each answer, and for each header and manages associations of said objects to populate the campaign dataset.

In one embodiment, said objects are managed in a first level of a hierarchical object structure.

In one embodiment, a second level of the structure comprises objects each for a question and answer combination.

In one embodiment. a third level of the structure comprises survey objects containing a plurality of second level objects.

In one embodiment, a fourth level of the structure comprises Campaign objects containing a first level campaign header object, at least one third level survey object, and an identifier for each answer.

In another embodiment, the data processor comprises means for dynamically generating a machine readable identifier for each combination of answer object and fourth level campaign object.

In one embodiment, the data processor comprises means for individual modification of objects. and the system propagates modifications to associated objects. ln a further embodiment, the data processor comprises means for managing a plurality of databases, and for including a database identifier into each machine readable identifier.

In one embodiment, said rules include rules for determining how other responses are to be processed such that processing or responses in a questionnaire is dynamically variable.

Detailed Description of the Invention The invention will be more clearly understood from the following description of some embodiments thereof, given by way of example only with reference to the accompanying drawings in which:— Figs. 1(a), l(b), and 1(0) are diagrams illustrating technical architectures of systems of the invention; and Figs. 2 to 6 are sample screen shots for operation of a system; Fig. 7 is a sample survey questionnaire generated by the system; Fig. 8 is a sample analysis output; Fig. 9 is a further sample survey questionnaire; and F igs.l0 and 1] are sample screen displays for managing surveys.

A survey management system of the invention comprises a computer having a survey question database. A record for each question has the question and also a set of possible multiple-choice answers. Also, each question record includes a classification attribute. allocating the question to a category. The system further comprises a client database.

A campaign may be generated by the system selecting a set of question records and saving them to a dataset for the campaign. The campaign dataset also includes client data. As the campaign dataset is populated with questions the system automatically generates a machine readable code for each answer and campaign pairing. Thus an answer may have one code when it is used in one campaign and another when used in a different campaign. Linkage of the machine readable code with objects in the database is described in more detail below.

The question records may be selected automatically by the system according to classification. Alternatively. the question records may be retrieved under user instructions from the database. Alternatively, questions may be selected by scanning bar codes of associated answers from a hard copy of a previous questionnaire.

For capturing answers. it is only necessary to scan the multiple choice answer boxes and the barcode for each question. The technical process of scanning a selected answer involves the system writing the response in a fixed, pre-set. path to the relevant database record. The fact that the bar code is uniquely associated with an answer and campaign pairing allows such direct routing.

In more detail. the system instantiates objects and the objects combine to form more complex objects on a tiered basis. The tiered levels are as follows: Level I: An object for each question. answer, client. user/control instructions (for example. “Read this specific line". “If answer to previous questions was no then skip to questions X”) Level 2: question and answer combination objects Level 3: survey consisting of multiple level 2 objects.

Level 4: Campaign consisting of a level 3 survey object + level 1 client object + a unique identifier object code.

Each question exists separately in level 1 as do the answers. An answer can be re—used with multiple questions. A single question can be combined with multiple answers to form a unique level 2 object.

A level 3 object is a survey and consists of multiple level 2 objects plus one or more level 1 user/control objects arranged in a specific order.

A level 4 object is a final campaign. This is a unique level 3 survey object combined with a level 1 client object and a unique identifying code. A user can run multiple different campaigns from the same survey, for instance for different products.

This design structure allows the system to quickly create active campaigns. each having any number or combination of the lower level objects. A graphical user interface (such as drag-and-drop) allows the user to instruct the system to combine these elements to form a level 4 campaign object. This can then be printed and provided to interviewers.

Each question can be designated to allow multiple answers or only one answer. for example, "do you buy any of the following papers” vs. “Gender”. This rule can be implemented in any or all levels. There are advantages in such a rule being implemented in each level. For instance, tying it to level 1 ensures that all uses of that question in higher levels will be so restricted regardless of the answers associated with it. doing so at level 2 restricts the usage of that Q+A object in all surveys. while doing so at level 3 or 4 limits it only within the Survey or Campaign. This rule can be included in the barcode itself.

The barcode is generated by the system at level 4 (the campaign). It is not necessary to store it within the database as an individual entity. It contains all of the elements necessary to disassemble the scanned answer back to its constituent base objects.

Thus. it identifies the relevant level l-level 4 objects. By scanning a specific barcode a number of actions are generated in the system: The code is parsed and checked for errors. This uses a checksum plus specific characters placed strategically within the code.

If an error has occurred then an alert is generated to prompt a re-scan.

Disassemble the code into its constituent pieces.

Check for this being a duplicate scan.

Check if this is a second answer to the same question. If so. check if that is allowed in the rL1les (the rule may be stored within the database or within the barcode itself).

Store the constituent objects together in a result table ready for analysis.

The barcode also allows for an “Other” option. This is a specially identified answer (level 1 object) which does not in fact contain an answer of its own. Instead. the system when it sees this answer allows the user to enter text. This provides the flexibility to combine multiple-choice and open-ended questions. It also allows multiple choice answers but allows capture of options which might not have been foreseen by the questionnaire designer.

The barcode may also contain a unique code which can identify an individual database as outlined below.

A typical barcode may be represented in Latin alphabet as follows: lllclq2alarsdbthle 1 is a scan check to ensure that a correct barcode is in fact being scanned.

C 1 is campaign 1 Q2 is question 2 of that campaign Al is answer 1 for that question ars is the object code signifying that the question in this campaign only allows a single answer and so second or later answers should be ignored. dbthl identifies database thl (see the multiple database section below) e signifies the end of the scan Not all of the above elements are strictly needed but it is important to include specific characters that we can check are where we expect them to be since bareodes can sometimes mis-scan.

Because of the underlying technical functionality of the invention, it can be implemented in any desired computing architecture, for example with either a stand- alone computer or a client/server arrangement as illustrated in Fig. 1(a).

The functionality is most commonly divided in two with the front—end containing the user interface and coding logic while the back-end is a database containing the level 1 to level 4 objects along with results from previous campaigns. The number of elements into which it can be broken can be reduced or increased.

Referring to Fig. l(b) the hardware architecture may comprise a local area network or a single machine, in which the following interactions take place: Step 1, front-end is initiated and looks for a pre-specified file. It then reads the file and extracts from it the path to the required back-end database. It connects to the database and may display the path and the campaign objects within the database.

Step 2, it then connects to the identified database and shows the path to the database and all campaigns contained inside it Referring to Fig. 1(c), there may be connections to multiple local network databases.

In this implementation each database contains its own objects and surveys. There may for example be individual databases for each client or business category. In this case when the results of a survey are to be scanned the front—end can connect to the relevant back-end and the barcodes scanned as described above. This does however leave open the possibility of error if the wrong back-end database is used. The invention solves this problem by incorporating a unique database identifier into each barcode along with the previously mentioned data. This works for example as follows: — A new database is set up for use with a new client. This can be a copy of an existing template database containing objects useful when working with a specific industry.

— The operational software is initiated and connected to the specified database.

The first thing it does is to check if a unique DatabaselD has been assigned. If so. it proceeds to the next step. If not, it generates a unique identifier and writes it into the allocated area of the database. This identifier will consist of a number of characters. If we take just the 26 letters of the alphabet and ten numerical digits then a 3-character identifier allows 46656 unique databases while 5 characters provides for more than 6 million databases.

— The unique identifier then becomes in effect another object and is incorporated into the barcode generated for each answer in level 4.

— When an answer is scanned the DatabaseID is also parsed and if it does not match that of the currently loaded database an error message is displayed.

— A variation of this approach allows the DatabaseID to automatically and dynamically load the correct database before inputting the results to the results table. This can be managed by giving the database the name matching its ID or by using a lookup table in a central database.

In this way the system can use multiple databases, allowing for example an individual database for each client. It also allows the work of scanning data into a database to be safely contracted to different parties without having to provide access to the companies’ servers. In this case a person hired to scan the data can be emailed the relevant database and conduct the scanning for instance at home. If the wrong database is emailed or the scanning party mixes it with another then the company can be sure that there is no mixing of data or other similar error.

It is also possible to connect the front and back-ends over a distance using a method such as the internet. In such a case a client program can run on a remote PC. This will contain many objects at different levels. for instance it can have many level 1 questions and answers as well as level 2 or 3 objects. It will also have a unique clientlD.

A server can contain level 3 and level 4 objects along with many different clientlDs.

On connecting to the server the client PC can download the objects which are linked to its own clientID. In this way a central server can supply new questionnaires and surveys to a specified client in a bandwidth-efficient manner since it only downloads the instructions on how to combine the existing objects. Similarly this is a secure way to transfer the data since only part of it is being transmitted at one time and. if intercepted. has no value on its own. Lower-level objects can be downloaded in the same ITl3.I’]l’l€I‘.

If desired. the client can also return completed data to the server when completed.

This will be tagged in the central server with the relevant clientlD. This clientID will ensure that the data is not mixed with those from other campaigns. The central office can then conduct analysis as required on data just from that client or from multiple clients running the same campaign. This is useful for instance for a large company having multiple outlets and who want to get both national and local results.

Referring to Fig. 1(a) the system may comprise a server having a master program for generating the database, and being accessible by clients to: — generate a survey by, on-line, instructing automatic selection of questions for a campaign, and downloading the survey with control parameters and the campaign client details, and — subsequently receive an upload of the survey results, the user having a bar- code reader for scanning the results selected.

It is possible to provide the end user with databases containing several objects with which they can interact. Because the user interface is separated from the technical aspects of the objects and barcodes the method of how the user interacts with those objects can be varied. It also makes it possible for the end user to create or modify objects. For instance. if a generic package was sold it could contain industry—specif1c questionnaires with some “dummy” questions and answers that could be modified by the end user to suit their own business.

An example of how such a system would operate and how a user interface could look is as follows.

Referring to Fig. 2. on initiating. the system can load an internal database or connect to and load data from a database specified in an external file. The display can identify to the user which database is currently selected and the information existing within that database.

Referring to Fig. 3, a user creates a question in the database by entering the question as it will appear on the survey. selecting a classification group if required. and writing in answers as they are to appear on the survey or select existing answers. If written in. they become answer objects and the system also automatically links them to the new level 2 question and answer object. The user writes a shorter (approximately l5 character) version to appear on charts. This allows the space on a graph to be used to exhibit the relevant data and not to be taken up with long labels. If an answer is to be left open for the person to enter their own information (e.g. an “Other" option) select this and write the associated text as it will appear on the survey. in this case “Other”.

Referring to Fig. 4, the user interacts with the system to create a campaign by entering a title of the survey, selecting a classification group to assign the survey to. and selecting a previously—created question from a drop-down menu.

Referring to Fig. 5, depending on the format of the survey the user may desire to enter instructions at various points, for example “If answer to question X is No then skip to question Y". The survey generator accommodates this. The user selects the survey. selects the question prior to which the instruction should appear. and enters the instruction at this point. For example, text typed at point 3 will appear between questions 2 and 3 on the printed survey.

Referring to Fig. 6. to create a campaign, the user enters the campaign title. assigns the campaign to a group, and selects the survey to be used from the previously—saved versions.

To run a campaign, the user simply selects the campaign from a drop down menu. and views and prints the survey.

It will be seen from Fig. 7 that the survey when printed has a set of questions, and for each question there is a set of potential answers. There is a printed bar code for each potential answer. This barcode meets the specifications described above.

Analysis of the results is according to algorithms selected for the campaign, and an example is illustrated in Fig. 8.

Referring to Fig. 9, another survey is illustrated, in this case for a radio station. This illustrates versatility of the system as there may be any desired number of potential answers to a question. This is at the discretion of the person generating the database of questions.

Referring to Fig. 10, for transferring the data in the completed questionnaire to the analysis system when the survey is completed the user or an automated system scans the barcodes which apply. Each campaign printed can also include a Start and End barcode. This is a generic code which can be incorporated into the first and last question barcodes or. as in Fig 10. printed separately. The first barcode to be scanned is the one that signifies the start of a new questionnaire. If a new questionnaire is scanned without firstly finishing the previous one then an error message can also be generated.

The selected answer to each question is then scanned. These can be scanned out-of- ordcr as the coding uniquely identifies the individual campaign. survey. question and answer. As specified previously individual questions can be configured so that only one answer is allowed or multiple answers are accepted. Open answers will display a further box in which to type the answer. As an aide to manual scanning, the scanned question and answer can be displayed on the screen along with buttons to delete previous entries or complete questionnaires. When all the answers on the questionnaire have been scanned a Finish Scanning code is scanned.

Analysing Data and Displaying Results At any point during the process the user may obtain a graphical report of results to date. With reference to Fig. 11, the user selects the campaign to be analysed and the dates for analysis along with the criteria for analysis such as by age, gender. residence. etc. The criteria available will depend on the questionnaire used. Results are displayed graphically for every question in multiple formats such as bar~charts or pie charts as shown in Fig. 8. There is automatic analysis by statistics such as gender or age.

Alternatively the results can be output in digital format for analysis by a dedicated graphing/statistical package.

User-Customisable Questions When creating a question for the question bank the author has the option of marking the question as customisable. The system then allows the end user to simply and quickly edit the question to allow them to for instance enter their own business name.

Alternatively. they can edit the responses to their own requirements such as a question that asks the person surveyed where they are from. This addition allows a generic package, with pre-designed questionnaires, to be sent to a number of different outlets without any customisation in advance.

It will be appreciated that the invention allows surveys to be created and carried out quickly and easily. This is accomplished by creating "banks" of objects from which surveys can be created. For instance, once a question and its associated answers are added to the bank they are always available for use by any later survey needing thcm.

Similarly. once a survey/questionnaire has been created then it can be selected again for different clients or campaigns.

Another aspect is that when the survey is printed for use it includes the question and answer in human-readable form but after each answer there is a barcode. This barcode contains all the information the system requires to process the result correctly. including unique identifiers and control codes for the campaign, the question, the survey. the answer, the client etc. In this way the usual data entry errors associated with the paper based interviews are eliminated.

These two elements together represent an object-oriented approach to the survey. The same question can be used in many surveys and similarly the same survey can be used in many campaigns but the barcode is comprised of elements, which automatically change to reflect the different usage. and so the system can input results from the different campaigns to different databases simultaneously without confusion. This allows the system to be modified such that it can be used as a stand-alone or networked platform. It can also be supplied with pre-created surveys or can download surveys as required. If downloading surveys, these can have unlimited usage or can have usage limited either by factors such as time or number of responses allowed.

The underlying technical processes of the invention allow use of the system in different business models. The following are some examples.

Associate Model A central agency (“Agency”) can divide an area into distinct regions. Each of the regions is assigned to franchised operators (“Associates”) who use the company name and expertise to sell market research to businesses in that area. Once a project has been agreed with a client the Associate devises a questionnaire and sends it to the Agency which holds and operates the system. The Agency encodes the questionnaire using the underlying technology and sends it back to the Associate. The Associate prints the questionnaire and conducts the interviews. When all interviews are completed the Associate returns the completed questionnaires to the Agency. The Agency scans the questionnaire data into the system and generates relevant graphical analysis. The Agency returns the completed reports to the Associate who in turn presents them to his client. The Associate therefore does not need to invest in software development. and the Associate gains the advantages of the system and the expertise of the Agency exclusively for their area. The Agency maintains control of the process and the survey engine.

A gencv using for its own in-house operation.

An Agency uses the system by designing questionnaires as described above and assigning campaigns using these to a client. The system prints hard copies and the survey is carried out. A scanning system scans the identified selected answers.

There is no need to separately select or identify the client, campaign, survey, question. or answer as these are all identifiable to the system from the information encoded into each answer barcode.

Advantages of this business model include reduced data entry time, reduced data entry error, quicker turn around from carrying out the survey to getting analysis. effective use of technology without having to invest in expensive technology in the field, and effective use of technology without having to invest in training for people in field.

Also. one can enter data from campaigns out of order since the encoded data in the barcode contains all relevant information Stand-alone version for businesses.

This involves providing to a business a discrete version of the software with some functionality disabled. This would not allow the business to create questions. answers or surveys. It would be supplied complete with questionnaires relevant to that industry, for example, a customer satisfaction questionnaire or a marketing analysis questionnaire. The business can then customise and print these and get customer feedback. The business then scans the data into the computer as before. Again, the barcode for any possible answer also has information encoded that allows the system to identify which questionnaire, survey, question, answer and campaign the data belongs to.

Advantages to small business include effective use of technology without having to invest in comprehensive technology, effective use of technology without having to invest in training for people in field, and ability to enter data from campaigns out of order since the encoded data in the barcode identifies all relevant information. Further. there is a quicker turn-around from carrying out the survey to getting analysis than using an agency, and the use of hardcopy surveys eliminates training requirements — employee simply has to ask question and tick relevant box.

The supplying agency can charge a modest fee and gain clients who would otherwise be unable to afford their services. hence providing a new revenue stream.

Distributing data over network (Fig. 1(a)) A server distributes a client program to a customer so that the customer system has the ability to provide survey requirements. The customer effectively buys individual surveys for conducting by themselves. When they outline the requirements the agency creates the survey and uploads it to the server. The client downloads the survey, prints it out, and carries out the survey themselves. They then scan in the answers as described above. The survey can have the option for the business to analyse the data themselves. Alternatively if the business has a number of branches each of which have downloaded the survey they can all upload the data back to the central server. The agency can then perform analysis on the results from each branch and present both branch and overall reports.

The client company therefore has the benefit of a professionally-created and tailored survey. and receives results back in a short time without having to invest in technology. Again, one can enter data from campaigns out of order since the encoded data in the barcode identifies all relevant information, and there is a quick turn-around from carrying out the survey to getting analysis.

In summary. it will be appreciated that the underlying technology of the invention provides banks of questions from which to generate surveys and the user just picks which of the previously created questions are required for this survey. The surveys can be used many different times in different campaigns, thereby allowing tracking of customer response over time. As the campaign, question, and answer data are all barcode encoded the system eliminates keying errors and reduce data entry time. Also. the system allows generation of simple multiple choice or open-entry questions. and it automatically prints a graphical analysis of the results.

The open-ended answers and the “Other” option are advantageous. This is a specially identified answer (level 1 object) which does not in fact contain an answer of its own.

Instead the system when it sees this answer opens a box on the screen for the user to enter text. This provides the flexibility to combine multiple-choice and open—ended questions. It also allows multiple choice answers but allows one to capture options which might not have been foreseen by the questionnaire designer.

This system allows the user to correct any errors using on-screen interface or using a barcode to for instance delete a previous answer scanned in error.

The system allows questions to be scanned in any order since the barcode identifies the relevant data.

The invention is not limited to the embodiments described but may be varied in construction and detail.

Claims (1)

1. Claims A survey management system comprising a data processor, and a user interface having a reader device. wherein the processor is adapted to: generate a plurality of questions in a database. each question including text for a survey question and at least one associated potential answer. generate a dataset for a survey campaign; populate the dataset with questions and with campaign header data, dynamically generate a machine-readable identifier for each answer in the dataset. the identifier being generated according to both the answer and the campaign header data, direct printing a questionnaire including text for a plurality of questions. text for the associated answers. and machine-readable identifiers alongside the answer text, capture survey campaign results from the reader device when machine reading the printed identifier of each answer selected by a respondent; automatically analysing each identifier, automatically determining the associated campaign for the identifiers, and automatically writing data representing the selected answers to the associated dataset, and analysing the selected answers to generate a report; wherein the data processor associates a component of a machine-readable identifier with rules, and automatically executes said rules upon machine reading of said identifier. A survey management system as claimed in claim 1, wherein the data processor is adapted to assign each question to a classification, and to automatically select questions for a campaign survey by receiving criteria from a user and automatically retrieving question records which satisfy said criteria. A survey management system as claimed in claims 1 or 2, wherein the data processor is adapted to populate a campaign dataset with question records by receiving identifiers from the reader device, determining the question and answer text from the database and writing said text to the dataset. A survey management system as claimed in any preceding claim, wherein the reader device is a bar code scanner. A survey management system as claimed in any preceding claim, wherein the data processor instantiates an object for each question, for each answer. and for each header and manages associations of said objects to populate the campaign dataset. A survey management system as claimed in claim 5, wherein said objects are managed in a first level of a hierarchical object structure. A survey management system as claimed in claim 6, wherein a second level of the structure comprises objects each for a question and answer combination. A survey management system as claimed in claim 7. wherein a third level of the structure comprises survey objects containing a plurality of second level objects. A survey management system as claimed in claim 8. wherein a fourth level of the structure comprises campaign objects containing a first level campaign header object, at least one third level survey object, and an identifier for each answer. A survey management system as claimed in claim 9, wherein the data processor comprises means for dynamically generating a machine readable identifier for each combination of answer object and fourth level campaign object. A survey management system as claimed in any of claims 5 to 10, wherein the data processor comprises means for individual modification of objects, and the system propagates modifications to associated objects. A survey management system as claimed in any preceding claim. wherein the data processor comprises means for managing a plurality of databases, and for including a database identifier into each machine readable identifier. A survey management system as claimed in any preceding claim. wherein said rules include rules for determining how other responses are to be processed such that processing of responses in a questionnaire is dynamically variable.