RETAIL CUSTOMER ANALYTIC SYSTEM
This invention relates to the collection of customer behaviour in retail outlets and an analytical system for use by retail management.
Background to the invention
Online retailing offers retailers the opportunity to easily monitor buyer behaviour in the online environment
Patents have been filed for monitoring and analysing online shopping customer behaviour. An example is USA patent 8898290 by Google.
In a physical store environment it is more difficult to assess the effectiveness of store displays and staff interaction with customers apart from measuring actual sales. One approach is to measure and track customers who carry a mobile phone.
In Australia it is estimated that 90% of shoppers carry a mobile phone. These phones are constantly checking automatically for a wifi network. It is possible to identify individual phones by their unique MAC address
USA patent application 20150026009 discloses a cloud based system utilising customer's mobile devices and an in store wifi.
USA patent application 20150025936 discloses a customer MAC address tracking and retail analysis system. At paragraph 113 a triangulation method is disclosed for tracking the location of a customer over time within the store.
One difficulty with these systems is that tracking a phone is not accurate in regards to tracking customers movements within a store because the signals may be coming from outside the store. Using instore video cameras is one means of verifying the information collected but use of cameras in this way may be regarded as breach of privacy laws.
It is an object of this invention to provide a more accurate method of tracking customers in a retail store environment.
Brief description of the invention
To this end the present invention provides a tracking system that utilises one or more sensors to collect movement by shoppers in the store in which each sensor includes a cell phone antenna, a WIFI antenna, a microcomputer and a passive infrared sensor in which the microcomputer is programmed to collect signals from shoppers cell phones via the WIFI antenna and to collect signals from the infrared
sensor and to correlate the two sets of signals to provide accurate movement information of shoppers in the store.
By combining the infra red and WIFI data the WIFI data is able to be verified as being within the store environment. Preferably more than one sensor more preferably 3 sensors are used so that by triangulation the movement of a cell phone can be tracked and used to create a heat map of the shoppers movements with in the store. By synchronising these movements with a map of the store layout information about the shoppers behaviour can be collected for analysis. Detailed description of the invention
A preferred embodiment of the invention will now be described with reference to the drawings in which
Figure 1 is an overview of the system;
Figure 2 is an example of an in store sensor;
Figure 3 is a flow chart illustrating the use of the sensor of figure 2;
Figure 4 is an example of a screen illustrating the analytics available .
With reference to figure 1 an individual store will preferably have 3 sensors which capture data from a cell phone which includes a MAC address and signal strength attenuation. The position of the smart phone at any instant is calculated using a triangulation estimation. The captured data is encrypted and sent to a server for processing.
The sensor shown in figure 2 includes a raspberry -PI microcomputer , a 3G dongle, a WIFI or WLAN antenna, a passive IR sensor and software.
The combination of a wifi antenna and a passive IR sensor enables the self correcting analytis of shopper motion in the store.
The WLAN antennas are used to determine the number of people in the store by analysing ping signal strengths and associated MAC addresses emitted by smart phones. This sensor alone may not differentiate between people in the store and outside the store because of variability in ping signal strengths. The PIR sensor determines the amount of human activity in the store and can be used to correct for errors in the WIFI signal analysis. The PIR sensors are an accurate measure of the number of people in the store. If the store is closed the PIR sensor can determine
that there are no shoppers in the store and discount the WiFi measurement accordingly.
As shown in figure 3 the information from the WiFi antenna and the PIR sensors are sent to the server and analysed to provide the shopper in store motion data for use in the analytic software.
The software on the server includes a map of the store layout and a calibration factor determined for each store. The server will also have point of sale data and other information such as a promotional calendar for the store.
The system is able to produce from this data four core metrics:
1. Store passers by
2. Visits into store
3. Time spent in store
4. Locations visited in store
These metrics are combined with point of sale data . The results may be displayed as heat maps of the visit. Accuracy is usually 1 to 2 metres.
By accumulating data statistical information can be derived and displayed. Figure
4 is an example of a screen displaying the analytics available.
On the database millions of data points that chart the relationship between the
PIR sensor output and the number of people in a store are accumulated. The database takes into account different types of retail environments and store area sizes. The relationship between the PIR output and Wi-Fi sensor output is different depending on the type of store, but we can classify the types of stores into several broad categories including:
• Small store (under 100 sqm), medium store (100 sqm - 500 sqm) or large store (over 1000 sqm) with low-shelves.
• Small store (under 100 sqm), medium store (100 sqm - 500 sqm) or large store (over 1000 sqm) with high-shelves.
• Outdoor stores.
• Temporary pop-up 'booth' stores
The system of this invention is able to detect when the output of the PIR sensor and Wi-Fi sensor are diverging and provide a notification alert, to investigate potential problems.
This method of notification has been a reliable method of identifying changing conditions in the retail store or potentially, a malfunctioning sensor.
The PIR sensor also helps identify Wi-Fi signals that should be classified as outside traffic rather than inside traffic. For example, if the PIR sensor detects no body heat within the store, it will be able classify the Wi-Fi signal signature (signal strength and dwell time) as attributable to outside visitors.
From the above it can be seen that this invention provides a unique means of providing accurate data on shoppers in store motion. Those skilled in the art will also realise that this invention may be implemented in embodiments other than that described without departing from the core teachings of this invention.