Turnkey platform for conducting shopper studies in virtual reality, based on physiological feedback and machine learning methods
Shopper research is at the threshold of technological evolution
Every FMCG manufacturer and retailer understands the importance
of researching customer behaviour at the point of sale, however, current approaches are inconvenient and costly due to their low technological effectiveness.
of researching customer behaviour at the point of sale, however, current approaches are inconvenient and costly due to their low technological effectiveness.
Research in a real store entails organizational difficulties: the need for an agreement
at each point, real traffic vs. expected traffic, the impossibility of changing the layout, practical impossibility of installing surveillance cameras, etc.
at each point, real traffic vs. expected traffic, the impossibility of changing the layout, practical impossibility of installing surveillance cameras, etc.
A hall test with an imitation of store shelves gives a limited number of test options (maximum of 5-6 options for the object of study:
layout planograms, POSM locations, etc.).
layout planograms, POSM locations, etc.).
Conventional quantitative polls about consumer behaviour only reflect an opinion about the fact:
"What would I buy while in the store?"
"What would I buy while in the store?"
A set of methods may be a solution, but it would mean long processing time and data compilation, and it is expensive.
25%
cheaper
50%
faster
* In comparison to conducting a similar study
at a real store
at a real store
1
2
3
A unique, fully automated platform
for conducting shopper research
for conducting shopper research
Ready-made leading retail chain stores of various formats
in a virtual reality of the highest level of graphics quality.
in a virtual reality of the highest level of graphics quality.
The maximum quality of data collection, unattainable during
traditional research (millimetres, milliseconds, fractions of degrees, etc.).
traditional research (millimetres, milliseconds, fractions of degrees, etc.).
Data is available for analysis and processing in real-time - right
in the process of research.
in the process of research.
1
Determine the research objective, store chain and its format.
5
SIMPLE STEPS
OF THE PROJECT
OF THE PROJECT
Adapt the location and coordinate the necessary
research quests.
research quests.
Select relevant respondents.
Set up a mobile research VR site at 4-16 stations
(in any convenient location).
(in any convenient location).
Quickly process the research results and provide them to the customer
in the usual presentation form or the form of a multimedia presentation.
in the usual presentation form or the form of a multimedia presentation.
2
3
4
5
(depending on the materials providedand the difficulty of the quests)
The time required for the preparation
and conduction of the research
and conduction of the research
10
30
Thanks to many years of experience in creating VR products with the highest level of graphics quality, we were able to achieve the effect
of complete immersion, paying particular attention to detail, lighting and physics of the objects. Respondents do not hesitate to transfer their skills and habits from real life to the virtual space, and the research process itself is no longer tedious and boring. The platform has the flexibility to adapt to a specific study.
of complete immersion, paying particular attention to detail, lighting and physics of the objects. Respondents do not hesitate to transfer their skills and habits from real life to the virtual space, and the research process itself is no longer tedious and boring. The platform has the flexibility to adapt to a specific study.
Maintenance of planograms.
Procedural changes of the prices on the price tags.
Adding POSMs and product installations.
Changes to the different categories locations within the store.
Procedural changes of the prices on the price tags.
Adding POSMs and product installations.
Changes to the different categories locations within the store.
PHOTOREALISTIC VR STORES
OF ALL FORMATS
OF ALL FORMATS
As well as the next additional parameters:
Changes to the level of cleanliness and order in the store.
Adding the required number of other customers in the store.
Reproduction of advertisement audio messages through the store's
speakerphone.
Adjustments to the navigation in the store.
Cashier's behaviour and more.
Changes to the level of cleanliness and order in the store.
Adding the required number of other customers in the store.
Reproduction of advertisement audio messages through the store's
speakerphone.
Adjustments to the navigation in the store.
Cashier's behaviour and more.
Using the Retaility platform, you can conduct any possible research in retail*:
STUDY OBJECTIVE
POS MATERIALS
What kind of materials do help customers? How to set the POS right?
STANDS, DISPLAYS, DMPs
Would investing in the equipment bring results? What is the customer conversion rate?
STORE REDESIGN
New stores? Or design in comparison
with the previous format?
with the previous format?
COMMUNICATION
What to say and how to say it to customers
to make sure their visit ends successfully
and creates a positive experience?
to make sure their visit ends successfully
and creates a positive experience?
RETAIL SPACE PLANNING
How to make a store a platform
for a positive interaction experience?
for a positive interaction experience?
ORGANOLEPTIC ASSISTANCE
(SOUND / ODOR)
(SOUND / ODOR)
How neuromarketing can create
a pleasant shopping experience?
a pleasant shopping experience?
* The relative error in researching via VR in comparison to a real store is less than 10% (confirmed by parallel studies using identical stores - real and virtual).
Automatic labeled data acquisition allowed us not just
to compare pre-assembled options with each other, but to create unique algorithm for automatic forming of efficient shelf impact recommendations. Our model highlights the most viable options
for further testing and eliminates weak options based on partially assembled data sample.
With real-time hypothesis modification during the research process machine learning methods will scale back the number
of test runs.
to compare pre-assembled options with each other, but to create unique algorithm for automatic forming of efficient shelf impact recommendations. Our model highlights the most viable options
for further testing and eliminates weak options based on partially assembled data sample.
With real-time hypothesis modification during the research process machine learning methods will scale back the number
of test runs.
OPTIMUM COMBINATION
PRESCRIPTION
PRESCRIPTION
Consumer behaviour research in other formats
AR solution for conducting remote package and marketing communication elements studies
GAMELAB
Interactive online platform for carrying out remote gamified research
ARTIFACT
