Do Virtual Stores Validate vs Real Store Tests?

Virtual Stores - Can They Be Trusted?

Shopping in computer simulated Virtual Stores, as a means of researching shopper behavior, is exciting much press coverage at the moment. It’s not that the technique is new, more that it’s particularly topical, given the dynamics in today’s stores.

Both Retailers and ‘Big Brand’ Manufacturers need to offer the best choice of product at the right price to an ever changing customer base - particularly difficult in today’s spatially challenged retail stores and tight profit margin constraints. Recognizing these challenges means there will always be keen interest in better, cheaper, faster ways of understanding the most cost-effective and profitable way to innovate product assortment and implement visual merchandising changes in-store.

‘The Use of Virtual Store Simulations in Marketing Research and Beyond’ outlines various aspects of Virtual Store simulation and its advantages in research, but little is said about the subject of validation - i.e. whether or not insights gathered in Virtual Store simulated shopping studies are accurate representations of what could be learned in real store studies, and whether or not such insights can be relied upon to accurately predict real shoppers’ behavior when subjected to the same test stimuli in a real store.

It’s a Good Question, and one that comes up whenever a client is interested in adopting this approach. The answer is - Yes; Virtual Store derived shopper insight does validate well with the real world. But, you must be mindful of the traps you can fall into along the way. It really does matter that the study is well designed, that the product category is appropriate for the technique and that the actual research work is carefully managed in-field. However, the question has been raised many times before, so you might be interested in how it’s been addressed over recent years.

First Attempts at Virtual Store Validation

The recognized academic expert in this field, Professor Raymond Burke, first published work in this area in 1992. His paper ‘Comparing Dynamic Consumer Choice in Real and Computer Simulated Environments - 1992′ describes an early study using two simple laboratory setups; one ‘text’ based, the other designed to simulate a basic but viable ’shopping’ experience.

In the simulation, dubbed their ‘Realistic Lab’, a computer showed photographs of products arranged on a shelf just as in a store display. Some product groups required several screens to display all the products and so the subjects were moved among screens by a simple trackball interface. Product ‘purchasing’ was simulated by the participant ‘zooming in’ on a selected item and confirming the purchase simply by incrementing a counter. Each of the studies ten participants shopped for 30 minutes.

A parallel study was also conducted in a real local store, recording the participants’ shopping behavior over a seven month period. The participants were also subjected to a lengthy, computerized questionnaire. This provided an essential benchmark for comparison with the computer simulated shopping tests.

In the paper Burke discusses many influences on product purchasing, distilling them into a set of hypotheses, which he could use as the first foundations for validating the technique. I have summarized the discussion below.

Virtual Store Simulation Factors - Burke’s Purchase Decision Process

Burke recounts, from earlier studies by others, that when purchasing products, shoppers typically go through the following process:-

• Search for information
• Evaluation of alternatives
• Specific choice
• Post-purchase evaluation and learning

Purchase Decisions - Search for Information

In the ‘Search’ stage, products are often selected on the basis of brand recognition, brand loyalty, referral, habitual routine or simply by lowest price. These factors can be relatively easily reproduced in a simulated environment and this led Burke to the hypothesis that:-

‘H1 - Consumer choice behavior in the laboratory will more closely match actual shopping behavior in product categories in which purchase decisions are based on cues that can be easily reproduced in the simulation, such as price and brand name.

He also suggested that where the product’s physical size is a relevant factor, the ability of a simulation to accurately predict purchase will depend upon the realism of the simulation; which led to the second hypothesis:-

‘H2 - The correspondence between laboratory and actual market shares of goods by sizes (rather than by brands) will be higher in more realistic simulations, which impart visual size information, than it will be in rudimentary simulations.’

Purchase Decisions - Evaluation of Alternatives

In the ‘Alternative Evaluation’ stage it was suggested that ‘price and in-store promotion often became separated from the brand image in the buyer’s mind and exert an independent influence on choice’. Also that Gabor’s work claimed that participants might be more price sensitive in a simulated environment than in a real store, particularly when the shopper’s awareness of all price options is less than 100% complete,which led to;

‘H3 - Subjects will be more sensitive to price promotions (i.e., features’) in the laboratory than to price promotions in the actual grocery store.’

However, if the store environment simulation was highly realistic, then other factors might lessen the effect of price, and so;

‘H4 - The correspondence between purchases made on promotion in the laboratory versus the supermarket will be higher in more realistic simulations than it will be in the rudimentary simulations.’

Purchase Decisions - Choosing a Product

…… to be continued in Part 2 …………

Tagged as: Consumer Behavior, Consumer Insights, Consumer Research, Market Research, Retail Insights, Shopper Insights, Virtual Shopping, Virtual Store