Can virtual shelves predict advertising effectiveness much better than surveys?


Will the ad increase sales? In most pre-testing protocols, sales impact is “measured” with explicit methods. However, the correlation between ‘real sales’ and ‘what people are saying’ is extremely low (-0.1 to 0.2). On the other hand, studies have shown that virtual shopping data has a much higher correlation to market shares (up to 0.95). Can virtual shelves be used to pre-test ads?
EyeSee conducted an experimental study of more than 1500 shoppers and 42 commercials to understand the potential of virtual shelves when predicting advertising effectiveness.


A questionnaire taps in the conscious mind – shopping is influenced by emotions and sub-consciousness

Millward Brown’s Link test is the most well-known pre-test for TV Commercials. Until recently, it only existed out of a survey. However, a survey is a tool to measure conscious factors, and is rarely sufficient to collect data for a person’s purchase process. Purchase decisions are far less rational than often believed; they take place in the subconscious mind and are driven by emotional and automatic judgment processes. For this reason, adding virtual shopping is a better way to pre-test TV Commercials.


Virtual shopping: a quick introduction

EyeSee Research Virtual Shopping Interactive platform for packaging testing

Interactive shelf testing platform

Shoppers are recruited online to conduct a virtual shopping trip at home with their own computer. They are exposed to multiple store environments, they can click on any product to zoom in and obtain more info about the product and indicate the quantity they want to buy. The platform combines the benefits of:

  • Real-life shop experience: visual stimuli and interaction (zoom in and navigation)
  • Online surveys: Fast, cost effective and globally scalable

You can read more about this technology here.


What we found out after testing 42 TV commercials with over 1500 people

EyeSee selected 21 well-known brands. For each brand we chose 2 TVCs for the same product. The commercials were present on YouTube and other social networks, so we measured each ad’s popularity based on the number of YouTube views it had. We divided the products into 9 categories and constructed 9 neutral shelves, one for each category. Each shelf included the best-selling brands and products for that category.

We recruited more than 1500 people via panel partner GMI Lightspeed, and divided them into 14 cells. In each cell, people were exposed to 4 TV commercials and they conducted a virtual shopping trip with 3 to 4 shelves, followed by filling out a small survey.

To obtain a base scenario for comparison, we have shown a virtual shelf to 3 different groups:

Virtual shelf EyeSee Reserach Eye Tracking Testing

3 respondent groups

We obtained 100 responses per group. Ideally, we would have pursued 300 to 500 responses per group, but we preferred to decrease the impact of a single commercial, showing 5 types of commercials instead (each with 100 responses). The main goal was to prove the concept on an aggregated level.


More popular TV commercials can double the increase in sales

Increase in virtual market share EyeSee Research

Increase in virtual market share

More popular ads increased the market share by 6.2%, while less popular ads only by 2.9%. Each group exists out of 20 ads with 100 responses each, so these results are significant.


Top 15% ads are responsible for 40% of the increase in sales value

The table below shows the increase of sales per ad level. Although this is only for illustration purposes as the sample is too low per ad, it shows an interesting distribution over the different ads:

  • the top 15% of ads increase sales by 17% on average
  • the upper middle 35% by 12%,
  • the bottom middle 35% decrease sales by 2%
  • the worse 15% of the ads decrease sales by 5%

The top 15% of commercials are responsible for 40% of the value that is created.

TVC testing EyeSee Research Eye Tracking copy

Difference in Virtual Market Share between group with and without TV commercial exposure


Measuring sales with virtual shelves could be the best answer.

In the end, the thing that matters most in a store is sales, but it is very hard to allocate ROI of advertising based on the actual sales only – other factors need to be controlled for as well. Take beer giant AB Inbev for example, in which of the following 2 scenarios will they sell better?

  • Great hot summer with a terrible ad
  • Terrible summer with a great ad

Furthermore, this approach cannot be used for pretesting.

At EyeSee, we are strong believers of triangulation of data. However, virtual shelves have a lot of (if not the most) potential to explain differences in ROI of a TV commercial in an experimental environment in which we can control a lot of factors (even the weather!). The results above show:

  • Impact of advertising on sales
  • Difference between ‘good’ and ‘bad ads’

EyeSee Research made virtual shelves accessible – fast, cost effective and global. So, what are you waiting for? Contact [email protected] to learn more!