Over the past two years, we’ve explored three different methodologies that are promoted as lower-cost“substitutes” for in-person eye-tracking. Here’s a quick synopsis of what we found, through our parallel testing:
VAS is an algorithm, which claims to predict visibility and viewing patterns without actually conducting shopper research. However, we found this system to be completely inaccurate (relative to in-person eye-tracking) – and quite likely to provide misleading findings. Thus, we cannot recommend it for any application.
"Click on What You Saw"
This approach relies on shoppers to self-report the first elements that they “saw” while viewing a package. In our experience, we found that shoppers were not able to accurately recall what they saw, given how quickly (and sub-consciously) the eye moves. Instead, they tended to click on what interested them on a package. Thus, the approach has some value, if it interpreted as an indication of elements of appeal (rather than a measure of visibility).
Web Eye-Tracking (via WebCam)
Our parallel studies suggested that, while webcam-based eye-tracking has significant issues tied to sampling and study execution, resulting in an end sample size that is more Qualitative than Quantitative (based on the size of the starting sample).
In addition, it has been shown to not accurately measure shelf visibility due to the wider error range (vs. in-person tracking) and the small size of packs (when shelves are shown on a computer screen for eye-tracking). Thus, it is not recommended as a measurement of shelf visibility.
We did find that this technique may provide reasonably accurate viewing patterns for individual packages. This may prove to be a useful application for web eye-tracking if the sample size challenge can be addressed.
To learn more about our parallel testing and recommended uses of eye-tracking, please read these articles: