Retail-Innovation.com recently wrote about Pizza Hut's "Subconscious Menu," which uses Tobii eye tracking technology along with "a mathematical algorithm" to determine which pizza toppings the customer is looking at in order to rapidly predict a pizza for them to order. You can see how it works in the video below.
Pizza Hut touted a 98% success rate of “recommending pizzas that customers enjoyed.” Sure, they enjoyed them, but was the predicted pizza EXACTLY what they wanted?
Just because you look at a menu item the longest, does it mean you want to choose it?
I'm not convinced that just because you look at something the longest, it means you want that option over others. Certain colors draw your attention and you may linger on a given item because it looks different from the others, or perhaps you're trying to figure out what it is. It would be interesting to play with this, unbeknownst to the user, and see the correlation between attention and sales.
I think this technology may be more applicable in a whole store, where you could track which items people look at the most while browsing. It would also be interesting to see the articles a person looks at a lot and yet does not buy. Did the user not buy because of price? Was the article appealing at first but then not upon closer inspection? Does the item constantly lose out to another comparable item?
Although it's less visually appealing, I suspect that using tracking on text on the menu would give more even weighting to items customers look at. When customers are trying to decide between items, they may spend more time reading the ingredients of the things they want. I wonder if they spend more time looking at the thing they ultimately do not choose because they may be mulling it over. If the camera could also judge facial expressions, it may be interesting because you could judge the reason someone was looking at an item and include that in the algorithm.
Predicting what the user is thinking is a marketer's dream, but I feel like Pizza Hut looks at too little data here to be truly predictive.
This technology is interesting, but I don't it's particularly useful in the Pizza Hut scenario. It did, however, spark a lot of debate and discussion within our design group. I think the fact that it could spark ideas means there is something to the idea.
The concept of predicting what the user is thinking is a marketer's dream. If you can predict something accurately, it means that you understand the drivers and if you can understand the drivers, you can target those drivers to better your offering. Simple eye tracking alone may not be enough to give really accurate predictions, but that information in concert with other information may provide great insight.
The verdict on this from my standpoint is it's cool technology but the algorithm is looking at way to small an amount of data to be truly predictive. I am also not sure there is a great customer need or pinpoint here. This is just something cool for the consumer to try out and have a bit of fun with.