Groupon Co-founders invest $4M in fast-growing E la Carte (YC S10)
techcrunch.com
techcrunch.com
About a year ago, I think I said something to the effect of "A/B tested restaurant menus would print cash." (Edit: here we go: http://news.ycombinator.com/item?id=2046426 ) If they're getting +12% sales just on superior interaction design (versus actually talking to a human -- and I have no difficulty believing that the average interaction with a device measurably beats the average interaction with a human for at least some restaurants on axes both customer and owner care about) I predict very, very, very nice things in their future when they decide to take an engineer or three aside and become a data company in addition to being a devices company.
Sprinkle a little code on a techphobic multi-billion dollar industry, watch wonderful stuff happen.
They gained 10% from those who stuck around and points to them for working the up-sell on side orders, but how many left never to return? To quantify the value of tabletop ordering systems is not as simple as saying "oh look, a 10% gain in average order value" it's much more complex than that. And without the customer numbers and order value variance it's hard to truly analyze the success of the proposition.
Why?
Personally, I'd love to see this. I particularly liked the bit about filtering and sorting the menu. And I could often use an option to randomly choose between a few options that sound appealing.
Really, though? The choice is so difficult that you need a battery powered LCD display tablet to do "rand()" for you?
Qualitatively: ask your waitstaff if they keep getting called over by customers saying "I pressed the devil box and it didn't work! I hate this!"
Quantitatively: you get credit card numbers and email addresses. See who comes back. (Bonus points: if they don't come back, ask them why not.)
If you're collecting their email addresses to give them their receipts, this is probably not as problematic as you think.
On the other hand, I'd love it if the places to which I've never returned asked me why, and then changed as a result and invited me to try again.
If you want a more efficient restaurant, this is certainly not the best way to get it. Put the menu up on a big board like McDonalds, Chipotle, Panera, or other already high volume eateries like street vendors. Want to be even faster? Cut down the menu complexity. Fewer menu items means less customer analysis paralysis and a more streamlined kitchen. Anyone in the restaurant business already does at least some data analysis on every sale and have regional and store by store A/B testing. I mean, they have to know what's selling when they do inventory and order more food.
What doesn't make you more efficient (i.e. turn orders and tables faster)? Giving just one menu to a table full of people then include a bunch of games and surveys on it. But if you want to have a table gimmick to keep people around ordering more appetizers and rounds of beer, this might be the ticket.
http://tech.fortune.cnn.com/2011/06/10/groupon-eric-lefkofsk...
I would presume there would be some retraining also such as how to handle technical failures during an order (what to say and how to still keep it streamlined through old manual methods) but if all that stuff is sorted this looks great !
I mean playing games when your waiting and nutritional information is great. I do that tim ferris diet and knowing if something has sugar or the forbidden ingredients in them would be good to know. Also it would provide important info to people with allergies to nuts and other similar kinds of foods.
Either way, I hope this takes off.
This would never replace "regular waiter service" but for those who want faster service, with no questions, it's a win-win.