Treating Your Idea Like It’s Sean Bean

Sean Bean does survive some of his roles, but as anyone that’s seen The Lord Of The Rings can attest, he doesn’t have to in order to make a successful product. In fact, where and how Sean Bean dies can be more important to a narrative than any of the surviving character’s contributions.

Last week I had an idea regarding ambiguity in online rating services. What if we made the current user trends transparent? The most straight forward application of this approach seems to be regarding ordering delivery online at aggregators like http://www.grubhub.com. Just show the traffic, the order numbers, the average amount spent on orders over a day/week. Let the numbers tick up, let the average shift by the cent in real time. Instant validity for high reviews, maybe something to look into if a poorly reviewed place has a lot of orders. This seems straight forward enough to prototype, but before I do I’d like to hit the idea with a hammer.

First, this is adding complexity to a visual field that seems to be functioning quite fine without it. Is there actually a need to validate a 1–5 rating compared to the cost of more clutter on the screen? And is there an easy answer to the clutter, before I bother with the experiment? There might be. What if, instead of a default sort-by-ratings tab there was a default sort-by-popularity tab. If you wanted more information you could click on the restaurant and get orders per day/week/whatever, but since all most people are interested in is food that at least meets a baseline of quality, a popularity metric factoring in both the customer rating and the amount of orders might just do it. Test that before adding clutter.

Second, given that I’m working on a fix for both ambiguity and potentially deliberate fraud, how do I know I’m not creating another security risk. What would happen if a restaurant ordered a bunch of fake stuff, packed their orders with cabbage, and sent them to the homes of their employee’s to boost short term visibility on the website? The employee’s presumably wouldn’t complain, the cost is low as a result of the fake ingredients, and for the minor cost of paying the driver the restaurant gets advertising. Insidious, but on the scale and the time frame GrubHub is operating on, ultimately irrelevant. With an entire cities worth of orders per day that amount of fraudulent data might boost a restaurant from the bottom five into the bottom 10, but it’s not going to create lasting impact on the popularity metric.

Sean Bean has survived the first act somewhat in tact. He may actually make it through this one.