As most of you probably don't know, Retrievr is an idea for a service that fosters relationships between artisanal coffee shops and their customers by combining a social check-in service with a simple feedback mechanism. We came up with it when we attended Lean Startup Machine a few weeks ago. Read more...
I think the Retrievr story is a worthwhile case study because it illustrates an interesting and unanticipated problem: when you build a two-sided product, you can't put the cart before the horse. We focused our efforts on the coffee shop owners first because they were the paying customers in our model. We had early success and even signed up paying customers using wireframes.
The coffee drinkers were supposed to be altruistically motivated to leave feedback, and if that was the case a wonderful loop would ensue. But now we're learning that coffee drinkers have entirely different attitudes towards feedback. In an ironic twist, they consistently validate our hypotheses that customers want to provide feedback and go out of their way to help their favourite coffee shops. But they thoroughly reject our proposed solution. And so the lynchpin that's holding both sides of the two-sided system together seems to be evaporating.
We made a crucial mistake: we didn't validate the driving side of the system first (giving feedback). Instead we focused on the paying customer (paying for analytics). Even worse, we accepted payments based on data we couldn't prove we could provide. We should have focused on the coffee drinker first, even though they wouldn't be the paying customers. Our challenge would then be to create value by integrating the coffee shop into whatever that looks like.
So, if Retrievr is going to continue, it will require a very severe pivot - definitely severe enough to send back the money we've collected so far. There might even be more promising products for us to run through the cycle.
Let me know what you think on twitter.