Lean Product Management: Test Before You Get In Too Deep
Imagine if you could order your favorite food or drink at a crowded bar, without having to wait for the waitstaff to notice your politely raised hand or your empty glass. This is the problem Uri Goren, co-founder of the startup ePsik, set out to solve, with a method for ordering via smartphones.
Uri and I met at the BizTec conference a couple weeks ago, when he attended one of my roundtable marketing sessions looking for ideas for how to penetrate the US market. He explained that his company already had reached an agreement with one of the leading Point-of-Sale vendors in the USA, meaning that his product would seamlessly integrate with the cash registers of all food service establishments using that POS system. His developers were poised to start the integration work, and he was already thinking ahead to the next step… once they have the integration done, how did I recommend that he start gaining traction in US restaurants?
Well, the best way to gain initial customers is to demonstrate that the system increases a restaurant’s sales, decreases its expenses, or increases customer satisfaction. Uri explained that ePsik didn’t have that data, yet, since the POS integration hadn’t been completed. Wait a second, I said to him, are you sure that restaurants will even want to use your solution? That their customers will change their ordering habits? You don’t need a fancy POS integration to show that.
The entrepreneurs at ePsik had already spoken with several local restauranteurs and pub owners, who had indicated they might be willing to try ePsik’s solution. The big question mark was what would the customers do? Would they be willing to change their habits and trust ordering from their smartphone? Or would they stick to their usual method of flagging down a server?
I suggested that they take a lean approach, that is, before they start integrating a single line of code (or perhaps in parallel), that they try out their system locally, here in Israel, using whatever it takes to make it work, even if Uri himself has to stand one night at a restaurant’s cash register, “translating” the orders coming in from people’s smartphones into usable data that could be entered manually into the cash register.
So they spoke with one of the bar owners they were already friendly with, who agreed to let them run the experiment at his bar one night. Unfortunately – or rather, fortunately – the experiment revealed several previously unconsidered flaws in their methodology. The end result of that one-night experiment was that not one single customer was willing or able to order using ePsik’s system. Some simply didn’t understand that they could order with their smartphones, and others understood but were not able to, due to the dim lighting in the bar. Clearly they needed to improve the solution and re-run the experiment.
While these results may seem disheartening, ePsik needs to think of them positively, since they are getting them one step closer to a product that works, and that meets the needs of their target audience. Imagine if they had spent weeks working on the POS integration, started convincing American bar owners to try them out, and then encountered the same results, having flown across the ocean to test out their new product in their first US beta user. Now that would be disheartening.
Let me boil this down to a few key tips:
1) Devise experiments to determine if your customers (and your customers’ customers) will change their habitual behavior in order to use your new solution.
2) Each experiment should have a defined hypothesis, or learning goal, in order to keep it focused.
3) Try your experiment(s) as close to home as possible. Why waste plane fare and hotel expenses to test something you can try in your own neighborhood?
4) Bear in mind that moving your experiment to a different geography will impact its results. That is, pay attention to the results you get close to home, but don’t let it make you expect that people in the USA will behave the same way. I have seen too many startups assume that early success in Israel will translate to success overseas, and unfortunately, that is not always the case.
Have you ever run a “lean” experiment while in product development? Please share in the comments what you did, and what you learned.
Image by flickr user Thomas Hawk