Thoughts on Product-Market Fit

Giff Constable startups

archerytargetHere are a few thoughts on product-market fit that came up when I was chatting this evening with Brant Cooper of Market by Numbers.

Revenue & Testing

To me, “lean startup” and “product-market fit” boil down to rigorously and continuously testing your assumptions as early as possible, and holding off heavy investments in scalability and growth until you feel confident in the drivers of your business.

Your revenue model is one of those assumptions, but it doesn’t have to be the very first thing empirically proven.  I am thinking specifically of consumer-focused services with variations of freemium (for example, Evernote or Zynga) or business models which separate the end-user and the revenue source / payer (for example, Daily Candy or Google).

In these cases you can build early confidence in your plan with customer development conversations (in the latter case above customer = both users and payers), as well as close observation of customer (or payer) behavior with similar products.

Take the example of a game which makes money from virtual goods.  You have a simple foundation — the game must be fun. Without that, you have nothing.  So you can first test fun, then virtual goods compulsion loops and most likely viral behavior, and then finally whether people will indeed pay for those virtual goods.  The key is that you are testing assumptions all the while, rather than waiting to do it all together after a longer product dev cycle. You are carefully choosing the order of what you test (for example, a more utilitarian product might want to test monetization earlier).  You are keeping an open mind and an eye out for the necessity to change/pivot.

The RIGHT Product-Market Fit
I like Sean Ellis’ target that 40% of your users should say that they would be “very disappointed” if they could not use your product.  It feels like a good stake in the ground.

I think that you must consider customer demographics (UPDATE: I’m using the word demographic loosely here — it’s about understanding customer commonalities and segmentation) when you do this kind of testing. Who is making up this data set and how does that fit into your assumptions?

If you were shooting for teens or 30/40-something women and you end up with the TechCrunch set all jazzed up, you need to ask yourself: “Should I give up on previous assumptions and instead aim at the TechCrunch set? Do I believe that I can effectively monetize that type of customer? Will I get stuck here, or will I make a leap to other types of users?” And if those answers are not coming back positive, “Am I willing to risk losing that momentum by changing my product to make it more compelling to mainstream users today?”  As with all this stuff, there is no universal right answer. Every startup has to navigate their own way.

Chris Dixon had a great post on this called “Techies and normals“.  I don’t believe that “techie” defines “early adopter” to the exclusion of all other types of people.  Technology is too pervasive in our culture now to be that simplistic.  You will have an adopter curve among “normal” (i.e. more mainstream) users too.

Steve Blank writes about not getting ahead of yourself by designing a product for the fat part of the adoption curve, and his argument makes sense.  You do need to think about your early adopters, but within a target demographic.  As always, this is about testing assumptions and being ready to pivot, either on design or on target market, if your assumptions were wrong.

(image from tony newall on flickr)