It takes a lot of skills to build successful analytic products. (Bradford Cross outlined the process well). The roles, defined in a spectrum from customer to researcher might look like this, although the continuum is not strictly speaking accurate (click to enlarge):
Note that customer participation during the entire process is not optional if you’re trying to make money.
In a big company, you might see one person occupying each of these roles. Of course, as team size increases communication overhead rises and rapid release becomes impossible. This can be mitigated by assigning only three people to a product at a time who’s skills cover most of these areas, and lending them expert resources in areas they are short, only as needed.
Analytic products are such a such a multidisciplinary undertaking that in a data startup a founding team is at minimum three people. Ideally all are founders. There are probably exceptions, but that is the minimum number of bodies required to flesh out all these areas with passionate people who share the vision and are deeply invested in the success of the company. Someone needs to be good at and enjoy each of these roles.
Of course, the split doesn’t have to look like that. It could look many ways:
In the second configuration, one of you is the customer. That is ideal because you are building for the guy ‘on the next bench,’ (Steve Blank’s phrase, refers to the next soldering bench and the fact that silicon valley always worked this way until quite recently) but it doesn’t replace customer development.
The third configuration is more eclectic. The point being that there is no right answer, but you’d better have these bases covered.
For an analytics product, this is the customer development team. Nobody is in charge and everyone participates. Build a minimal v0.1 inside of one month that expresses your core insight and some key feature of the data. Then drive, fly or swim to a real customer and introduce them to your application in person. Give them access and keep them clicking each alpha release. Log everything. Monitor and evolve your metrics. If you lose your alpha users, pivot.