Picking your first customers: the gradient of influence

At Bloomberg Beta, we often ask founders who they will target first to use their product.

Your choice of first customers is one of the most underappreciated, important decisions in the early life of a startup. It’s a one-way door, and one-way doors are expensive (it can be excruciating to turn back the clock once you start serving customers).

Palantir will always be known as a company that got started serving government security agencies. Facebook is a child of Harvard, now and forever. Bloomberg was used first by traders of US government bonds. LinkedIn’s first community will always have been the technorati. OUYA, where I’m chairman, is still known as a project born on Kickstarter.

Since Bloomberg Beta focuses on “day zero” companies that are just getting going, and we focus on helping our companies with going to market, this issue often comes up for our founders just after we invest.

How do most companies answer this question? Plenty of unsatisfying ways:

  • We’ll try a lot of different customers and see who bites (at least they’ll get data!)
  • We’re still trying to figure it out by interviewing potential customers (fair enough)
  • “Early adopters” (um… often true by definition)
  • We like these customers because they’re the most complex, and if we can figure them out we can get everyone else (often times a bad idea, since you fall over and may never end up succeeding with anyone)
  • We’re focusing on [industry X] because… (this is a start, at least it’s a hypothesis)

The “top 10%” answer is to think about the kinds of customers that (a) have greatest need for the product and (b) are easiest to attract and serve.

Yet this approach might be misguided. It works for maximizing the value of *those customers* to the company. It may fail to maximize the value of the company, though, because it neglects a critical effect: some kinds of customers have a powerful effect on the future customers you might get, a spillover effect. Others, less so.

One way to think of it: your first customers are your first hires as marketers. You want them to be as good at their job as possible.

Imagine all your customers as you did before (in terms of how valuable they’ll find you, and how easy they’ll be to attract). Then, consider that some of them have influence on others. You’ll find pockets where one kind of customer influences another, and some customer classes who are widely influential, or influential to several important adjacent customer classes.

Think of this as a gradient of influence among your customers.

You want to target a class of customers, at first, who are as high up on that gradient as possible — even if they’re harder to attract (as I imagine security agencies were for Palantir) or you make less money from them (as OUYA did from its Kickstarter backers, where it priced the product as low as possible before knowing the cost to serve each customer).

Often, the type of customer who will ultimately pay your bills is not the best first customer, and that’s counterintuitive.

Notice I’m describing customer classes vs. individual customers. Many startups overweight the value of that one Fortune 50 logo, vs. the value of being known for serving one class of customers well. Better to dominate some initial market and then expand, than to have dispersed pockets of usage where each customer is famous.

There are many ways to slice the universe of potential customers into classes:

  • Industry — “tech companies”
  • Function — Bloomberg got its start focusing on bond traders
  • Reputation — “Top 100 best places to work companies”
  • Community or event — Foursquare sparked a fire at SXSW

Think about it like this: what’s the class of customers most admired by the entire universe of customers you might have? Define that admired class as narrowly as possible, to make it easier to find and serve them, and to get critical mass in at least one market.

One advantage of this approach: it makes it easier to avoid having to launch. Yes, freedom from launching is an asset — “launch” is another expensive, risky, often-unnecessary one-way door.

If you start to see gradients of admiration among your potential customers, it might prompt you to think differently about where to begin.

  1. roybahat posted this