Track 13 · Founding · lesson 4
Distribution, before product, again
11 min
Track 13 · Founding · lesson 4
11 min
Ask a hundred founders how their first ten customers arrived and the answers are boringly similar. Someone they already knew. Someone that person introduced. Someone they messaged directly, one at a time, having found them by hand. A group where the problem was already being discussed.
Almost never: a channel that scales. Almost never: people who found the product on their own.
The first ten are a manual, unglamorous, personal process, and the reason to say so plainly is that founders who expect otherwise conclude that something is wrong with their product when nothing is.
Track 3 made the case that distribution beats product more often than anyone admits. This is the same claim, moved earlier: distribution is a constraint on what you should build, not a phase that begins after you have built it.
The question how would the hundredth customer hear about this has an answer before anything is made, and the answer disqualifies ideas. A product with no plausible route to its buyers is not an early-stage business with a marketing problem. It is a different business you have not designed yet.
People you already know. Fastest, warmest, and a limited pool that does not refill. It also produces the most misleading feedback, because some of them are buying you rather than the product.
Introductions from those people. One step out, still warm, and the pool is several times larger. The cost is that each one requires an ask, and the asks are finite too.
Places where the problem is already discussed. Forums, trade groups, local associations, the comment section under a competitor's announcement. Slow, cheap, and the people there have pre-identified themselves as having the problem.
Direct outreach, one at a time, by hand. Find the specific person, write something that proves you understand their situation, ask for twenty minutes. Low response rate, high information per response, and it works in exactly the markets where nothing else does.
Borrowing someone else's audience. A partner, a supplier, a consultant who already sells into your buyers. Fast when it works, and it hands a share of the relationship — and often the margin — to someone else.
Predict
Every route above works better when the person doing it built the thing, and for reasons that have nothing to do with sales ability.
The founder can change the product mid-conversation. When a customer says the thing is nearly right except for one detail, a salesperson takes a note and a founder makes a decision.
The founder can hear the objection properly. A salesperson hears we said no. The founder hears the specific sentence that preceded it, and that sentence is usually the most valuable output of the whole month.
And the founder can price badly on purpose. The first customers are being paid for information, and a discount, a pilot or a free month is a reasonable price for it — as long as everyone understands that is what is happening.
Not the same thing
Manual, personal work per customer: writing each message yourself, onboarding by video call, doing part of the job by hand.
A route where money or content goes in and customers come out, without a specific person being persuaded each time.
Which is which? Put each one on a side.
The first ten customers of a new business
Customer four hundred, in a business with a known buyer and a stable message
Entering a second market where the buyer is a different kind of person
Before the first sale, one figure decides whether any of this is viable, and it is knowable in advance: how many potential customers exist at all.
Eleven companies with the problem is a consulting business. Four hundred is a small software business. Four hundred thousand individuals is a different animal entirely, with different economics and a different route to every one of them.
Counting the market by hand — actually listing names until you get bored — is the cheapest research available and it is skipped almost universally, because the answer is often discouraging and the exercise takes an afternoon.
Check
Nobody writes a case study about six hundred messages. It is repetitive, most of it produces nothing, and the parts that work are indistinguishable from the parts that do not until weeks later.
What gets written about instead is the channel that eventually worked at scale, described as though it had always been there. The story runs from the product to the growth curve, and the eighteen months of one-at-a-time conversations in between are compressed into a sentence.
That compression is not dishonest. It is just the same survivorship filter as everywhere else: the interesting part of the story is the part that scaled, and the part that scaled was built on top of a year that was not interesting at all.