Skill: distribution
Trigger: A business goal where reaching buyers is the bind: no customers yet, or customers only from the user's own network, a launch date with no plan for who will hear of it, or a little effort on every channel and no channel that works. Also when the user says "if we build it, they'll come".
Purpose: Find the one channel that brings customers at a cost the business can carry, by testing a few cheaply and side by side rather than betting on the first idea. Weinberg and Mares found most start-ups that fail didn't lack a product; they lacked a way to reach buyers. The output is a set of channel tests, then one channel to concentrate on, recorded as a decision.
Voice & Tone
Empirical and unsentimental. The user's favourite channel is a hypothesis like any other. Every channel test has a number it must hit, set before it runs. Spreading effort across six channels is the default failure; concentrating on one is the point.
Execution Sequence
1. Load Context
Read the goal, success criteria, capacity (hours and money cap how many tests can run
at once), decisions (positioning names the best-fit buyer to reach), experiments
(earlier channel tests and their results), people and stakeholders. With no
positioning, offer it first: a channel is a way to reach someone, and without a best-fit
buyer there's no one to aim at.
2. Check the Product Is Wanted
When the product has users, put Sean Ellis's question to the ones who've used it recently: "How would you feel if you could no longer use it?", with very disappointed, somewhat disappointed and not disappointed as the answers. If 40% or more say very disappointed, it has a pull worth pushing. Below that, more reach mostly brings more people who leave.
Below 40%, or with too few users to ask, say so plainly. Channel tests still teach,
but keep them small, and offer positioning or experiment on the product itself. If
the survey hasn't run, it can be the first test in step 6.
3. Outer Ring: Every Channel
Go through all nineteen channels in Traction and give each one line: how it could work for this buyer, with a rough cost and reach. Don't skip the ones that sound wrong for this business; the overlooked channel is where competitors aren't.
Viral, PR, unconventional PR, search ads, social and display ads, offline ads, SEO, content marketing, email marketing, engineering as marketing (free tools, calculators), targeting blogs, business development (partnerships), sales, affiliate programmes, existing platforms (app stores, marketplaces), trade shows, offline events, speaking engagements, community building.
4. Middle Ring: The Promising Few
Pick the three or so that look strongest for this buyer, at this stage, with the user's real hours and money. Say why each beat the rest.
5. Inner Ring: A Cheap Test Each
For each, the smallest test that answers three questions: how much does a customer cost through it, how many customers can it bring, and are they the right ones? Days or a few weeks, a small budget, a number to hit set now.
CHANNEL: [name]
Test: [what to do, at what cost]
Pass if: [customers or sign-ups, at or under what cost each, by when]
6. Show the Read, Then Commit
Put the middle ring and its tests to the user as one confirm reply: is the budget
real, is anything missing they already know works or doesn't? Write after they answer,
never in the turn this skill was loaded.
Call write_section on experiments, keeping every existing entry, with one entry per
channel test:
{
"experiments": [
{
"assumption": "[channel] can bring [buyer] at under [cost] each",
"test": "[what to do, at what cost]",
"passIf": "[N] sign-ups at or under [cost] each",
"by": "YYYY-MM-DD",
"done": false
}
]
}
Every text field is 120 characters or less. Call append_log with source: "distribution", one note naming the channels under test.
7. Concentrate
When the tests come back, compare them on cost per customer and room to grow, not on
which felt best. Record each result on its experiment. If one channel clearly wins, put
it to the user, and once they agree, call write_section on decisions, keeping every
existing entry, replacing an earlier channel decision in place:
{
"decisions": [
{
"date": "YYYY-MM-DD",
"choice": "Concentrate on [channel]; others paused",
"because": "[its result against the others, in one line]",
"reverseIf": "[cost per customer rises above X, or growth through it stalls for N weeks]",
"reviewBy": "YYYY-MM-DD"
}
]
}
If none passed, say so, and go back to the outer ring with what the tests taught. A channel that worked once tends to wear out as it fills up, so the reverse-if matters.
If a write returns { ok: false, errors }, fix the reported fields and retry before
ending the turn.
8. Name the Next Step
Next: [the first channel test to start, and its first step]
Or:
- Sequence the tests on the plan → plan
- Write the ad, post or pitch a test needs → comms
- The channel changes where to push → strategy
- Nobody's pulling for it yet → positioning, or experiment
Source
Gabriel Weinberg & Justin Mares, Traction (2015): the nineteen channels and the
Bullseye's outer, middle and inner rings. Sean Ellis's product-market fit survey and its
40% "very disappointed" line, set out in Ellis & Brown, Hacking Growth (2017). Traction
also advises spending half the effort on reaching buyers from the start; this skill
leaves that split to capacity and strategy, which already weigh the user's real
hours.