Traction, and the bullseye framework

What is the bullseye framework?

Three rings, worked from the outside in.

The outer ring is everything possible. The book names nineteen channels: viral marketing, public relations, unconventional PR, search engine marketing, social and display ads, offline ads, search engine optimization, content marketing, email marketing, engineering as marketing, targeting blogs, business development, sales, affiliate programs, existing platforms, trade shows, offline events, speaking engagements, and community building. You write an idea for each one, including the ones you consider absurd. The point of the exercise is not the absurd ideas. It is that you cannot default to the channel you already know if you have been made to write the other eighteen down.

The middle ring is what looks probable. From the long list you pick a handful, and you run a cheap test on each: not a campaign, a test, designed to answer one question for the smallest amount of money and time that can answer it.

The inner ring is what works. The book's strong claim is that at any one time a single channel drives most of a company's growth, so once a test comes back positive you stop spreading and go deep on that one channel until it is exhausted.

Running alongside all of it is the book's other argument, the 50 percent rule: spend half your time on product and half on getting customers, from the beginning, rather than building for a year and then thinking about distribution.

What does the book get right?

Three things, and the first one is the reason to read it.

It makes channel choice an explicit decision. Almost every early company picks its channel by accident: whichever one the founder has done before, or whichever one a competitor is visibly doing. Writing nineteen lines on a page breaks that, and it takes an hour.

It separates the cost of a test from the cost of a channel. Teams routinely refuse to test a channel because running it properly would be expensive, which confuses two different numbers. What you need first is the price of an answer.

The 50 percent rule is the part most founders skip and most need. Half your time on product, half on getting customers, from the first week. The useful version of it is a calendar test rather than a principle: open last week's calendar and add up the hours that went to building against the hours that went to finding customers. Almost every technical founder I have done this with finds a ratio closer to nine to one, and almost none of them knew it before they counted. The book's claim is not that the right answer is exactly half. It is that a founder who has never counted is always further from half than they think, and the gap gets more expensive the longer the product takes.

It takes a channel's ceiling seriously. The book pushes you to ask how many customers a channel could ever give you, rather than only what one costs today. This is the question that gets skipped, and skipping it is how companies build a team around a channel that runs out at a tenth of the volume they need. Three questions per channel, and the third one is the expensive one to get wrong: can you reach these people here, can you afford an answer, and could this ever carry the volume you will need in two years.

What has dated?

The framework survived. Two of its premises did not.

The cheap test is no longer cheap, because the platforms learn. In 2015 you could put a few hundred dollars into a paid channel and read the result. Now the paid platforms need a minimum volume of outcomes before their delivery settles. Meta's ad sets leave the learning phase after roughly 50 optimization events within seven days of the last significant edit (Meta Business Help Center, accessed Oct 2026). Google recommends judging a Target CPA campaign over 30 days with at least 30 conversions in it (Google Ads Help, accessed Oct 2026). A test below those thresholds returns a number that moves every week, and a founder reading it will usually conclude the channel does not work when what they have measured is the variance.

The measurement the test depends on got worse. Apple's App Tracking Transparency arrived in 2021, and four years in, about half of iOS users consent to tracking (AppsFlyer, April 26, 2024). Google removed the first click, linear, time decay and position-based attribution models from Google Ads and Analytics across 2023, leaving data-driven and last click (Google Ads Help, accessed Oct 2026). None of that stops you testing channels. It does mean a small test now produces a signal you have to interpret rather than a number you can read, which is a different skill than the book assumes.

One of the nineteen is not a channel. Viral marketing, as the book has it, is mostly a property of a product: whether using it requires or rewards telling someone else. You cannot test it the way you test search ads, and treating it as a channel sends teams off to build share buttons instead of looking at why nobody wants to share. That work belongs in referral.

The cost of the obvious answers rose. The two channels most founders shortlist are now the most competitive. Across thousands of Google Ads and Microsoft Ads accounts, the average cost per click was $5.42 and the average cost per lead $66.69 (LocaliQ, updated June 1, 2026). The book's instinct to look at the strange channels in the outer ring is worth more now than when it was written, not less.

So how do you run a channel test that produces an answer?

Four decisions, made before you spend anything.

Pick the event you will optimize toward, then do the arithmetic. Multiply the cost of that event by 50 and you have the weekly budget floor for one Meta ad set. If the answer is more than you will spend, choose an earlier event that happens five to ten times as often, and accept that you are then testing whether your message creates interest, not whether it creates customers.

Write the question and the kill date on the same line. "Does a developer audience on LinkedIn produce leads that clear our definition at under $900, judged on November 14." A test without a date becomes a budget line nobody will defend or cancel.

Decide what you will judge early and what you cannot judge yet. Lead quality against your written definition is readable in weeks. Whether those leads became revenue is not readable until a full sales cycle has passed. Confusing the two is how good channels get killed in week six.

Run one channel at a time if the budget is small. Two parallel tests on a small budget give you two unreadable results. The book's inner-ring logic is right here: sequencing beats coverage when you are short of money.

Who should read it, and who should skip it?

Read it if you have no channel yet, or one channel that arrived by accident and you have never asked whether it is the right one. The hour you spend on the outer ring will be the best hour of the quarter.

Skip it if you already have a channel producing qualified leads under your ceiling. The book will tempt you to diversify before you have exhausted what is working, and the second channel almost always costs more and teaches less than a founder expects. Go deeper instead: the creative testing loop is the work that gets more out of a channel you already have.