Traction
Gabriel Weinberg & Justin Mares
Traction is Gabriel Weinberg and Justin Mares's field guide to getting customers, built on one finding: almost every failed startup had a product, and what it lacked was a repeatable way to reach people. The book catalogues nineteen traction channels and gives a method, Bullseye, for testing three cheaply and committing to the one that works, while spending half your time on traction from day one.
You are building a product because building feels like progress, and marketing is something you will get to when the product is ready. Traction's authors interviewed the founders of dozens of companies and found that this instinct is how most startups die. The product gets finished; the customers never arrive, because nobody spent as much finding them as building for them.
The book in essence
The book's first rule is the fifty percent rule: from the beginning, half the founder's time goes to product and half to traction, because traction work is what tells you whether the product is worth building, and because channels take longer to find than anyone expects. The second contribution is the catalogue: nineteen channels, from viral marketing, search ads, SEO, content and email to the ones founders forget, engineering as marketing, business development, sales, existing platforms, trade shows, offline events, speaking, community. Which one works for a given company is close to unpredictable in advance, which is why guessing fails. The third is Bullseye, the method that replaces the guess: brainstorm all nineteen, rank them into rings, test the three most promising cheaply and in customers rather than traffic, then concentrate everything on the winner until it saturates.
Weinberg founded DuckDuckGo and wrote the book with Mares from interviews with the founders of Wikipedia, Reddit, Evernote, Mint and others. Each channel chapter is a case study plus the practical mechanics, which keeps the book useful long after the specific tactics have aged.
Who it's for
- Founders whose product is nearly done and whose marketing plan is a vague intention
- Teams that have tried three channels at a quarter effort each and concluded that marketing does not work for them
- Anyone who has never seen the full list of nineteen channels and is about to default to social media
Who it's not for
- — Readers who want depth on a single channel such as SEO or paid search; each gets a chapter, not a book
- — Companies past product-market fit with a channel already saturating; the book is about finding the first one
Key lessons
01
Half your time on traction, from day one
The fifty percent rule feels wrong to a founder with an unfinished product, and it is the book's most important idea. Traction work early tells you whether anyone wants the product before you finish it, teaches you what to build by putting you in front of customers, and starts the slow work of finding a channel while there is still money to test with.
A founder in Izmir selling accounting software to freelancers had built for eight months and marketed for none. Under the fifty percent rule she gave every afternoon to traction, and the channel that worked, a free plug-in for the invoicing tool freelancers already used, brought more customers in a month than the previous eight.
02
There are nineteen channels, not three
Most founders can name a handful of channels and default to the one they know. The catalogue exists to break that habit: nineteen ways to reach customers, each with its own economics, including the unfashionable ones like trade shows, speaking and business development that often work precisely because competitors ignore them. Naming them turns marketing from a mood into a set of experiments.
A B2B startup in Istanbul assumed Instagram was its channel and spent six months on content that produced two dental clinics. Listing all nineteen put trade shows in the middle ring. One stand at a dental exhibition in Ankara produced forty demos and nine paying clinics in a week.
03
Run Bullseye instead of guessing
Brainstorm every channel with a sentence on how it could work for you, pick the three most promising, and test each cheaply, measured in customers acquired and cost per customer rather than in traffic. One usually stands out. That channel gets the whole effort until it stops yielding, and only then does the process restart. The discipline is concentration: one channel at full effort teaches more in a month than six at a quarter each in a year.
A consumer app in Ankara tested three channels for three weeks each: influencer posts, search ads and a referral loop. Only the referral loop produced customers who stayed. The founders cancelled the other two and spent the next year making the loop better.
04
Measure channels in customers, not attention
Impressions, followers and traffic are the vanity metrics of marketing. A channel test is judged by how many customers it acquired, what each cost, and whether they stayed. A channel that produces attention and no customers is a channel that does not work, however good the numbers look on a screen.
A Turkish e-commerce startup's PR push earned coverage in three national outlets and a spike of forty thousand visits. Eleven of them bought. The founders logged the channel as a failure at that cost and moved on, which the coverage alone would never have told them.
Try this today
Write all nineteen traction channels on one page with one honest sentence each on how it could work for your product, circle three, and design a one-week test for each measured in customers and cost per customer.
Quick check
What is the fifty percent rule?
How does Bullseye choose a channel?
Selected quotes
“Almost every failed startup has a product. What failed startups don't have is traction.”
“Traction is basically quantitative evidence of customer demand.”