Justin Mares
Author of Traction
Where to hear and read about Justin Mares1 book
Justin Mares wrote Traction (2015), a playbook on how startups find customers across nineteen channels. The thing most worth learning from it is to test a few channels cheaply, then put everything behind the one that works.
If it is not on the path, don't do it!
Traction: How Any Startup Can Achieve Explosive Customer Growth
Justin Mares: the story and the lessons
1What Justin Mares's Traction argues
Mares wrote Traction because startups fail from too few customers. The book holds that nearly all failed startups built a product but lacked a big enough customer base. Traction means sales for a paid product or a growing user base for a free one.
The usual path is to build, launch, expect customers, then try familiar tactics such as Facebook ads, local PR and blog posts.
He says to spend half your time on traction from day one. Dropbox tested search ads while still building and found each customer cost $230 for a $99 product. It built a referral program instead, which became its biggest growth driver.
Marketo blogged about the problems it meant to solve before building and reached market with a pipeline of more than fourteen thousand interested buyers.
Early cold customers show where the product leaks. Cold customers have no tie to the founders, so they give more honest feedback than friends. Early spend leaks out of an unfinished product, and the leaks show where the holes are.
In fall 2008 Gabriel Weinberg posted DuckDuckGo to Hacker News. The product was not good enough to switch to, but the replies pointed him toward search privacy.
He names four traps that kill a wanted product. The traps are no business model, a market too small, customers too costly to reach, and a hypercompetitive market. Products also pass through three phases: make something people want, market it, then scale.
In phase I, send only a small flow of customers through using unscalable tactics, and scale only once product-market fit means customers stay.
Mares advises studying engaged bright spots before deciding to pivot. A few deeply engaged customers may hold promise that the core metrics hide. Ask whether they are early adopters in a huge market or outliers.
DuckDuckGo offered privacy from 2009, and growth exploded only after the 2013 NSA leaks. With no bright spots at all, it may be time to pivot.
2Bullseye: Traction's method for choosing a channel
Bullseye starts by writing a strategy for all nineteen channels. A channel is a broad category such as social ads. A strategy is a specific move inside it, such as running ads on reddit.
Keep every idea in a spreadsheet with educated guesses, so losing ideas can return in later rounds. The book says the framework grew out of talks with over forty founders.
He tells founders to audit their bias against unfamiliar channels. Some channels are out of sight, some disliked, and some skipped as schlep, such as business development and trade shows. Every one of the nineteen has given initial traction to both enterprise and consumer startups.
"Just because you hate talking on the phone doesn't mean your customers do." Jason Cohen bets competitors will refuse certain channels and prefers winning customers there.
Test about three channels for under $1,000 and a month each. These tests collect data, not growth. Each should show what a customer costs, how many are available, and whether they are the right customers now.
Mint's spreadsheet estimated traffic, click-through and conversion per strategy. Its cheap tests included a small newsletter sponsorship and some Google ads. A $250 AdWords test gives a rough read on search ads.
Set up tracking before spending anything on a test. A spreadsheet or cohort tool is enough, but it must exist before money goes out. Rank strategies by customers available, conversion rate, acquisition cost and lifetime value.
If tracking can't be automated, ask new customers how they found you.
Pour all effort into the winner and cut weaker channels. If search ads win clearly while trade shows and publicity work less well, the book says to spend everything on search ads. Other channels may feed the core one, but several separate strategies at once is the mistake.
Mint's blogs brought its first forty thousand customers. A rerun focused on publicity then took it to 1 million users within six months of launch. Bullseye is to traction what Lean is to product.
3Mares's rule for goals: one number and a Critical Path
Pick one traction goal that would change the company. Examples are one thousand paying customers, or enough growth to raise money or reach profit. Break it into dated, numeric subgoals and share one calendar with product work.
DuckDuckGo set goals in sequence: real switching, then 100 million searches a month for about break-even, then 1 percent of general search.
Check a channel's ceiling before winning it. DuckDuckGo ranked first for "new search engine," but the phrase drew about fifty searches a day when it needed about five thousand new visitors a day.
A tactic that brings real users can still be a blip against the target. Recheck as you grow.
The Critical Path is mainly a filter for what not to do. It is the fewest-step route to the goal, written as milestones. For 100 million searches a month, DuckDuckGo's path was a faster site, a stronger mobile offering and more broadcast TV coverage.
Users kept asking for images and auto-suggest, and the team left both off. The rule: "If it is not on the path, don't do it!"
Rebuild the path whenever the goal changes. For the 1 percent goal, images and auto-suggest became required milestones. Mainstream users forgive far less than loyal early ones.
The first milestone list is often wrong, so challenge every item and use market feedback to add or skip steps.
4How Traction wins first customers where attention already sits
Small blogs and link-sharing sites are the fastest route to first customers. Mint's prelaunch "I Want Mint" badge appeared on 600 blogs and brought fifty thousand signups. Its "Can I send you $500?" emails bought tracked ads on personal blogs and drew more than ten thousand preregistrations.
The badge traded VIP access for promotion and the new links lifted Mint.com in search. The surest way to find niche blogs is to ask target customers what they read.
Press climbs from small sites to big outlets, so pitch small first. Pitch only real milestones, keep pitches short with demo links, and pick an angle people will argue about. DonorsChoose.org went from local coverage to Newsweek to Oprah's favorite things, then won Gates Foundation sponsorship.
Gabriel Weinberg talked with a Time reporter on Twitter long before any pitch. That reporter later put DuckDuckGo on Time's top fifty sites of 2011.
Stunts and delight are uncrowded, so plan for flops. Half.com got a town renamed for a year at a cost of $70,000 and two hires. The stunt drew over 40 million impressions, and eBay bought the company for more than $300 million six months later.
Grasshopper's $65,000 chocolate grasshopper mailing raised visits over 3,000 percent, but its March Madness promotions and dancing videos flopped. Dollar Shave Club's $5,000 launch video was the main source of over 12,000 customers in two days.
Events and talks reach customers online ads miss. Twitter spent $11K on hallway screens at SXSW 2007 and daily tweets rose from twenty thousand to more than sixty thousand. Enservio ran its own summit because it could not reach top insurance executives any other way.
Speaking slots go to founders who ask organizers what they want and reuse one or two story talks. Dan Martell keeps two core talks, one for entrepreneurs and one for partners.
5Buying attention in small tracked bets
Search ads test positioning, and near breakeven in weeks is a signal. Archives.com's first campaign broke even within weeks, with no optimized pages. It later spent over $100,000 a month on search, and Ancestry.com bought it for $100 million.
Judge every ad by cost per acquisition, not clicks. At 15 cents per click, earning 13 cents loses money, while 20 cents gives about 25 percent margins.
Social ads should push shareable content, not sales pages. Airbrake put $15 of Twitter ads behind an interview with Stripe's CTO and got hundreds of retweets and tens of thousands of visits. CareOne's social leads were 217% more likely to make a first payment.
"Paid is fundamentally only as good as the content you put behind it."
Offline ads can't be predicted, so track each one. Give each ad its own URL, promo code or offer. A $7,000 DuckDuckGo billboard drew national press and doubled its users that month.
Smart Bear's Jason Cohen admitted "I have zero ability to predict what's going to work." Winners decayed after a quarter or a year.
Affiliate programs shift the risk of bad traffic to the affiliate. You define the conversion, so invalid leads earn nothing. Blogger Maneesh Sethi referred more than three thousand people to RescueTime.
Start on an existing network with a flat payout set below what a customer is worth, and recruit your own customers first.
6Assets that keep producing customers after the work is done
Long-tail SEO works when a page costs less than a customer is worth. Patrick McKenzie paid $3.50 for an owls of East Asia bingo page that earned about $60 over three years. Short fat-head terms make up about 30 percent of searches and the long tail about 70 percent.
Links are the hard part because they depend on outsiders, and bought links work briefly, then get penalized.
A blog needs six months and hand promotion before it pays. Unbounce started blogging the day it began building, and its first hire was a blogger. It had 20,000 monthly visitors and over 5,000 emails at its summer 2010 launch.
OkCupid wrote long data-driven posts that took a month each, and about a year later ranked first for "online dating."
Free tools and triggered email turn visitors into qualified leads. HubSpot's Marketing Grader, built in a few days, has been used by more than 3 million sites. Customer.io's founder email, sent thirty minutes after signup, gets about a 17 percent reply rate.
WP Engine trades a free speed report for an email, then sends a seven-email course and pitches near the end.
Measure the viral coefficient and fix the weakest step first. K is invites times conversion. Above 1 means exponential growth, and doubling either factor doubles K. myZamana grew from 100,000 users to over 4 million in under a year.
Even expert teams need one or two engineers for two to three months, and only one to three of every ten A/B tests are likely to win.
Communities need strict quality rules set early. Stack Overflow posted rules in its FAQ and users enforced them harder than Jeff Atwood would have. He called having no Meta his biggest mistake.
Alexis Ohanian emailed thanks to anyone who wrote about reddit and sent early users shirts and stickers.
7Borrowing other companies' customers
Business development needs a pipeline of fifty to one hundred partners. Most deals fail, so focus on ten to twenty chosen by revenue, reach or inventory, not brand name. Reach the person who owns the metric your deal moves, through a warm intro.
Delicious built a custom interface for The Washington Post, and a Mozilla promotion more than tripled its users. A pursued Wikipedia deal failed.
Hand-made deals must prove value before a low-touch API scales. Delicious opened its API only after the Post and Mozilla deals worked, then integrated with thousands of sites.
A public API alone attracts no one. Early deals create the demand.
Sales scales through a funnel built from the buyer's side. JBoss traded free documentation for contact details and got over ten thousand leads a month. Its inside team closed about 25 percent, against 7 to 10 percent industry averages.
Run calls in SPIN order and close with a deliverable, a timeline and a yes or no.
Platform bets pay when you fill a gap or arrive early. Evernote's Android app was featured for six weeks straight and brought hundreds of thousands of customers. Its Nokia, Windows and BlackBerry bets did not move the needle.
Alex Pachikov judges it worth it, because one win like Android covers several failures.
Trade shows pay only with set goals and meetings booked ahead. A $10,000 show turns profitable around the third sale of a $5,000 product. At RJMetrics only 28 percent of conversations came from walk-ups, and going outbound tripled or quadrupled the show's value.
Jason Cohen aimed for five meetings a day at Smart Bear and mailed discount cards redeemable only at his booth.
8Why Traction says every channel decays
Every strategy loses power as competitors crowd in. Early banner ads drew click-through rates over 75 percent, which plummeted once banners became common. Zynga grew on Facebook when it was uncrowded, and a gaming company today basically cannot repeat that.
Andrew Chen's line: "Over time, all marketing strategies result in shitty click-through rates."
Run at least one A/B test a week inside the core channel. That pace can make a channel two to three times more efficient. Testing has two jobs: tune the current strategy and find better ones in the same channel.
A blog-centered company can write guest posts elsewhere and buy Twitter and Facebook ads to amplify its best posts.
Each 10x jump in scale needed a fresh channel. DuckDuckGo went from roughly one hundred searches per day to more than ten million. It moved from content marketing to social and display ads, then publicity, and lately business development.
Tech news articles worked in phase I but later could not convert enough people to change its search numbers.
If repeated Bullseye rounds find nothing, fix the product. Saturation or rising costs signal it is time to rerun the process, drawing on data from earlier tests.
Several failed rounds point to a leaky product, not a missing channel.
Traction: How Any Startup Can Achieve Explosive Customer Growth
Startups grow by finding the one traction channel that works at their current stage, using a three-step Bullseye process: brainstorm all nineteen channels, test a few cheaply, then focus on the winner. It is worth reading for its concrete playbooks and founder cases for every channel, from SEO and viral loops to trade shows and business development.
Mares wrote Traction because startups fail from too few customers. The book holds that nearly all failed startups built a product but lacked a big enough customer base. Traction means sales for a paid product or a growing user base for a free one.
The usual path is to build, launch, expect customers, then try familiar tactics such as Facebook ads, local PR and blog posts.
He says to spend half your time on traction from day one. Dropbox tested search ads while still building and found each customer cost $230 for a $99 product. It built a referral program instead, which became its biggest growth driver.
Marketo blogged about the problems it meant to solve before building and reached market with a pipeline of more than fourteen thousand interested buyers.
Early cold customers show where the product leaks. Cold customers have no tie to the founders, so they give more honest feedback than friends. Early spend leaks out of an unfinished product, and the leaks show where the holes are.
In fall 2008 Gabriel Weinberg posted DuckDuckGo to Hacker News. The product was not good enough to switch to, but the replies pointed him toward search privacy.
He names four traps that kill a wanted product. The traps are no business model, a market too small, customers too costly to reach, and a hypercompetitive market. Products also pass through three phases: make something people want, market it, then scale.
In phase I, send only a small flow of customers through using unscalable tactics, and scale only once product-market fit means customers stay.
Mares advises studying engaged bright spots before deciding to pivot. A few deeply engaged customers may hold promise that the core metrics hide. Ask whether they are early adopters in a huge market or outliers.
DuckDuckGo offered privacy from 2009, and growth exploded only after the 2013 NSA leaks. With no bright spots at all, it may be time to pivot.
Bullseye starts by writing a strategy for all nineteen channels. A channel is a broad category such as social ads. A strategy is a specific move inside it, such as running ads on reddit.
Keep every idea in a spreadsheet with educated guesses, so losing ideas can return in later rounds. The book says the framework grew out of talks with over forty founders.
He tells founders to audit their bias against unfamiliar channels. Some channels are out of sight, some disliked, and some skipped as schlep, such as business development and trade shows. Every one of the nineteen has given initial traction to both enterprise and consumer startups.
"Just because you hate talking on the phone doesn't mean your customers do." Jason Cohen bets competitors will refuse certain channels and prefers winning customers there.
Test about three channels for under $1,000 and a month each. These tests collect data, not growth. Each should show what a customer costs, how many are available, and whether they are the right customers now.
Mint's spreadsheet estimated traffic, click-through and conversion per strategy. Its cheap tests included a small newsletter sponsorship and some Google ads. A $250 AdWords test gives a rough read on search ads.
Set up tracking before spending anything on a test. A spreadsheet or cohort tool is enough, but it must exist before money goes out. Rank strategies by customers available, conversion rate, acquisition cost and lifetime value.
If tracking can't be automated, ask new customers how they found you.
Pour all effort into the winner and cut weaker channels. If search ads win clearly while trade shows and publicity work less well, the book says to spend everything on search ads. Other channels may feed the core one, but several separate strategies at once is the mistake.
Mint's blogs brought its first forty thousand customers. A rerun focused on publicity then took it to 1 million users within six months of launch. Bullseye is to traction what Lean is to product.
Pick one traction goal that would change the company. Examples are one thousand paying customers, or enough growth to raise money or reach profit. Break it into dated, numeric subgoals and share one calendar with product work.
DuckDuckGo set goals in sequence: real switching, then 100 million searches a month for about break-even, then 1 percent of general search.
Check a channel's ceiling before winning it. DuckDuckGo ranked first for "new search engine," but the phrase drew about fifty searches a day when it needed about five thousand new visitors a day.
A tactic that brings real users can still be a blip against the target. Recheck as you grow.
The Critical Path is mainly a filter for what not to do. It is the fewest-step route to the goal, written as milestones. For 100 million searches a month, DuckDuckGo's path was a faster site, a stronger mobile offering and more broadcast TV coverage.
Users kept asking for images and auto-suggest, and the team left both off. The rule: "If it is not on the path, don't do it!"
Rebuild the path whenever the goal changes. For the 1 percent goal, images and auto-suggest became required milestones. Mainstream users forgive far less than loyal early ones.
The first milestone list is often wrong, so challenge every item and use market feedback to add or skip steps.
Small blogs and link-sharing sites are the fastest route to first customers. Mint's prelaunch "I Want Mint" badge appeared on 600 blogs and brought fifty thousand signups. Its "Can I send you $500?" emails bought tracked ads on personal blogs and drew more than ten thousand preregistrations.
The badge traded VIP access for promotion and the new links lifted Mint.com in search. The surest way to find niche blogs is to ask target customers what they read.
Press climbs from small sites to big outlets, so pitch small first. Pitch only real milestones, keep pitches short with demo links, and pick an angle people will argue about. DonorsChoose.org went from local coverage to Newsweek to Oprah's favorite things, then won Gates Foundation sponsorship.
Gabriel Weinberg talked with a Time reporter on Twitter long before any pitch. That reporter later put DuckDuckGo on Time's top fifty sites of 2011.
Stunts and delight are uncrowded, so plan for flops. Half.com got a town renamed for a year at a cost of $70,000 and two hires. The stunt drew over 40 million impressions, and eBay bought the company for more than $300 million six months later.
Grasshopper's $65,000 chocolate grasshopper mailing raised visits over 3,000 percent, but its March Madness promotions and dancing videos flopped. Dollar Shave Club's $5,000 launch video was the main source of over 12,000 customers in two days.
Events and talks reach customers online ads miss. Twitter spent $11K on hallway screens at SXSW 2007 and daily tweets rose from twenty thousand to more than sixty thousand. Enservio ran its own summit because it could not reach top insurance executives any other way.
Speaking slots go to founders who ask organizers what they want and reuse one or two story talks. Dan Martell keeps two core talks, one for entrepreneurs and one for partners.
Search ads test positioning, and near breakeven in weeks is a signal. Archives.com's first campaign broke even within weeks, with no optimized pages. It later spent over $100,000 a month on search, and Ancestry.com bought it for $100 million.
Judge every ad by cost per acquisition, not clicks. At 15 cents per click, earning 13 cents loses money, while 20 cents gives about 25 percent margins.
Social ads should push shareable content, not sales pages. Airbrake put $15 of Twitter ads behind an interview with Stripe's CTO and got hundreds of retweets and tens of thousands of visits. CareOne's social leads were 217% more likely to make a first payment.
"Paid is fundamentally only as good as the content you put behind it."
Offline ads can't be predicted, so track each one. Give each ad its own URL, promo code or offer. A $7,000 DuckDuckGo billboard drew national press and doubled its users that month.
Smart Bear's Jason Cohen admitted "I have zero ability to predict what's going to work." Winners decayed after a quarter or a year.
Affiliate programs shift the risk of bad traffic to the affiliate. You define the conversion, so invalid leads earn nothing. Blogger Maneesh Sethi referred more than three thousand people to RescueTime.
Start on an existing network with a flat payout set below what a customer is worth, and recruit your own customers first.
Long-tail SEO works when a page costs less than a customer is worth. Patrick McKenzie paid $3.50 for an owls of East Asia bingo page that earned about $60 over three years. Short fat-head terms make up about 30 percent of searches and the long tail about 70 percent.
Links are the hard part because they depend on outsiders, and bought links work briefly, then get penalized.
A blog needs six months and hand promotion before it pays. Unbounce started blogging the day it began building, and its first hire was a blogger. It had 20,000 monthly visitors and over 5,000 emails at its summer 2010 launch.
OkCupid wrote long data-driven posts that took a month each, and about a year later ranked first for "online dating."
Free tools and triggered email turn visitors into qualified leads. HubSpot's Marketing Grader, built in a few days, has been used by more than 3 million sites. Customer.io's founder email, sent thirty minutes after signup, gets about a 17 percent reply rate.
WP Engine trades a free speed report for an email, then sends a seven-email course and pitches near the end.
Measure the viral coefficient and fix the weakest step first. K is invites times conversion. Above 1 means exponential growth, and doubling either factor doubles K. myZamana grew from 100,000 users to over 4 million in under a year.
Even expert teams need one or two engineers for two to three months, and only one to three of every ten A/B tests are likely to win.
Communities need strict quality rules set early. Stack Overflow posted rules in its FAQ and users enforced them harder than Jeff Atwood would have. He called having no Meta his biggest mistake.
Alexis Ohanian emailed thanks to anyone who wrote about reddit and sent early users shirts and stickers.
Business development needs a pipeline of fifty to one hundred partners. Most deals fail, so focus on ten to twenty chosen by revenue, reach or inventory, not brand name. Reach the person who owns the metric your deal moves, through a warm intro.
Delicious built a custom interface for The Washington Post, and a Mozilla promotion more than tripled its users. A pursued Wikipedia deal failed.
Hand-made deals must prove value before a low-touch API scales. Delicious opened its API only after the Post and Mozilla deals worked, then integrated with thousands of sites.
A public API alone attracts no one. Early deals create the demand.
Sales scales through a funnel built from the buyer's side. JBoss traded free documentation for contact details and got over ten thousand leads a month. Its inside team closed about 25 percent, against 7 to 10 percent industry averages.
Run calls in SPIN order and close with a deliverable, a timeline and a yes or no.
Platform bets pay when you fill a gap or arrive early. Evernote's Android app was featured for six weeks straight and brought hundreds of thousands of customers. Its Nokia, Windows and BlackBerry bets did not move the needle.
Alex Pachikov judges it worth it, because one win like Android covers several failures.
Trade shows pay only with set goals and meetings booked ahead. A $10,000 show turns profitable around the third sale of a $5,000 product. At RJMetrics only 28 percent of conversations came from walk-ups, and going outbound tripled or quadrupled the show's value.
Jason Cohen aimed for five meetings a day at Smart Bear and mailed discount cards redeemable only at his booth.
Every strategy loses power as competitors crowd in. Early banner ads drew click-through rates over 75 percent, which plummeted once banners became common. Zynga grew on Facebook when it was uncrowded, and a gaming company today basically cannot repeat that.
Andrew Chen's line: "Over time, all marketing strategies result in shitty click-through rates."
Run at least one A/B test a week inside the core channel. That pace can make a channel two to three times more efficient. Testing has two jobs: tune the current strategy and find better ones in the same channel.
A blog-centered company can write guest posts elsewhere and buy Twitter and Facebook ads to amplify its best posts.
Each 10x jump in scale needed a fresh channel. DuckDuckGo went from roughly one hundred searches per day to more than ten million. It moved from content marketing to social and display ads, then publicity, and lately business development.
Tech news articles worked in phase I but later could not convert enough people to change its search numbers.
If repeated Bullseye rounds find nothing, fix the product. Saturation or rising costs signal it is time to rerun the process, drawing on data from earlier tests.
Several failed rounds point to a leaky product, not a missing channel.