Growth is the discipline of getting a product in front of the people who need it, repeatedly and without going broke doing it. It's not one tactic — it's a portfolio of channels, loops, and experiments, each with its own economics, and the job is picking the ones that fit your product and running them like science instead of superstition. Use this guide to understand the vocabulary before you spend a dollar or a week chasing a "growth hack" — then go run the missions to practice the choosing.
Acquisition Channel
An acquisition channel is a repeatable path new customers take to find you — search, ads, referrals, sales, partnerships — and every channel has its own cost, volume, and speed. Gabriel Weinberg and Justin Mares, in Traction, catalog nineteen of these and make a blunt observation: most founders only ever consider the channels they already know, or the ones "people like us" are supposed to use, which means the best channel for your specific business is often one nobody on your team has tried. Their Bullseye framework fixes this by forcing you through three rounds — brainstorm all nineteen channels for your product, run cheap parallel tests on the handful that seem most promising, then go all-in on the one that's actually working.
Dropbox is the canonical cautionary tale: they started with search engine marketing, discovered they were paying $230 to acquire a customer for a $99 product, and only found their real channel (referrals) after that expensive detour.
Watch out: picking a channel because a competitor uses it, or because it's the one your
team feels comfortable with, isn't a strategy — it's a guess wearing a strategy costume.
→ practice this in the acquisition-channels mission.
Channel Fit
Channel fit is the match between a channel's economics and your product's price point and audience — a $30/month product cannot afford a channel that costs $3,000 to acquire a customer. This is where acquisition-channel enthusiasm meets unit-economics reality: a channel isn't "good" or "bad" in the abstract, it's only good for your numbers. Paid social might be a phenomenal channel for a $2,000 enterprise contract and a terrible one for a $9 consumer app, because the same click costs the same dollar amount regardless of what you're selling.
Worked example: if a channel's blended cost per acquired customer is $150, and your product's gross margin per customer over its lifetime is $120, the channel doesn't pencil out no matter how much volume it can deliver — you'd be paying to lose money faster.
Watch out: "it's working" usually means "it's bringing in users," not "it's profitable." Check
the cost against what a customer is actually worth before you scale spend.
→ cross-check with the cac and ltv-cac-ratio calculators, and the channel-fit mission.
Growth Loop and Flywheel
A growth loop is a closed cycle where the output of one cohort of users — invites sent, content published, revenue earned — becomes the input that acquires the next cohort, so growth compounds instead of resetting to zero every day. This is the structural upgrade from a funnel, which is linear: you pour traffic in the top, some comes out the bottom as customers, and tomorrow you have to pour in fresh traffic again from scratch. A loop, once built, keeps turning on its own — Andrew Chen's The Cold Start Problem frames this as the difference between a marketing campaign (temporary) and a network effect (permanent).
A flywheel is the broader version of the same idea applied to the whole business: a self-reinforcing system where each part feeds the next — more customers fund more product value, which attracts more customers — and gains momentum as it turns, exactly the way a heavy wheel gets easier to spin once it's moving. A flywheel is usually made of several smaller loops (an acquisition loop, an engagement loop, a monetization loop) stacked together.
Watch out: most things people call a "growth loop" are actually a funnel with a referral step
bolted on. A real loop has to feed itself — check whether cohort N+1 is caused by cohort N, not
just correlated with it.
→ practice this in the growth-loops mission.
Virality and Viral Coefficient
Virality is growth where your existing users expose new users to the product as a byproduct or incentive of using it — not as a separate marketing effort bolted on top. Traction breaks viral loops into three repeating steps: a user experiences value, that experience prompts them to invite or expose others, and those others convert into new users who repeat the cycle. The mechanism varies — inherent virality (Google Docs is more useful the moment you invite a collaborator), incentivized virality (Dropbox gives both parties free storage for a referral), and broadcast virality (a Spotify listen shows up in a friend's feed automatically).
The viral coefficient (K) is the average number of new users each existing user brings in:
K = invites sent per user × conversion rate per invite. If each user sends 3 invites and 20% of
those convert, K = 0.6 — meaningful growth, but not self-sustaining. Above K = 1.0, each user
replaces themselves and then some, and the product grows on its own without additional spend. The
second variable that matters just as much is viral cycle time: the same K compounds far faster
if the loop closes in a day than if it takes a month, which is why shortening signup friction is
often a bigger lever than raising K itself.
Watch out: K above 1.0 sounds like a magic number, but true sustained virality (K > 1 forever)
is rare — Traction notes K above 0.5 is already a meaningful growth contributor even without
crossing 1.0.
→ run the numbers in the viral-coefficient calculator and the virality-mechanics mission.
Referral Program and Word of Mouth
Word of mouth is unpaid, unprompted recommendation between customers — the cheapest channel there is, and the hardest one to fake, because it's earned entirely by a product people actually want to talk about. A referral program is the designed, incentivized version of this: a structured mechanism (extra storage, account credit, cash) that turns satisfaction into a measurable, trackable acquisition channel instead of leaving it to chance. Dropbox's give-space, get-space referral loop became its single biggest growth driver, generating tens of millions of signups — but it worked because the underlying product already had people who wanted to recommend it.
That ordering matters: a referral program amplifies advocacy that already exists, it doesn't manufacture advocacy that doesn't. If your product isn't generating unprompted word of mouth, bolting a referral widget onto it will mostly incentivize people to spam their contact list once for a reward and never again.
Watch out: treating a referral program as a fix for weak organic word of mouth is backwards —
diagnose why people aren't already talking about you before you pay them to.
→ practice this in the referral-programs mission.
Paid Acquisition
Paid acquisition means buying attention through ads — search, social, display — which is fast to start and fast to scale in a way organic channels never are, since you can turn spend up or down and see the effect within days. The catch is that it's only sustainable while a customer's lifetime value clears what the auction charges to reach them, and ad auctions are competitive by design: if a channel is working brilliantly for you, other advertisers notice and bid the price up until it stops working brilliantly.
Worked example: if you're paying $40 CAC through paid search and your gross margin per customer is $150 over their lifetime, you have real room to scale — but if a competitor with deeper pockets enters that same auction, your $40 can become $70 within a quarter with no change on your end.
Watch out: paid acquisition is "rented," not owned — the moment you stop paying, the channel
stops producing, unlike SEO or word of mouth which keep compounding after the initial investment.
→ check the math with the cac calculator and the paid-acquisition mission.
Content Marketing and SEO
Content marketing is publishing useful material that attracts your segment while they're researching their problem — guides, tools, comparisons — and it's slow to compound but cheap to sustain, and unlike paid acquisition, you own it: a blog post published two years ago can still bring in customers today with zero incremental spend. Traction points out content marketing's secondary effect is often as valuable as the direct traffic: a strong company blog opens doors to partnerships, press, and business development that a cold outreach email never would.
SEO is the specific practice of making your pages the best available answer to what your customers are already typing into a search engine, so it sends them to you for free — it's best understood as a part of content marketing, the distribution layer that makes the content findable rather than a separate channel. It rewards patience: ranking for a competitive term can take six months to two years, which is exactly why it's worth starting before you need the traffic, not after.
Watch out: content marketing measured week-to-week looks like it's failing, because its payoff
curve is compounding, not linear — judge it over quarters, not weeks.
→ practice this in the content-engine mission.
Partnerships
Partnerships mean borrowing another company's audience or distribution in exchange for value you provide them — an integration, co-marketing, or a revenue share. Traction frames business development as sales' quieter cousin: sales exchanges dollars for a product, business development exchanges value for reach. Google's early growth leaned heavily on exactly this — default-search deals with Netscape and Yahoo! delivered more distribution than any ad campaign could have at the time.
A partnership also carries a second, less obvious payoff: trust transfer. When a respected partner integrates with you or recommends you, some of their credibility rubs off on you before a prospective customer has evaluated you at all — which is why an integration listed on a bigger platform's marketplace often converts better than the same feature marketed cold.
Watch out: partnership deals take real strategic clarity to work — know the metric you need
the partnership to move before you start pitching, or you'll sign deals that feel good and change
nothing.
→ practice this in the partnerships mission.
Product-Led Growth
Product-led growth (PLG) is a go-to-market motion where the product itself — free tiers, self-serve onboarding, in-product sharing — does the acquiring, converting, and expanding, instead of a sales team doing it manually. Wes Bush, in Product-Led Growth, frames the central decision as choosing your weapon: a free trial (full or partial product, time-limited) suits products whose value is obvious fast, a freemium model (permanently free tier, no time limit) suits products with a very large addressable market since giving the product away only pays off at scale, and a demo still suits complex, high-touch sales motions where self-serve would undersell the product's actual value.
PLG success depends on two things happening well before a prospect ever talks to a salesperson: onboarding that gets them to first value fast, and activation that gets them using the core feature, not just poking around the settings page. Bush's rule of thumb: freemium generally needs a market on the order of tens of millions of potential users to work, because conversion rates from free to paid are low by design — a niche product with fifty potential customers has no business giving itself away.
Watch out: PLG isn't "remove the sales team" — it's "let the product qualify and convert
before sales ever gets involved," and most successful PLG companies still layer sales on top for
their biggest accounts.
→ practice this in the plg-basics mission.
Network Effects
Network effects occur when each new user makes the product more valuable to every other user — the dynamic behind marketplaces, social products, and workplace tools. Andrew Chen's The Cold Start Problem insists the term gets thrown around as a buzzword far more often than it's actually built with intention, and offers a sharper starting unit: the atomic network, the smallest possible group of users who are stable and self-sustaining on their own. For Zoom, that's two people on a call. For Slack, it's roughly three coworkers. For Airbnb, it takes hundreds of active listings in a single market before the network stops feeling empty and starts feeling reliable.
The strategic implication is sequencing: you don't launch a network effect broadly, you win one atomic network completely, then use it as a base to tip the next one — which is exactly how Tinder expanded campus by campus rather than launching nationally on day one.
Watch out: a sub-scale network is fragile, not just small — users who show up and find nobody
they know will leave and not come back, so "grow slowly everywhere" is often worse than "dominate
one small network first."
→ network effects are one of the strongest forms of moat; see the growth-loops mission for the
mechanics of building one.
Growth Experiment
A growth experiment is a small, time-boxed test of one growth idea against one metric, run from a prioritized backlog instead of chasing whatever tactic is loudest that week. Lean Analytics calls the disciplined version of this growth hacking — not a synonym for "clever trick," but a specific process: find a metric measurable early in a user's lifecycle, confirm it's correlated with a business outcome you care about, predict the outcome from where the metric stands today, then change the product to move the metric and watch whether the outcome follows.
The examples are instructive because they're all leading indicators, not vanity numbers: Facebook's growth team found that a user who added seven friends within ten days became durably engaged; Twitter tracked when a new follow got followed back; Dropbox tracked one file saved in one folder. None of these are big swings — they're small, specific, testable claims about what predicts retention, tested fast.
Watch out: a backlog of experiments beats a single "big bet" — prioritize by expected impact
and cost of testing, and treat a failed experiment as data, not a wasted sprint.
→ practice this in the growth-experiment-backlog mission, and use the conversion-funnel
calculator to see where an experiment would move the needle most.
Go deeper
- Gabriel Weinberg & Justin Mares, Traction — the Bullseye framework and all nineteen traction channels in depth, with the viral-coefficient math worked through in detail.
- Andrew Chen, The Cold Start Problem — atomic networks, tipping points, and how network effects actually get built (and how they die).
- Nir Eyal, Hooked — the Hook Model (trigger, action, variable reward, investment) for the product-side habit loops that make retention-driven growth loops possible.
- Wes Bush, Product-Led Growth — the free-trial vs. freemium vs. demo decision framework and how to build a product that sells itself.
- Alistair Croll & Benjamin Yoskovitz, Lean Analytics — growth hacking as a disciplined process, and how to find a genuine leading indicator instead of a vanity metric.
- Rob Walling, Start Small, Stay Small — bootstrapped, low-budget takes on SEO, partnerships, and word of mouth for founders without a growth team.