Growth hacking is a term coined by Sean Ellis in 2010 to describe a mindset and approach to growth that prioritizes rapid experimentation across the full customer journey — acquisition, activation, retention, revenue, and referral — rather than relying on a fixed set of traditional marketing channels. The “hack” in growth hacking refers to finding unconventional, scalable paths to growth by identifying leverage points that traditional marketing overlooks or underinvests in. The term has been stretched and misused to the point of near-meaninglessness, but the underlying concept is genuinely useful: systematic experimentation on growth levers, with a bias toward measuring results and doubling down on what works.
The original context for growth hacking was early-stage B2C technology companies where traditional marketing was either too expensive or too slow for the growth demands they faced. Dropbox’s “refer a friend, get more storage” program, Hotmail’s “Get your free email at Hotmail” signature appended to every outbound email, and Airbnb’s integration with Craigslist to post listings cross-platform are the canonical growth hacking examples. Each of these was a product-level lever that produced viral or network-effect growth at a cost that paid media could not replicate.
The Growth Hacking Framework: AARRR
Dave McClure’s AARRR framework (Acquisition, Activation, Retention, Revenue, Referral) organizes the growth levers by stage in the customer lifecycle. Traditional marketing tends to focus heavily on Acquisition — driving traffic and leads. Growth hacking looks at all five stages and asks where the highest-leverage improvement opportunity exists. A company with strong acquisition but low activation (many users sign up but few reach the aha moment) will not grow through more acquisition marketing; fixing activation is the highest-leverage growth lever at that stage.
The diagnostic question for each AARRR stage is: what would a 10% improvement here do to the overall growth rate? If conversion from acquisition to activation is 20%, improving it to 22% effectively increases the output of all acquisition spending by 10% without spending another dollar on acquisition. If retention at 6 months is 60% and improving it to 66% retains enough revenue to fund an additional acquisition channel, then retention is the lever worth optimizing. Growth hacking as a practice identifies these leverage points through data analysis and experiments against them.
Viral Growth and Referral Loops
Viral growth — where existing users refer new users, who refer more new users — is the growth mechanic most associated with growth hacking, because it produces acquisition at effectively zero marginal cost per referred user. The viral coefficient measures how many new users each existing user generates: a viral coefficient above 1 means the product is growing on its own through referrals alone; below 1 means the referral loop supplements but does not replace paid or organic acquisition.
Building a referral loop requires that the product has a natural sharing mechanism (inviting a teammate to a collaboration tool, recommending a useful tool to a colleague), an incentive that makes sharing worthwhile (Dropbox’s extra storage, Uber’s ride credit), and a friction-free referral path (a unique link that can be shared in one tap). Most products do not have natural viral coefficients above 1; the growth hacking work is designing incentive and friction reduction to maximize referral rate from the fraction of users who are inclined to share.
Growth Hacking and Attribution
Growth hacking experiments require measurement to distinguish from guessing — and measurement in a growth context means attributing outcomes to specific experiments. When Dropbox tested its referral program, the growth team tracked referral signups separately from organic signups to measure the incremental user acquisition generated by the referral mechanic. When a growth team tests a new onboarding flow, they measure activation rate before and after the change, ideally via A/B test, to isolate the experiment’s effect from external factors.
The attribution challenge for growth experiments is cleaner than the multi-touch channel attribution problem in marketing: experiments are typically designed to produce measurable outcomes on specific metrics in a defined time window, and the experimental design (control vs treatment group) provides the counterfactual. What is harder is attribution across the entire AARRR funnel: a growth experiment that improves activation rate produces downstream effects on retention, revenue, and referral that take months to fully materialize. Teams that measure growth hacking experiments by short-term conversion metrics and ignore downstream effects may declare experiments successful on leading indicators that do not translate to lasting revenue growth.