Yes, with one precondition everyone skips
Growth hacking still works, but only after the precondition everyone skips: once your product is habit-forming, then and only then should you start thinking about growth hacks. A hack pours users into whatever retention you already have. Habit-forming products compound the pour. Leaky ones convert it to churn and a spent idea. The graveyard of growth hacking is not failed hacks, it is successful hacks on products that could not hold the users.
I say this as a reformed extremist. I once made $50,000 a month spamming Google before the whole edifice was deindexed in one update, which is the compressed lesson of the discipline: hacks that fight the platform die by the platform. The hacks that survive exploit an overlooked legitimate mechanic, not a policy gap.
What a real hack looks like now
Where do durable hacks come from? My sourcing rule: 70 percent from reverse-engineering competitors, 30 percent from past experience. A current favourite of the first kind: when a competitor spends heavily advertising their brand but forgets to own their own review queries, ranking for those alternative-and-review searches captures their paid demand at the moment of maximum doubt, a mechanic that now pays twice because AI assistants lean on review content when they recommend. Non-obvious channels count too: the best sales hack I ever saw was hiring 50 relentlessly hungry salespeople from Serbia and Turkey, a labour-market arbitrage no tool could find.
The meta-hack is process. At Superpower we ran a $250 pitch challenge, paying anyone to present a growth tactic, because paying for ideas outperforms hoping for them. But every idea still enters the ranked experiment queue and lives or dies by the number, which is what separates a hack from a stunt. Growth hacking is one move inside growth marketing, not a substitute for strategy, and the moment a hack needs fake accounts or platform deception, walk away, that ending is already written.
And keep the definition honest inside your team. The moment growth hacking becomes the label for ordinary work, discounting, a paid campaign, a redesign, the queue fills with theatre wearing a clever name. Reserve the word for mechanics with genuine asymmetry, where a small input buys an outsized, measurable return against the number the strategy owns.
FAQ
Does growth hacking still work in 2026?
Yes, on habit-forming products, sourced mostly from reverse-engineering competitors, and run through a measured experiment queue. What no longer works is platform deception, which dies at update cadence, and hacks poured into products that cannot retain users.
What is an example of a growth hack that works?
Ranking for a competitor's review and alternatives queries while they spend heavily on brand ads: you capture their paid demand at the moment of doubt, and the same content is what AI assistants cite when asked for recommendations.
When should a startup start growth hacking?
Only once the product is habit-forming. Before retention exists, hacks convert to churn and spent ideas. Fix activation and habit first, then pour.