The original growth hacker reveals his secrets | Sean Ellis (author of “Hacking Growth”)

Sean Ellis says fix activation before buying growth

Sean Ellis argues that growth starts with proving a product is a must-have, and only then with scaling it. His survey measures that early, and fixing onboarding and activation comes before spending on acquisition. Companies like Lookout and LogMeIn show that these fixes multiply the return on channels that already existed.

Takeaways

• The very-disappointed survey measures product-market fit early. Ellis asks users how they would feel without the product. Around 40% answering very disappointed signals fit. The survey leads and retention lags. The 40% mark works more as a focusing target than a firm cutoff, and Nubank reportedly uses 50%.
• Onboarding to the right experience beats retention tactics. Lookout's score went from 7% to 40% in two weeks after it repositioned around antivirus and streamlined onboarding, then hit 60% six months later. Teams should ignore somewhat disappointed users and focus on must-have users.
• Fix activation first, acquisition last. Ellis sequences growth as activation, then engagement and referral, then the revenue model, then acquisition channels. He says most companies start with acquisition too early and pour money into a leaky product.
• LogMeIn's roadmap freeze unlocked scale. Most sign-ups never used the product, so a CEO froze the roadmap for three months. Sign-up-to-usage rose 1000%, and the same channels then scaled from $10,000 to a million dollars a month.
• Asking non-converters why reveals cheap wins. In a LogMeIn demand test, people who skipped the download doubted the free product was real. Showing a paid trial beside the free option lifted downloads by 300%. Blending qualitative and quantitative research produces better tests.
• North Star metrics should start from must-have value. Ellis derives the metric from the value the survey uncovers, with revenue as a byproduct. His examples are Airbnb nights booked, Amazon monthly purchases, Eventbrite weekly tickets and Dropbox files accessed.
• Channel choice depends on the product. Search did not work at Dropbox because nobody was looking for it, while Bounce needed both SEO and street signage. Dropbox had strong word-of-mouth before its intern-built referral program.

Running the survey well

• Sample users who have used the product two or more times and recently, drawn at random.
• Tools such as SurveyMonkey with emailed surveys, Qualaroo or Intercom all work.
• Follow up by asking why the primary benefit matters, and use multiple-choice benefit statements.
• Culture affects how optimistic answers are, so read scores with that in mind.
• Superhuman won over somewhat disappointed users without diluting the product for must-have users.