The situation
Demand was not the problem. The platform had enormous self-serve top-of-funnel and a genuinely useful research tool underneath it. But a large share of SMB customers signed up with intent and left before getting anything out of it.
The product worked well for experienced SEO practitioners. It did not work for people who lacked the strategic maturity to turn keyword research into a business result — and that was most of the paying base.
What the real problem was
Customers were choosing keywords based on aspiration rather than competitive reality. They misread domain authority and backlink strength, targeted the wrong geography, and cancelled before the compounding benefit of content had time to appear.
Every exit survey said the product was missing something. It was not. It was failing to walk anyone to a first win. That is a design problem, and it shows up on the income statement as churn.
What I changed
- Redesigned onboarding from a static setup flow into a guided, visual value exchange — the product asked for less and gave back something usable immediately.
- Built site scraping and brand inference into onboarding so the product started from the customer’s actual business context instead of a blank field.
- Improved keyword selection with competitor context, keyword-difficulty guidance, and geolocation targeting for users who could not self-assess winnability.
- Embedded playbooks for lower-maturity users who needed strategy, not more data.
- Partnered early with a generative AI provider to build structured long-tail content workflows — moving the product from telling customers what to do into helping them do it.
- Added reminders, publishing cadence support, and recrawl-based feedback loops so progress became visible before results arrived.
- Introduced higher-value packaging and laid the groundwork for usage-oriented monetisation as AI compute costs rose.
- Extended acquisition and activation through a contextual Chrome extension and adjacent brand intelligence.
What happened
Voluntary churn more than halved. Average tenure roughly tripled. Trial-to-paid conversion and ARPU both improved by double digits, and the paid customer base grew by a multiple. Average tracked keywords per customer went up roughly fourfold — the leading indicator that told us onboarding was actually working before the retention numbers moved.
Why this one matters
It is the cleanest demonstration I have that retention is a product design outcome, not a lifecycle-marketing outcome. Nothing here was a discount, a save offer, or a win-back campaign.