Case study

Acquisition was growing. Revenue wasn't.

SaaS product serving the freelancing market

Acquisition was growing, but the final growth number was not moving with it.

New customers were coming in. Too many were also leaving.

That meant part of the acquisition work was simply replacing customers we had already paid to acquire.

What looked obvious

The obvious answer was to improve retention campaigns and send more communication before renewal or cancellation.

But by the time somebody was cancelling, the real problem had often started much earlier.

What I noticed

I spoke to customers who had churned and looked at the lifecycle from signup onward.

The pattern was not simply price or competition. Onboarding, support and value realization mattered. Product usage often started changing before cancellation became visible.

Because switching friction was low, waiting for a cancellation signal was waiting too long.

What I designed

I designed a customer information and intervention system around the lifecycle rather than around cancellation.

It connected documentation, automated onboarding communication, product usage tracking and an internal dashboard.

We tracked signals such as signup, logins, usage frequency, time in product, important feature usage and changes relative to each customer's normal behaviour.

Intervention

The system was built to tell people when intervention was needed.

Support, account management, customer success and sales could see when an account was moving away from its own normal pattern and act before the risk became a cancellation.

Email was only one delivery mechanism. The real product was the information and intervention system behind it.

My role

I designed the system and led the implementation.

I was hands-on with dashboard logic, tracking, automation design and system behaviour. Developers implemented significant parts of the database, backend and application layer.

The technology included React, Java, shell scripts, PostgreSQL, Gmail API integration, internal automation and product usage tracking.

What changed

The result was not a single retention campaign. It was a different way of operating around customer risk.

The business could see change earlier and intervene while there was still something useful to do.

The numbers

57% reduction in new-account churn

25% reduction in established-account churn

These are reductions in the number of accounts churning relative to the previous baseline. They are not revenue-churn measures.

What stayed with me

Churn is often the last visible event in a problem that started much earlier.

If you want to change the outcome, you have to find the point where the behaviour first changed.

Back to Work