The situation
The business had strong growth in the United States and much weaker performance internationally, with GDPR compliance becoming a hard requirement for European expansion. At the same time it was under pressure to improve profitability and Rule of 40 performance ahead of a likely acquisition process.
The product also carried real technical debt. A custom JavaScript page builder with non-standard DOM behaviour limited flexibility and constrained adjacent product expansion. Commercially, the platform was positioned as a traditional A/B testing tool at exactly the moment machine learning began reshaping conversion optimisation.
What the real problem was
This was not a roadmap question. It was a strategic inflection point with two live options and only enough capacity for one.
The industry consensus at the time favoured AMP — Google-backed, low-risk, easy to explain to a board. The alternative was to spend the same engineering capacity on data readiness, GDPR, and a machine-learning conversion capability that did not yet have a market category, let alone demand.
Choosing the second meant telling the executive team we were deliberately not doing the thing everyone expected.
What I changed
- Deprioritised AMP in favour of GDPR readiness and a multi-year conversion-intelligence strategy — a conviction call made against the prevailing market view.
- Repositioned the platform from static A/B testing toward AI-driven, context-aware conversion optimisation.
- Led the creation of Smart Traffic, which used UTM and visitor context to route each visitor to the variant most likely to convert for them.
- Expanded the platform with adjacent conversion tooling, including exit intent and pop-ups, to widen the wedge rather than deepen a single feature.
- Restructured engineering around clearer technical ownership, and introduced product-owner training and Scrum of Scrums to fix coordination across a growing org.
What happened
ARR grew more than threefold over the period at a strong compound rate, while net profit multiplied and churn fell by double digits. The AI conversion capability shipped as one of the earliest of its kind in MarTech, outperformed human conversion experts in head-to-head testing, and generated two patents. The company was positioned for, and completed, an exit to a private equity buyer.
The AMP bet the rest of the category made did not pay off.
Why this one matters
It is the clearest example I have of a conviction-led product bet: choosing the harder, less legible option, resourcing it properly, and holding the line long enough for it to become the reason the company was worth buying.