Conversion Rate Optimisation

We start with a quantitative pass: where users drop out by step, device and traffic source, which pages carry the most revenue potential, and which segments convert far below their peers. That tells us where to look, not what is wrong.

The qualitative pass tells us why. Session recordings, heatmaps, on-site polls, customer and sales-team interviews, and reviews of support tickets consistently surface objections the business had stopped noticing: unclear delivery terms, missing specification detail, pricing that requires a phone call, or a form asking for information nobody wants to give at that stage.

Only then do we write hypotheses, each one stating the problem, the proposed change, the expected effect and how we will know. That format is what makes the programme auditable later, when someone asks whether any of it actually worked.

We score hypotheses on expected impact, confidence in the evidence and cost to implement, then run the short list. Most sites do not have the traffic to test everything, and pretending otherwise produces a backlog of inconclusive experiments that erodes the team's belief in the whole exercise.

For high-traffic pages we run proper A/B tests with pre-calculated sample sizes and a fixed read window. For lower-traffic pages and long sales cycles we use sequential before-and-after implementation with clean measurement windows, or test upstream where volume is higher, and we are explicit about the weaker evidence that produces.

Losing tests are documented as carefully as winners. A well-run programme produces a body of knowledge about your customers that outlives any individual change.

Across consumer and commercial sites the same themes recur: clarity of offer above the fold, proof close to the decision point, removing steps and fields, making price and terms explicit rather than requesting contact, and matching the page to the promise made in the ad or search result that brought the visitor there.

For ecommerce, delivery and returns clarity, product detail depth and checkout friction dominate. For considered purchases and lead generation, the biggest wins are usually in the form itself, the strength of the reason to enquire, and the speed and quality of what happens immediately afterwards.

We also test the offer, not just the page. Changing what you are asking someone to do — a quote instead of a call, a trial instead of a demo — frequently outperforms every layout change put together.

A CRO programme that lives in a testing tool decays. We implement winners properly in the codebase, document the reasoning, and feed insights back into ad creative, email and sales messaging, because a message that converts on a landing page usually converts in an ad too.

Reporting is tied to revenue rather than conversion rate alone, since a change that lifts conversion while cutting average order value is not a win. We agree that guardrail metric at the start.