Most businesses do "CRO" as a series of random acts: change the button colour, move the form, A/B test the headline — then wonder why conversion barely moves. That's not conversion rate optimisation; that's decoration with statistics. Real CRO is a strategy: a systematic process for understanding why visitors don't convert and fixing the actual causes, in priority order. Here's how to build one.
Why tactics-first CRO fails
The button-colour test is the classic cargo cult. Someone read that orange buttons convert better, so they test orange buttons, see a 2% wobble, and declare CRO doesn't work. The problem: button colour was never the reason visitors weren't buying. The reasons are usually structural — wrong traffic, unclear value, missing trust, friction in the journey — and no amount of tactical tinkering fixes a structural problem.
Strategy-first CRO flips the order: research before testing, hypotheses before changes, and a prioritised backlog driven by evidence rather than opinions. The tests get smaller; the wins get bigger.
Phase 1: Understand (before you change anything)
Four research streams, none optional:
- Analytics forensics. Where do visitors drop out? Segment by traffic source, device, new vs returning. A checkout that converts on desktop but dies on mobile isn't a copywriting problem. Look at the shape of the funnel before theorising about it.
- Session recordings and heatmaps. Watch real visitors use the site. You'll see rage clicks, confused scrolling, forms abandoned at the same field — the qualitative truth behind the quantitative drop-offs. Ten sessions teach you more than ten reports.
- Voice of customer. On-site polls ("what's stopping you from enquiring today?"), post-purchase surveys ("what nearly stopped you?"), sales team debriefs, review mining. Your visitors will tell you what's wrong if you ask — most businesses never do.
- Heuristic and competitor review. An expert pass over the key journeys against known usability principles, plus a look at how competitors handle the same journeys. Not for copying — for calibrating what's normal in your market.
This phase typically surfaces 30–50 issues. That's normal. The next phase decides which ones matter.
Phase 2: Prioritise ruthlessly
You can't fix everything. Score each issue on three axes:
- Impact: how many visitors does it affect, and how severely? A friction point on the checkout page outranks one on the about page, always.
- Confidence: how strong is the evidence? Analytics + recordings + customer quotes beats one person's hunch.
- Ease: how hard is the fix? Quick wins build momentum and fund the bigger projects politically.
Work the high-impact, high-confidence items first regardless of ease — but sequence a few easy wins early to prove the programme's value to sceptical stakeholders. CRO programmes die without visible early results; plan for that reality.
Phase 3: Hypothesise and test properly
Every change gets a hypothesis in plain English: "We believe [change] will cause [effect] because [evidence]." If you can't write the "because", you're guessing — which is fine occasionally, but call it what it is.
On testing discipline — the unsexy part that determines whether your results mean anything:
- Test one thing at a time (or run proper multivariate tests if you have the traffic). Changed the headline, image and form in one test? You learned nothing.
- Run tests to significance with adequate sample sizes. Ending tests early because the first week's numbers look good is how businesses "learn" false lessons.
- Test big things. Headline tweaks move conversion by percentages; fixing the value proposition, removing process friction or adding missing trust signals moves it by multiples. Bold tests, not timid ones.
- Document everything. Winners, losers, inconclusives — all of it, in a shared log. A test that "failed" taught you something about your customers; that's an asset if you record it.
Phase 4: The continuous loop
CRO isn't a project; it's a capability. The mature rhythm: monthly research review, prioritised backlog, 2–4 tests running continuously, results fed back into understanding. Over a year, this compounds — each test teaches you something that sharpens the next hypothesis.
And keep perspective on what CRO can and can't do. It can't fix bad traffic (visitors who were never going to buy), a weak offer, or a broken product. If conversion is catastrophically low, the problem is usually upstream — targeting, positioning, pricing — not the shade of the CTA button. The strategy tells you honestly when the answer is "fix the offer", which is precisely why strategy comes first.
What good looks like
A healthy CRO programme for an SME: a living backlog of 15–30 prioritised hypotheses, tests running most weeks, a win rate around a third (if everything wins, you're not testing boldly enough), and conversion trending up quarter over quarter. Expect the programme to pay for itself within two quarters if the research phase was honest — if it doesn't, the problem is usually traffic quality or offer strength, and the research will have told you that too. The secondary benefits — deep customer understanding, a testing culture, evidence-based decision making — often outlast any individual test result.
CRO and SEO: allies, not rivals
CRO and SEO are often run as separate workstreams that occasionally sabotage each other — CRO removes "clutter" that was ranking content; SEO adds copy blocks that hurt conversion. The mature approach treats them as one discipline: search-optimised pages that convert.
The reconciliation is straightforward: keep the ranking content (Google needs the words), but design it to serve conversion — specs and guides that answer buying questions and build trust, FAQ sections that handle objections and capture featured snippets, comparison tables that rank and sell. Test CRO changes with SEO impact in mind (watch rankings after big template changes), and brief SEO content with conversion intent from the start. Our service page template was built on exactly this principle: pages that rank and convert, because a page that does only one is half a page.
Our conversion rate optimisation service runs the full strategy for UK businesses — research, prioritisation, testing and implementation, all tied to revenue rather than vanity metrics. If your site gets traffic but not enquiries, get a free consultation and we'll show you where it's leaking.
