What conversion optimization actually is
Conversion rate optimization (CRO) is the practice of increasing the proportion of visitors who complete a valuable action on your website — a purchase, an enquiry, a trial sign-up or a booking. Its reputation has suffered from a focus on superficial tweaks, but done properly, CRO is closer to applied customer research than to design trivia.
The central question is simple: why do people who could become customers leave without doing so? Sometimes the answer is that they were never a good fit, and no page change will help. Often, though, they leave because something was unclear, missing, difficult or untrustworthy. Those are problems you can solve.
It helps to think of conversion as the outcome of three forces. Motivation is how much the visitor wants what you offer. Clarity is how easily they understand it. Friction is the effort and risk involved in acting. Most effective improvements increase clarity or reduce friction; motivation is largely shaped before the visitor arrives.
Research comes before testing
A common mistake is to begin with a list of test ideas gathered from best-practice articles. Some of those ideas may help, but they are generic by definition. Your visitors, offer and context are specific. Research identifies the problems that actually exist on your website, which means your tests are far more likely to matter.
Good CRO research combines two kinds of evidence. Quantitative data tells you where and how much. Qualitative data tells you why. Neither is sufficient alone.
Quantitative signals: where people drop off
Start by confirming that tracking is accurate — many CRO programmes have been derailed by broken events. Then map the key journey step by step and look at drop-off between steps. Useful cuts include:
- Device type: mobile experiences frequently underperform for reasons that are invisible on a desktop review.
- Traffic source: visitors from different channels arrive with different expectations.
- New versus returning visitors: first-time visitors usually need more reassurance.
- Landing page: pages that receive significant traffic but rarely lead onward deserve attention.
- Form field interactions: individual fields that cause errors or abandonment.
Page speed belongs here too. Slow pages, layout shifts and unresponsive interactions frustrate visitors before they read a word. Google’s Core Web Vitals provide a reasonable starting framework for measuring these issues.
Qualitative signals: why they hesitate
Once you know where problems occur, qualitative research helps explain them:
- Session recordings show real behaviour: hesitation, repeated clicks on non-clickable elements, scrolling past key information.
- On-site surveys ask a single, well-timed question, such as what is stopping someone from continuing today.
- Usability testing asks representative people to complete a task while thinking aloud. Even a handful of sessions tends to reveal recurring issues.
- Customer and sales conversations surface the questions and objections people raise before buying.
- Heuristic review by experienced practitioners assesses clarity, trust and usability against established principles, including accessibility.
Session recording and survey tools collect information about visitors, so they need to be configured with privacy in mind — masking form inputs, respecting consent choices and being disclosed in your privacy and cookie policies.
Writing a hypothesis worth testing
Research findings become useful when they are turned into specific hypotheses. A good hypothesis connects an observation to a change and an expected result:
Because we observed that mobile visitors frequently abandon the quote form at the “property details” step, and survey responses suggest they do not have this information to hand, we believe that allowing visitors to save progress and return later will increase completed quotes among mobile visitors.
This format forces clarity. It records the evidence, names the change, identifies the audience and states what you expect to happen. When the test ends, you learn something whatever the result, because you know what you believed and why.
Prioritising what to test
Research usually produces more ideas than you can test. A simple scoring approach helps teams agree on order. For each hypothesis, consider:
- Evidence: how strongly the research supports the problem.
- Impact: how many visitors the change affects and how close it is to the conversion point.
- Effort: design, development and approval time required.
Avoid letting effort dominate. Easy tests are attractive, but if they only affect a small part of the journey, they are unlikely to move overall results.
A/B testing discipline
A/B testing compares two or more versions of an experience by randomly splitting traffic between them. It is a powerful method, and an easy one to misuse. A few principles prevent most problems:
- Estimate sample size in advance. Use your baseline conversion rate and the smallest effect you care about to estimate how many visitors each variation needs.
- Do not stop early because results look good. Repeatedly checking and stopping at the first significant result greatly increases the chance of a false positive.
- Run for full business cycles. Behaviour often differs by day of week; running for whole weeks reduces that bias.
- Choose one primary metric. Decide before launch which metric determines the winner, and treat others as supporting evidence.
- Quality-assure every variation across devices and browsers. A broken variation produces misleading results.
Expect many tests to be inconclusive. That is normal, and it is still information: the change did not matter as much as you thought, which helps redirect effort.
When you cannot run a test
Lower-traffic websites often cannot reach reliable sample sizes in a reasonable time. That does not mean CRO is out of reach. Research-led improvements — fixing clear usability issues, clarifying offers, simplifying forms — can be implemented directly and monitored over time. The evidence standard is lower than a controlled experiment, so be cautious in attributing changes, but the improvements themselves are usually sound.
Usability testing becomes especially valuable here, because its value does not depend on traffic volume.
Ethical optimisation
Some techniques can increase short-term conversion by misleading or pressuring people: fake countdown timers, invented stock warnings, pre-ticked add-ons, hidden fees revealed at checkout and cancellation flows designed to confuse. These are often called deceptive design patterns.
Beyond damaging trust, several of these practices can breach consumer protection and competition law, and advertising platforms prohibit destinations that mislead users. Ethical CRO uses only true information: real deadlines, actual stock levels, clear prices and honest descriptions. Clarity is both the ethical choice and, over time, the effective one.
Building an ongoing programme
The value of CRO compounds when it becomes a routine rather than a project. A sustainable programme typically includes a shared backlog of hypotheses, a regular cadence of research and testing, a central record of results — including losses — and a habit of sharing learnings with media, creative and product teams.
That last point is often overlooked. A test that shows which reassurance matters most to visitors is useful for ad copy, sales conversations and product decisions too.
If you would like help understanding where your website loses potential customers, our conversion optimization team can help. You can also read about building a performance marketing strategy that sends the right visitors in the first place.
