Conversion rate optimisation: how to increase your conversionsConversion rate

The calculation is simple: conversions divided by visitors, multiplied by 100. If 1,000 people visit your website and 30 make an enquiry, your conversion rate is 3 per cent.

You decide what counts as a conversion: a submitted form, a phone call, a purchase from an online shop, a booked appointment or an application through the careers page.

Before making a change, define what you can measure and where gaps will remain. Use tracking only with a suitable legal basis, the required consent and transparent information; qualitative tests can complement quantitative data.

And forget industry benchmarks. No average can tell you whether your figure is ‘good’. Compare your own results: today against last month, and the variant against the original.

The calculation is clear: if the measured conversion rate rises with the same number of visitors and the quality of the conversions remains unchanged, the number of conversions rises mathematically. Whether a particular change achieves this remains to be seen.

More traffic creates ongoing work. Website improvements also incur analysis, development and testing costs, and they do not automatically affect every visitor in the same way.

With paid traffic in particular, it is worth checking whether the landing page and process contain unnecessary points of friction. A commercial improvement is possible, but not guaranteed.

CRO is therefore a systematic testing process, not a guaranteed lever for revenue.

CRO is not a gut-feeling project. It is a four-step cycle that you repeat.

Step 1: Analyse. Where do visitors leave? Which pages have high bounce rates? At what point do they abandon the form? Use web analytics, heatmaps and session recordings only with a suitable legal basis, consent and privacy-friendly configuration. The data provide clues, not a definitive cause.

Step 2: Form a hypothesis. State a testable assumption. For example: if we shorten the form from seven fields to three, enquiries will increase because the barrier is lower.

Step 3: Test. Half of the visitors see the original and the other half see the variant. Run the test until the result is robust rather than a product of chance.

Step 4: Evaluate and implement. If the variant wins, make it the standard. Document the result, then begin the next cycle.

Sound familiar? Measure first, then reallocate, then implement. We use the same principle to distribute advertising budgets: the figure comes before the decision.

Not every adjustment is worth the effort. These six areas provide a framework; analysis and testing will show which are relevant.

What distinguishes a landing page from a standard subpage? Focus. No navigation, no detours, one objective.

A campaign-specific landing page can provide a consistent link between an advert and its destination. Near the top, it should clearly reinforce the substantiated promise made in the advert; a test will show whether it performs better than an existing page.

The structure follows a clear logic: a headline stating the main benefit, a sentence that makes the promise specific, trust signals, benefits and a form. Nothing else.

Every element has a job. Anything without one goes.

What exactly is an A/B test? Two versions of the same page run in parallel and visitors are assigned at random. What matters in the end is which version generates more conversions.

The most important rule is one change per test. If you change the headline and form at the same time, you will not know which change made the difference.

Define the sample size, duration, evaluation method and stopping criteria before the test. An early interim result or gut feeling is not a robust basis for a decision.

Do not expect a fixed uplift. A test may show a difference, no reliable difference or measurement problems. Document ambiguous results and their uncertainty as well.

Prioritise by potential impact. Start with headlines, calls to action and forms because they affect every visitor's path.

You need fewer than tool providers claim. Google Analytics provides the fundamentals: traffic, behaviour and conversions. Heatmaps show where people click and scroll. Session recordings show individual user journeys.

Specialist A/B testing tools automatically assign visitors to variants. The right one depends on your setup, not the length of its feature list.

Then there is the most underrated tool: conversations. Sit beside real customers while they use your website. We do this on site in Saarbrücken or Neunkirchen, or at our office in Ottweiler. Open the laptop, watch and take notes.

Usability tests can reveal causes that aggregated dashboards alone cannot explain. Their scope and evidential value depend on the method and sample.

With low traffic, traditional A/B tests are often difficult to assess robustly. Qualitative usability tests, error analysis or a limited traffic test may then be more useful; these methods do not promise results either.

For fundamental technical or design problems, rebuilding with sound web design may make more sense than running many individual tests. Which route is faster or more economical depends on the current site.

Anyone who promises a fixed uplift percentage before seeing your figures is guessing. No credible provider can make that promise.