In short
Conversion rate optimisation (CRO) refers to the systematic, data-driven process of increasing the share of website visitors who carry out a defined desired action, the so-called conversion. In e-commerce this is typically completing a purchase, but it can also be a newsletter sign-up, a contact enquiry, a download or creating an account.
The conversion rate itself is a simple metric: the number of conversions divided by the number of visitors, as a percentage. With 5,000 visitors and 100 orders, the conversion rate is 2%.
CRO is the discipline of improving this rate, not by buying more traffic but by getting more out of the existing traffic. This makes CRO particularly cost-effective: every additional conversion with the same advertising budget lowers acquisition costs and increases returns.
Why CRO is an economic lever
A simple calculation illustrates the effect: a shop with 50,000 visitors a month, a conversion rate of 2% and an average order value of €80 makes 1,000 orders and €80,000 in revenue. If optimisation raises the conversion rate to 2.5%, the same traffic produces 1,250 orders and €100,000, i.e. €20,000 more without a single cent of additional advertising budget. It is precisely this leverage effect that makes CRO so attractive in digital marketing.
CRO is a process, not a trick
Serious CRO follows an iterative cycle instead of random "best practices":
- 1. Analyse: where do users drop out? Data from web analytics, heatmaps, session recordings and funnel analysis show the weak points.
- 2. Form a hypothesis: formulate a reasoned assumption, for example: "Because the note on shipping costs only appears in the last step, many people abandon. If we show it earlier, the abandonment rate will fall."
- 3. Test: measure the change against the existing version, usually via an A/B test (see below).
- 4. Evaluate: did the variant perform statistically significantly better? Only then is it adopted.
- 5. Repeat: the cycle starts again. CRO is never "finished".
A/B tests and multivariate tests
The most important tool of CRO is the A/B test: two variants of a page (A = original, B = change) are served at the same time to a random half of the visitors each, and you measure which variant converts better. Statistical significance is important: a difference must be large enough and based on enough data not to be mere chance. In a multivariate test, several elements are tested in combination at the same time, which requires considerably more traffic.
Typical levers in e-commerce
| Area | Typical lever |
|---|---|
| Checkout | Guest checkout, fewer fields, clear shipping costs, trust signals |
| Product page | Better images, clear CTAs, reviews, availability |
| Performance | Faster loading times (Core Web Vitals) |
| Trust | Seals, reviews, transparent returns, visible contact details |
| Mobile | Touch-friendly operation, short forms |
A concrete example
In its funnel analysis, a Shopware shop finds that 60% of users leave the basket as soon as the shipping costs appear at checkout. The hypothesis: the late surprise puts people off. In an A/B test, half of the visitors see the shipping costs clearly shown on the product page and in the basket. After three weeks with sufficient data volume, the variant shows a significantly lower abandonment rate at checkout. The change is rolled out. Important: the result is context-specific; another shop could see a different result with the same intervention, which is why it is tested rather than guessed.
Common misconceptions
First: CRO is not copying "best practices". What works for one shop can fail for another; target group, range and brand context are decisive. That is why tests are run. Second: a higher conversion rate is not automatically better if it comes at the expense of order value or customer quality; CRO has to look at the overall return, not just the rate. Third: a single A/B test that was won is no proof of a truth; only a series of consistent results produces reliable knowledge. Fourth: CRO is not a one-off project but an ongoing process.
Distinction from SEO and performance marketing
SEO and paid advertising bring visitors to the site; CRO ensures that these visitors become customers. The disciplines complement each other: more traffic with poor conversion burns budget, and good conversion without traffic has no effect. In practice, good teams work on both in parallel and look at the entire chain from the first click to the order. This chain only becomes measurable once tracking & analytics have been set up cleanly.
Outlook
CRO is increasingly shaped by AI and personalisation: systems automatically serve different user segments the variant that converts better for them, and analysis tools recognise points of friction faster. The core, however, remains unchanged: understand where users fail, form a hypothesis, test cleanly, evaluate. A sound introduction to test procedures and methodology is offered, for example, by the overview of conversion rate optimisation on Wikipedia.
FAQ
How do you calculate the conversion rate?
The number of conversions divided by the number of visitors, times 100. Example: 100 orders from 5,000 visitors = 2%.
What is an A/B test?
A procedure in which two variants of a page are served at the same time to random groups of visitors in order to measure which converts better.
What is a "good" conversion rate?
There is no universal value; it depends heavily on industry, product, price and traffic source. What is meaningful is the development of your own value over time, not a comparison with blanket figures.
Is CRO worthwhile with little traffic?
Qualitative methods (usability tests, heatmaps) always help. Statistically reliable A/B tests, however, need a minimum number of visitors to deliver meaningful results.