Conversion rate: what to measure before a redesign

Conversion rate describes how often website visits lead to an action you want to measure. For a service business, that might be an enquiry. For an online store, it might be an order. Start by defining exactly what counts as a conversion.

How is it calculated?

If you compare enquiries with visits, divide enquiries by visits and multiply by 100. Hypothetical example: three enquiries from 100 visits gives a rate of 3%. Keep the definition and reporting period consistent when comparing results.

A visit is not always a different person: someone can return several times. Some reports use visits and others use users. Check the definition before comparing figures from different reports.

Different actions have different value

Clicking a phone number does not prove a call took place. Sending a form does not prove the enquiry is relevant. Adding an item to a basket is not a sale. Track these steps separately from the outcome that matters to your business.

What should you check before rebuilding?

Start with basic checks. Does the form work, and does its message arrive? Can the page be used on a phone? Is the offer understandable? Can a visitor find the price, delivery information or next steps without searching around?

Check where visits come from as well. A redesign will not fix an advertisement attracting people who want something you do not offer. Incorrect measurement can also report too many or too few conversions.

A more useful first intervention

Hypothetical example: a site receives visits but few enquiries. On a phone, its form asks for too much information and shows an error that is difficult to notice. Repairing and simplifying that form is a better first intervention than rebuilding every page. Then check whether visitors can complete it more easily.

How do you evaluate a change?

Compare relevant periods and account for traffic sources, campaigns and seasonal differences. When there are few visits, a handful of enquiries can substantially change the percentage. Avoid firm conclusions from very small samples.

Comparing two page versions can help when enough data is available. On a small site, testing the form and speaking with customers may identify obvious obstacles sooner.

See user experience and conversion improvement and data and reporting checks.