Conversion rate is a fraction. Most effort goes into the numerator, trying to persuade more of the people who arrive.
The larger opportunity is usually in the denominator: the people who arrived, could not establish whether you were relevant to them, and left without registering as anything. They are not in your conversion rate. They are why it is low.
Web conversion optimisation works best when treated as removing reasons to leave rather than adding reasons to stay.
Web Conversion Optimisation: A Practical Business Guide
1. Map the actual journey first
Before changing anything, write down the steps between somebody arriving and completing your primary action.
Count the page loads. Count the decisions. Count the fields. If the total is four steps and could be two, every extra one is a place to lose somebody, and no amount of persuasive copy compensates for an unnecessary step.
Then find where people leave, using drop-off by step in analytics and by watching five people attempt the task. Analytics reports where; observation explains why. Neither works alone, and most businesses have only the first.
2. Answer the qualifying questions immediately
Most visitors are not deciding whether to buy. They are deciding whether you are relevant, which is a faster and harsher judgement.
Four things settle it: what you do, who it is for, roughly what it costs where you can say, and what happens if they get in touch. Sites answer the first and leave the rest to be inferred, which forces people to hunt or guess.
Put those where they need no searching. Uncertainty about price deters more enquiries than a price does.
3. Reduce the friction in the mechanism
One primary action per page, placed where the decision occurs rather than only at the bottom.
Wording that names the outcome. “Send my enquiry” answers what the reader is agreeing to; “Submit” describes what the form does for you.
Forms limited to what routes the enquiry. Every field is answered by everybody, so a question useful once in twenty costs you the other nineteen. Validate inline so nobody retypes after an error.
And make sure the whole journey works with a keyboard. An enquiry form failing keyboard navigation turns away somebody who had already decided, which is lost revenue rather than a compliance question. Work towards WCAG 2.1 Level AA and test the enquiry path specifically.
4. Speed is part of the funnel
A page nobody waits for converts at zero, and the visitors most affected are frequently those on the worst connections.
Test on a real mid-range phone on mobile data rather than a narrowed desktop window. Where performance and environmental impact both apply, treat them as the same lever: reducing unnecessary data transfer and processing supports a faster experience while lowering estimated digital carbon, though environmental figures remain modelled rather than directly measured.
What happens after the enquiry counts too
Conversion is usually measured at the form submission, which is where the website’s job ends and where the actual outcome is still undecided.
An enquiry answered in four hours converts differently from one answered in three days. A confirmation screen saying only “thank you” leaves somebody wondering whether anything was received, and some of them enquire elsewhere while waiting.
Two changes cost nothing. Say who will respond and roughly when, on the confirmation screen and in the automatic email. And make sure somebody owns the inbox, because a form performing perfectly into an unmonitored address is a conversion problem no amount of page optimisation will find.
This is worth checking before spending anything on the funnel above it. Businesses regularly discover their real leak is here.
Test against outcomes
Measure enquiries rather than clicks. A change producing more clicks and fewer qualified enquiries has made things worse, and click-through rate will report it as a success.
Change one thing at a time so you can attribute the result.
Be honest about volume. Most small businesses lack the traffic for a meaningful split test in a fortnight, and a two per cent difference across ninety visits is noise. Below that threshold, five people struggling with the same field is better evidence than a test measuring randomness.
Our sustainable web design work treats journey, speed and accessibility as one brief rather than three separate projects.