11 conversion rate optimisation mistakes, and how to fix them
The CRO mistakes that keep coming up: testing button colours, chasing lead volume, ignoring mobile and speed. Here's what goes wrong and how to fix each one.
Conversion rate optimisation has a reputation problem. Mention it in a marketing meeting and half the room pictures somebody A/B testing the shade of a button for six weeks. The other half pictures a consultant’s slide deck full of heatmaps. Neither is what good CRO looks like, and both come from the same small set of mistakes.
Here are eleven we see often enough that we’ve stopped being surprised by them, and what to do instead.
1. Optimising the conversion rate instead of the outcome
A conversion rate is a ratio, and ratios are easy to flatter. Remove the phone number field and form completions go up. Did revenue? Often it goes down, because sales now has a pile of email-only leads they can’t call. Pick the number that pays the bills (booked appointments, qualified opportunities, orders) and optimise towards that. We made a longer version of this argument in the case against more leads.
2. Testing trivia first
Button colour, headline font, the word “Submit” versus “Send”. These are easy to test and almost never matter. The changes that move numbers are structural: what you ask for, when you ask, what happens next. If a test wouldn’t change how a visitor experiences the page in any way they’d notice, it probably won’t change their behaviour either.
3. Calling a test after three days
A test that’s run for a weekend has measured a weekend. Traffic on Monday morning behaves differently from traffic on Saturday night, and a campaign email on Wednesday can swamp everything. Decide how long a test runs before you start, run it over full weeks, and don’t peek at the dashboard with a view to stopping early when it looks good.
4. Running tests on pages with no traffic
If a page gets a few hundred visits a month, an A/B test on it won’t reach a meaningful answer this year. That’s fine. Low-traffic pages can still be improved, just not by split testing. Use judgement, watch session recordings, ask the sales team what customers say, and make the change. Not every decision needs a p-value.
5. Forgetting that most visitors are on a phone
It’s common for the team building a website to review it on large monitors while the bulk of visitors arrive on a phone, often one-handed, often half-distracted. A form that looks tidy on a desktop can be a scrolling slog on mobile. Do your reviews on a phone. Better still, do them on a cheap phone on a train.
6. Ignoring speed
A page that takes several seconds to become usable loses people before your clever copy has a chance. Third-party scripts are the usual culprit: three analytics tools, two chat widgets from previous vendors nobody removed, a heatmap tool from a project that ended last year. Audit what loads, remove what nobody uses, and check again after every new tool goes in.
7. Treating the form as the finish line
The visitor fills in the form, sees “Thanks, we’ll be in touch”, and the CRO team counts a conversion. Then the lead sits in an inbox for a day and a half. From the customer’s side, the experience hasn’t ended; it’s stalled. Look at what happens after submission: how fast someone replies, whether they can book a time straight away, whether the lead reaches the right person. If you sell property, Uprealer’s lead routing and booking is one example of closing that gap. Speed to lead often matters more than anything on the page itself.
8. Asking for commitment before giving value
“Book a demo” as the only call to action on every page assumes every visitor is ready to talk to sales. Most aren’t. Offer steps of different sizes: a quick question answered, a price range, a short guide, a comparison. Someone who’s had a useful exchange is much more likely to take the bigger step later.
9. One generic call to action everywhere
A visitor on a pricing page has a different question from one reading a how-to article. Same button, same wording, same form on both is a missed opportunity. Make prompts page-aware: on pricing, offer to help choose a plan; on a product page, offer a specific answer about that product. A conversational form or AI agent that knows which page it’s on (the kind of thing Serviceform is built for) does this far better than a static form, and we’ve written more on page-aware conversations.
10. Making changes nobody can measure
A new chat widget goes live, the form gets shortened, and the homepage is redesigned, all in the same fortnight. Leads go up. Which change did it? Nobody knows, so nobody knows what to keep. Space changes out, write down what you changed and when, and make sure leads carry the source page and the tool that captured them all the way into the CRM.
11. Stopping at the website
The most expensive CRO mistake is assuming the job ends when the lead is captured. A website that converts beautifully into a CRM nobody works is a vanity project. Close the loop: tie website conversations to what sales did with them and what they turned into. That’s what tells you which pages, offers and prompts are worth more effort, and which ones just generate noise.
A quick example of what fixing this looks like
Picture a B2B services firm with a single “Contact us” form on every page and a 1.5% form conversion rate. The numbers are made up, but the setup is common. A conventional CRO project might test the button and the headline. A better one would ask: what are visitors on each page trying to find out? Then it would replace the generic form with a short conversational flow on the three highest-intent pages, add a booking step for anyone who qualifies, and route each enquiry to the right team. Even if the headline conversion rate barely moved, the number of booked first meetings could change a lot, and that’s the number the sales director cares about.
What good CRO actually is
Strip away the tooling and conversion optimisation is a simple discipline: understand what visitors are trying to do, remove what gets in the way, and make the next step obvious and easy. Measure what matters, not what’s convenient. And remember that the visitor doesn’t experience your website, your form tool and your sales team as three separate things. They experience one company, and they judge it by how quickly and helpfully it responds.