In short
If your website gets visitors but almost no enquiries, work out first which problem you have. There are two, they look identical in a traffic report, and they have opposite fixes. Either the right people are arriving and the site is failing to convert them, or the wrong people are arriving and the site is converting as many as it ever will. Most businesses assume the first and redesign pages for an audience that was never going to buy.
The conversion problems themselves are usually mundane: no obvious next step, a form asking far too much too soon, a phone number invisible on mobile, slow pages, nothing that makes a stranger trust you, no price information anywhere. Then the part nobody counts as a website problem at all — how fast you reply once an enquiry lands.
A warning about method, too: at the volumes most UK small businesses have, A/B testing cannot give a meaningful answer. Judgement and customer feedback beat pretending you have the numbers for statistics.
First: is it conversion, or is it intent?
Almost everyone gets this backwards, and it is expensive. The same low conversion rate can mean a page is badly built, or simply that visitors were searching for something adjacent to what you sell. A joinery firm ranking for “how to fix a sticking door” gets traffic and almost no enquiries, and no amount of button colour changes that — the visitor is not a bad prospect, they are not a prospect. So look at what people searched for before arriving: Search Console shows the queries reaching each page.
- Informational queries dominate (“how to”, “DIY”) — an intent problem. The page is doing its job; the job is not enquiry generation.
- Commercial queries dominate (“near me”, “cost”, place names) — a conversion problem, and the rest of this applies.
- Your brand name dominates — these people already know you, so something is actively losing them.
Then compare commercial pages against the site average: if the services page converts many times better than the blog, which converts at nothing, the sitewide conversion rate is meaningless.
You cannot fix what you are not measuring
Plenty of sites cannot say how many enquiries arrived last month, let alone from which page. Fix that first, because everything after it is guesswork. Track form submissions as events rather than inferring them from a thank-you page, plus phone taps, email clicks and chat clicks, each attributable to a landing page and source. Google Analytics 4 has a recommended event for exactly this: generate_lead.
Then count by hand the enquiries that reached your inbox and phone last month and compare. If they diverge badly, the tracking is wrong and every decision from it will be too. That reconciliation is the foundation of our analytics and tracking work.
The friction points, in the order they usually bite
No clear next step
Read your main service page and ask what you want the visitor to do. If the answer takes more than four words, the page does not know either. One primary action, near the top and again at the bottom.
Forms that ask too much, too soon
Every field is a reason to leave. A form asking for company size, budget range and a full project description is for people who have already decided. Name, one contact method and one free-text box gets you further. Qualify on the phone — qualifying first means you never qualify at all.
A phone number nobody can find on a phone
For trades and anything urgent the phone is still the main channel, yet the number is often buried in the footer, not tappable, or two taps away. Put a tappable number in the mobile header — ten minutes of work, routinely one of the highest-yield changes there is.
Pages that are simply slow
Slowness costs you visitors before they read anything. Google’s Core Web Vitals give concrete targets: Largest Contentful Paint within 2.5 seconds, Interaction to Next Paint of 200 milliseconds or less, and Cumulative Layout Shift of 0.1 or less, at the 75th percentile of page loads. Google documents how these relate to Search. Most sites fail on mobile for unglamorous reasons: uncompressed images, a page builder loading scripts it does not need, too many tracking tags. Layout shift is the sneaky one, because a button that moves as the page settles gets mis-tapped. That is development work.
Nothing that makes a stranger trust you
A visitor who has never heard of you is making a risk judgement in seconds. Give them something real: reviews with names, photographs of actual work and actual people, the town you are based in, a full address. It must all be true — UK marketing must be legal, decent, honest and truthful, per gov.uk and the CAP Code.
No pricing information at all
A site with no price signal filters out serious buyers as effectively as unrealistic ones, because they will not book a discovery call to find out whether you are in their league. You need a bracket, not a price list. That is why we publish a pricing page.
Why “request a quote” converts worse than a specific ask
It is the default call to action on British small business sites and weaker than it looks. It is vague about what happens next — a number in an hour, a call tomorrow, a fortnight of chasing? It implies commitment, which someone still comparing options is avoiding. And it asks for effort without offering value. Specific asks answer all three:
- “Check if we cover your postcode” — trivial, useful, self-qualifying.
- “Get a same-day price for a standard boiler service” — concrete and bounded.
- “Book a 15-minute call” — the time cost is capped and explicit.
- “Send a photo and we’ll say if it’s repairable” — the visitor gets a real answer.
Each is a smaller commitment than “request a quote” while producing a better-qualified lead, and that is the substance of conversion rate optimisation. Mobile and desktop need different treatment here: mobile visitors are in motion and want the number and a price fast, while desktop visitors compare suppliers with tabs open. Split conversion reporting by device. If desktop converts well and mobile barely converts at all, the blended figure is not a result — it is a mobile problem hidden inside an average.
The bit that happens after the enquiry lands
You can fix everything above and still lose the work, because the enquiry sits unanswered until Thursday. Someone who submits a form has usually submitted two or three, and at that moment they are receptive. Several hours later they are not. None of this is a website problem, which is why website reporting never surfaces it.
- Send notifications to a phone, not only an inbox checked twice a day.
- Acknowledge automatically, stating truthfully when a human will respond.
- Name one person responsible for first response, with a target — same working day — that you check.
- Check the spam folder weekly; form notifications land there more often than expected.
If you follow up beyond answering the original enquiry, the ICO’s guidance on electronic mail marketing sets out the rules under PECR: replying to an enquiry is not the same as marketing to that person later.
Testing honestly when your traffic is low
If your site gets a few hundred visits a month and a handful of enquiries, you cannot run a valid A/B test. Detecting a realistic improvement needs conversions in the hundreds per variant, so at ten a month the test runs for years. Running it anyway is worse than not testing: you read noise as signal and never know.
- Fix the obvious things untested. A visible mobile phone number does not need validating.
- Ask real people. Six customers asked “what nearly stopped you getting in touch?” will tell you more than a year of underpowered testing.
- Read session recordings qualitatively. Twenty recordings where people hesitate at the same point is a finding, even if it is not a statistic.
- Compare over long windows, and date every change. Leave a change a quarter, compare to the same quarter last year, and treat it as directional. The dated log is the only way to interpret a shift six months later.
Frequently asked questions
What conversion rate should my website achieve?
There is no credible universal number, and anyone quoting one is describing a different business. It depends on sector, price point and traffic mix. The only useful benchmark is your own site last quarter.
Is it a traffic problem or a website problem?
Read the queries reaching each page in Search Console. Mostly informational means an intent problem, fixed with different content or channels. Mostly commercial and still not converting means the site is losing buyers.
Is a redesign the answer?
Usually not as a first move. Redesigns change everything at once, so when numbers move you cannot tell what caused it, and they often remove what was working. If the site is fundamentally slow or unmaintainable, rebuild on those grounds instead.
How long before I see a difference?
It varies too much to promise anything, and we will not. A tappable phone number can change behaviour the day it goes live; other changes take a quarter of data before you can tell.
Key takeaways
- Decide whether you have a conversion problem or an intent problem before changing anything. The queries reaching each page will tell you.
- Sitewide conversion rate blends audiences with nothing in common. Segment by landing page and device, or you optimise an average that describes nobody.
- Reconcile tracked enquiries against the ones that reached your inbox and phone; broken measurement produces confident, wrong decisions.
- Replace “request a quote” with an ask that states what the visitor gets and what it costs them in time.
- Response speed after the enquiry lands sits outside the website entirely and can undo everything it did right.
If you would like a second pair of eyes on where your site loses people, that is what our conversion work is for, and the industries page shows the sectors we work in most.