The visits come in, traffic grows, enquiries don't. And the first instinct is nearly always the same: redesign the site.
A website isn't converting when it gets traffic but doesn't produce the action it exists for: a quote request, an order, a phone call. That describes a symptom, not a cause. The path that takes a visitor to a conversion has three points where it can break, and a fix only helps if it starts from the right one.
Below is the sequence we use to isolate it: the benchmarks, and the report to open at each step.
First question: is your number really low?
A conversion rate only means something when you compare it with something similar. And similar doesn't mean your industry: it means the same device, the same traffic channel, the same kind of visitor. The gaps between those segments are wider than the gaps between industries.
Every year Contentsquare publishes a benchmark built on 99 billion sessions from more than 6,500 websites, this time covering Q4 2024 to Q4 2025. It's an international sample, not a UK one, and should be read as an order of magnitude.
| Segment | Conversion rate | Note |
|---|---|---|
| Desktop | 3.4% | 30% of traffic |
| Mobile | 2.0% | 70% of traffic |
| Returning visitors | 2.9% | |
| New visitors | 1.7% | |
| Paid search | 2.8% | |
| Traffic from AI assistants | 1.3% | +55% year on year |
| Organic social | 0.7% |
Between 0.7% for organic social and 3.4% for desktop there's a factor of five. An overall conversion rate, read without segmenting it, tells you nothing: a site with 70% of its traffic from social that converts at 1% is doing fine, while the same 1% on paid search traffic is a problem that costs money every day.
For ecommerce there's a second reference point closer to home: IRP Commerce measures an average rate of 2.26% across its UK and Irish client base in July 2026, against 1.94% in July 2025.
In 2026 the denominator moved
Part of the problem isn't on your site. In the same sample, over twelve months, visits fell by 3.8%, engagement by 10% and the conversion rate by 5.1%. Average order value rose by 6%.
Fewer people, more decided. Average time per session is 4 minutes 46 seconds on desktop and 2 minutes 20 seconds on mobile; scroll rate stops at 50.5% on desktop and 45.2% on mobile. Half of visitors never reach the middle of the page.
Then there's the new channel. Traffic from AI assistants is still only 0.2% of the total, converts at 1.3% and has grown 55% in a year. It's small, and that's exactly why you should isolate it now: by the time it weighs enough to move the overall number, it will be too late to work out where it was coming from. If organic traffic is falling while your rankings haven't changed, the explanation more often lies in how Google answers above the results than in the site.
The UK context adds a figure that the international benchmarks don't have. According to the ONS Digital Economy Survey (published April 2023, data for 2021), UK businesses made £459.2 billion of website sales, but even in retail, the sector with the highest share, only 34.5% of businesses sold through a website. Online now accounts for 28.8% of retail sales in Great Britain (ONS, August 2026, published 18 September 2026), up from 28.4% in July. Most UK company websites don't sell: they collect enquiries. And a site that collects enquiries is measured differently.
Three breaking points, in the order you measure them
- Upstream. The people arriving weren't in the market. The site works, the traffic doesn't.
- On the page. They were in the market and got lost: they didn't understand what you sell, didn't trust you, couldn't find the next step.
- At the exit. They wanted to convert and couldn't: the form, the checkout, costs that only appear at the end.
The order isn't decorative. Each point contains the next: if the traffic is wrong, improving the form produces nothing measurable, because the people who would fill it in aren't there. You work through the sequence from the top and stop at the first point where the numbers fall outside the benchmarks.
Upstream: you're being found for the wrong searches
The articles that rank for this question hardly ever mention this cause. On Google UK, for "why is my website not converting" (checked on 10 October 2026), the snippets of the seven organic results on page one talk about navigation, mobile, product descriptions, UX, trust, messaging and calls to action; only one, Listing Mirror, mentions the audience. In the projects we work on, it's the most frequent cause of all.
The mechanism is simple. A page can rank very well for a search that has nothing to do with buying. "What is an LED wall" and "LED wall price" are two different worlds: the first brings students, the curious and competitors, the second brings people putting a budget together. If most of the clicks come from the first kind, the site is doing exactly what it's built to do and the conversion rate will stay low whatever you change.
You can check it in ten minutes. Open Google Search Console, Performance report, last three months, sort by clicks. Take the top twenty queries and mark next to each one whether the person typing it is buying or researching. If more than two thirds are research searches you've found the breaking point, and it isn't on the site: it's in the match between search intent and the pages you've published.
The fix isn't rewriting what's there. It's adding the pages that capture the commercial searches you don't cover today, and they're nearly always the ones the company avoids. For Macropix, a Milan manufacturer of custom LED screens, "ledwall prezzo" (LED wall price) gets 720 searches a month in Italy and has very strong commercial intent: most manufacturers leave it alone so they don't have to show their price lists. You answer it by explaining what the price depends on, without publishing figures. In the UK the demand is smaller but just as commercial: "led screen price" gets 90 searches a month and "led video wall price" 40 (Google Ads data via DataForSEO, October 2026).
The result of that work is documented in the full case study on SEO for manufacturing: "monitor pubblicitario" (advertising display) went from position 88 to position 2 between 2020 and 2025, and in July 2026 macropix.it held a 25.55% share of voice on the LED wall cluster, ahead of amazon.it at 18.08% (source: Semrush). None of the nine keywords that bring in the most enquiries contains the company's name.
If this is the part that applies to you, the work you need is called keyword research, not conversion rate optimisation.
On the page: benchmarks, not opinions
This is where the literature is full of design advice. The things you can actually measure are fewer, and worth more.
Speed. Google publishes precise thresholds: LCP within 2.5 seconds, INP within 200 milliseconds, CLS within 0.1, measured at the 75th percentile of page loads and separately for mobile and desktop. They aren't a matter of opinion and you can read them at no cost in PageSpeed Insights. Page speed matters precisely because it's the only item on the list with an official threshold.
Mobile. Mobile brings 70% of traffic and converts at 2.0% against 3.4% on desktop, with half the time per session. If the site was designed for the big screen and adapted afterwards, this item alone explains a good share of the gap.
The first page they see. A figure from the same benchmark that rarely gets quoted: one visit in three starts on a product detail page, and 61% of those visits bounce. The homepage is no longer the main way in. Plenty of sites have a polished homepage and inner pages that don't say who you are, why anyone should trust you or what happens when they click.
The pages to look at first are the three that get the most organic traffic, not the homepage. If they're landing pages built to convert, the work has already been done; if they're blog posts, what's missing is the next step inside the text. The rest of the routine maintenance belongs to on-page work.
At the exit: the point where they'd already decided
Baymard Institute keeps a collection of 50 independent studies published between 2006 and 2025, last updated in September 2025, and derives from them an average cart abandonment rate of 70.22%. Seven in ten people who put something in their basket leave.
These are the reasons they give, leaving out the 42% who were only browsing.
| Reason for abandoning | Share of abandoners |
|---|---|
| Extra costs too high (shipping, tax, fees) | 40% |
| Delivery too slow | 20% |
| Didn't trust the site with card details | 19% |
| Had to create an account to buy | 18% |
| Checkout too long or complicated | 17% |
| Website errors or crashes | 17% |
| Returns policy not clear | 13% |
| Couldn't see the total cost up front | 12% |
Six reasons out of eight have nothing to do with design. They're about what the site didn't say earlier and says at the last step. That's why showing delivery costs on the product page, rather than at checkout, moves the numbers more than a redesign.
In B2B there's no basket and the exit point is the form. Two changes pay off straight away: add to the form the fields your sales team asks for on the first call anyway (quantities, sizes, timings, budget), and say what happens after someone submits it. The first qualifies the lead and shortens the cycle; the second raises the submission rate, because hardly anyone fills in a form when they don't know what comes next. It's the same reasoning behind a B2B lead generation strategy.
Symptom, breaking point, report to open
| Symptom | Likely breaking point | Where to look | Benchmark |
|---|---|---|---|
| Lots of clicks, almost no enquiries | Upstream: wrong intent | Search Console, queries sorted by clicks | more than two thirds informational queries |
| The rate collapses on mobile only | Page: mobile experience | GA4, conversions by device | desktop 3.4% / mobile 2.0% |
| They land and leave from the same page | Page: speed or content | PageSpeed Insights, scroll rate | LCP within 2.5 s, scroll above 45% |
| They reach the basket and disappear | Exit: costs or friction | Checkout journey in GA4 | average abandonment 70.22% |
| The form gets opened but not submitted | Exit: too many fields or uncertainty | Session recordings | form opens vs submissions |
| Everything looks fine and it still doesn't convert | The offer or the price | Objections collected by the sales team | cost per lead vs margin |
How much traffic you need before you can say anything
This part is missing from nearly every article on the subject, and it changes the priorities. With 500 sessions a month and a 2% rate you get ten conversions. Taking them from ten to thirteen is a 30% improvement that no statistical tool can tell apart from chance.
Below a thousand sessions a month on a single page, A/B tests are no use: the sample is too small for the result to mean anything. Those are the numbers most small business websites live with. The practical consequence is that for those companies the lever isn't conversion rate optimisation, it's bringing in more of the right traffic (SEO and campaigns) and fixing the things you can see with the naked eye without having to prove them: speed, mobile, hidden costs, the form.
The time for serious CRO comes later. If you don't yet know which numbers you're looking at, start with which KPIs to measure and how Analytics is set up: a fair share of the sites that "don't convert" have broken tracking, and the conversions are happening but not being recorded. It's worth ruling that out before you rewrite a single line.
When the problem isn't the website
That leaves the case almost nobody puts in writing: the site is fine, the traffic is right, and it still doesn't convert. Then the problem sits upstream of marketing, and it's the offer or the price.
The quickest way to spot it is to compare acquisition cost with margin. On Macropix's Google Ads account, which runs in Italy, €16,805 produced 374 conversions between July 2025 and June 2026: €44.93 per lead, with an average CTR of 3.59% and a CPC of €0.81. There the sums add up, because a custom LED installation is worth many times that figure. On a product with a £30 margin, a similar cost per lead makes the channel unsustainable, and no change to the website will alter that.
It's a calculation to make before signing off a project, not after: the return on a digital activity depends more on the average value of a customer than on the conversion rate. If that's the missing piece, it's the work we start every lead generation project with.
One thing the industry rarely says: a conversion rate is a ratio, and ratios can improve by getting worse. A site that loses 30% of its traffic and keeps the same number of conversions shows a rising rate while the business takes less money. It has happened to plenty of sites over the past year, as informational traffic moved into AI-generated answers. Read on its own, month on month, that number is the most elegant way of not noticing anything: always look at it next to sessions and conversions in absolute terms. If your traffic is falling, that's the first thing to fix.
FAQs
It depends on device and channel before it depends on industry. In the Contentsquare 2026 benchmark desktop converts at 3.4% and mobile at 2.0%; paid search at 2.8% and organic social at 0.7%. For a B2B site that collects quote requests, the useful number isn't the percentage but the cost per lead compared with the average value of a customer.
Almost never as the first step. Rebuilding the site changes the second of the three breaking points and leaves the other two untouched: if the traffic comes from informational searches, or if delivery costs only appear at checkout, the new site will convert just like the old one. A diagnosis takes a few hours; a rebuild costs several thousand pounds.
Changes at the exit point (form fields, showing costs earlier, speed) can be measured within two or three weeks, if there's enough traffic. Upstream changes, meaning new pages to capture commercial searches, follow SEO timescales: the first movement in three or four months, the full result well beyond that.
Not directly, and expecting it to leads to disappointment. An article captures research searches, made by people who by definition aren't buying right now. Its job is to get found and pass the reader on to the right page: if an article has no link in the text to a relevant commercial page, the traffic it brings stops there.
Compare the conversion rate by channel in Analytics. If direct and returning traffic convert reasonably well and organic or social traffic doesn't, the site works and the problem is who's arriving. If every channel converts badly, it's the site.
