The state of small business websites in Croydon: 3,241 homepages checked
Original research. Between 11 July and 16 August 2026 I ran the same automated check over every small business homepage I could find in Croydon and the towns around it. About half of them tell a search engine nothing about being a local business.
- Of 1,934 live small business homepages inside the London Borough of Croydon, 1,003 (51.9%) carry no structured data identifying them as a local business. Search engines and AI assistants are left guessing at the name, address and hours.
- Speed is not the local weak point. 30.1% of Croydon homepages score under 60 on mobile, against 38.5% across the rest of the same sample.
- What Croydon does worse is looking alive: 15.8% of homepages have nothing newer than 2023 on them, and 25.3% link to no social profile and no Google Maps listing.
- The gaps are cheap to close. A meta description, a LocalBusiness block and a link to your own Google Business Profile are an afternoon of work, not a rebuild.
Between 11 July and 16 August 2026 I ran the same automated check over the homepage of every small business I could find a website for in Croydon and the towns around it. It came to 3,241 sites. This page is the result: how many loaded at all, how many are readable on a phone, how many are secure, how fast they are, and how many say anything that a search engine or an AI assistant can actually use to work out that the business is local.
Nothing here identifies a single business. Every figure below is a count or a percentage, and any group of fewer than 30 sites has been dropped rather than shown. The script that produces these numbers is kept so every figure can be recomputed.
The sample, in plain terms
Two sets of numbers appear on this page. The first is the London Borough of Croydon on its own: 2,054 businesses with a website, of which 120 (5.8%) had a homepage that did not load at all when it was checked, leaving 1,934 live homepages. That is the denominator for every borough percentage here, because if the page never loaded, none of the other checks could run.
The second is the wider area this studio covers, the borough plus 20 neighbouring towns from Banstead round to Wimbledon: 3,241 businesses with a website, 210 (6.5%) down, 3,031 live. The town and trade tables further down use that wider set, because the borough on its own does not give enough businesses per trade to report honestly.
About half of Croydon's homepages are invisible as local businesses
Of the 1,934 live homepages inside the borough, 594 (30.7%) carried no structured data at all. A further 409 (21.1%) carried some structured data, but none of it described a local business: no name, address, opening hours or service area in a form a machine can read. Together that is 1,003 homepages, 51.9% of the borough.
Structured data is a small block of machine readable detail in the page source. It is invisible to a human visitor and it is the single most direct way of telling Google, Bing and an AI assistant what the business is called, where it is, and when it is open. Without it, all of that has to be inferred from prose, and an assistant answering "who does this in Croydon" has nothing it can quote with confidence. It is not a ranking guarantee, but it is the difference between being described accurately and being guessed at. I have written separately about how AI search picks which businesses to recommend.
What else is missing
Share of the 1,934 live homepages in the London Borough of Croydon. Every check is defined at the bottom of this page.
The two that cost the most for the least effort are near the top. A missing meta description (465 homepages) means Google writes the snippet under your listing for you, from whatever text it finds first. No link to a social profile or a Google Maps listing (490 homepages) means the site is a dead end: nothing connects it to the Business Profile that actually appears in the local pack.
Further down, 305 homepages (15.8%) have nothing newer than 2023 anywhere on the page. A customer reading one has no way of knowing the business is still trading.
Speed is not the local weak point
Page speed is the thing small business owners are told to worry about most, and in Croydon it is not where the damage is. Google PageSpeed Insights returned a mobile score for 735 of the 1,934 live borough homepages. The median was 66.
Share of the 735 borough homepages that returned a mobile score. Scores are Google PageSpeed Insights, measured once per site during the audit window.
So roughly a third are slow enough to lose visitors, and one in ten is genuinely fast. That is worth fixing, but it is a smaller and more expensive problem than the missing half sentence of structured data above it.
What Croydon runs on
Share of the 1,934 live borough homepages. "Not identified" means the page source carried no platform fingerprint: hand built sites, custom systems and heavily customised installs all land here.
WordPress runs more of Croydon than everything else combined that could be identified. Of those, only 5 were on a WordPress version older than 5.0, so the old horror story of the abandoned WordPress 3 site is now rare. The more common problem is a current WordPress with nothing filled in.
Croydon against the rest of the sample
The same audit ran on 13,139 businesses outside this area, 11,836 of which had a live homepage. That makes a useful control: it shows which Croydon numbers are a local story and which are just what small business websites look like everywhere.
| Check | Croydon borough | Rest of sample |
|---|---|---|
| Live homepages measured | 1,934 | 11,836 |
| No structured data at all | 30.7% | 31.7% |
| Structured data, nothing identifying a local business | 21.1% | 20.6% |
| No meta description | 24.0% | 26.6% |
| No link to social media or Google Maps | 25.3% | 19.9% |
| Looks abandoned | 15.8% | 12.5% |
| On a drag and drop builder | 12.0% | 9.9% |
| Scores under 60 on mobile | 30.1% | 38.5% |
| Mixed content on a secure page | 11.7% | 13.8% |
| Not on HTTPS | 6.8% | 7.0% |
| Not mobile friendly | 4.4% | 4.3% |
| AI crawlers blocked in robots.txt | 3.1% | 3.3% |
| Homepage did not load at all | 5.8% | 9.9% |
Croydon sites are not slower, not less secure and no less mobile ready than the rest of the sample. They are more likely to look abandoned, more likely to sit on a drag and drop builder, and noticeably more likely to link nowhere else at all. The picture is of sites that were built once, often on a builder, and then left.
Town by town
Across the wider 3,031 live homepages, 25 towns had at least 30 audited sites and are shown below. Seven more had fewer than 30 and have been left out entirely rather than reported on a handful of businesses. Sorted by sample size, largest first.
| Town | Sites | No structured data | No meta description | On a builder | Looks abandoned |
|---|---|---|---|---|---|
| Coulsdon | 420 | 27.1% | 22.9% | 14.0% | 18.8% |
| South Croydon | 361 | 29.9% | 18.6% | 12.7% | 15.2% |
| Purley | 229 | 34.1% | 26.6% | 17.9% | 18.8% |
| Selsdon | 228 | 28.9% | 25.9% | 9.2% | 17.1% |
| Croydon | 168 | 32.1% | 25.0% | 11.3% | 11.3% |
| Worcester Park | 167 | 32.3% | 29.9% | 9.0% | 8.4% |
| Kenley | 143 | 33.6% | 27.3% | 13.3% | 15.4% |
| Streatham | 132 | 31.8% | 28.8% | 12.1% | 3.8% |
| Sanderstead | 129 | 27.1% | 24.8% | 8.5% | 14.7% |
| Crystal Palace | 113 | 29.2% | 25.7% | 7.1% | 9.7% |
| Beckenham | 112 | 24.1% | 23.2% | 14.3% | 14.3% |
| Norbury | 91 | 35.2% | 24.2% | 4.4% | 5.5% |
| Bromley | 81 | 34.6% | 19.8% | 9.9% | 18.5% |
| Penge | 58 | 27.6% | 12.1% | 5.2% | 8.6% |
| Addiscombe | 57 | 36.8% | 31.6% | 3.5% | 5.3% |
| Sutton | 53 | 18.9% | 20.8% | 9.4% | 15.1% |
| Thornton Heath | 52 | 34.6% | 28.8% | 9.6% | 15.4% |
| Orpington | 44 | 34.1% | 9.1% | 9.1% | 15.9% |
| Wallington | 44 | 18.2% | 15.9% | 2.3% | 11.4% |
| Wimbledon | 38 | 23.7% | 7.9% | 23.7% | 10.5% |
| Epsom | 35 | 34.3% | 28.6% | 5.7% | 28.6% |
| West Wickham | 34 | 32.4% | 17.6% | 14.7% | 8.8% |
| Shirley | 32 | 43.8% | 34.4% | 9.4% | 15.6% |
| Mitcham | 31 | 22.6% | 12.9% | 6.5% | 12.9% |
| Morden | 30 | 33.3% | 23.3% | 0.0% | 10.0% |
On a phone this table scrolls sideways; the town name stays put as you scroll. Twenty five towns had 30 or more audited sites and are shown. Seven had fewer, so I left them out rather than publish a percentage built on a handful of homepages.
Shirley and Addiscombe come out worst on structured data, Wallington and Sutton best. Epsom stands out for abandonment, with more than a quarter of its homepages showing nothing newer than 2023. Wimbledon has by far the heaviest use of drag and drop builders. These are small samples once you get past the top few rows, so treat the bottom of the table as a hint rather than a fact.
Trade by trade
The same 3,031 live homepages, grouped by what the business does. 39 trades had at least 30 audited sites; 36 more were suppressed for being too small. No town and trade figures are combined anywhere on this page, because that would start to point at individual businesses.
| Trade | Sites | No structured data | No meta description | On a builder | Looks abandoned |
|---|---|---|---|---|---|
| Restaurants | 241 | 31.1% | 32.8% | 17.8% | 12.0% |
| Pubs | 155 | 27.1% | 27.7% | 9.0% | 3.2% |
| Landscapers | 108 | 26.9% | 15.7% | 16.7% | 15.7% |
| Caterers | 98 | 25.5% | 18.4% | 11.2% | 16.3% |
| Carpenters | 96 | 29.2% | 24.0% | 14.6% | 17.7% |
| Roofers | 95 | 26.3% | 20.0% | 10.5% | 17.9% |
| Window cleaners | 89 | 38.2% | 23.6% | 11.2% | 13.5% |
| Scaffolders | 88 | 34.1% | 28.4% | 17.0% | 20.5% |
| Cafes | 87 | 36.8% | 31.0% | 13.8% | 10.3% |
| Kitchen fitters | 72 | 26.4% | 25.0% | 18.1% | 12.5% |
| Glaziers | 71 | 35.2% | 14.1% | 4.2% | 15.5% |
| CCTV installers | 71 | 31.0% | 14.1% | 8.5% | 23.9% |
| Charities | 67 | 41.8% | 58.2% | 16.4% | 11.9% |
| Skip hire firms | 66 | 21.2% | 21.2% | 3.0% | 18.2% |
| Takeaways | 64 | 43.8% | 42.2% | 3.1% | 4.7% |
| Bathroom fitters | 64 | 29.7% | 25.0% | 26.6% | 26.6% |
| Handymen | 60 | 20.0% | 26.7% | 20.0% | 23.3% |
| Dentists | 58 | 13.8% | 8.6% | 12.1% | 20.7% |
| Nail salons | 58 | 32.8% | 22.4% | 12.1% | 15.5% |
| Paving contractors | 54 | 22.2% | 22.2% | 9.3% | 11.1% |
| Plumbers and heating engineers | 50 | 24.0% | 12.0% | 12.0% | 22.0% |
| Barbers | 48 | 56.2% | 50.0% | 14.6% | 20.8% |
| Guttering firms | 48 | 20.8% | 16.7% | 12.5% | 16.7% |
| Funeral directors | 45 | 24.4% | 26.7% | 6.7% | 4.4% |
| Florists | 44 | 34.1% | 20.5% | 6.8% | 2.3% |
| Chiropodists | 43 | 32.6% | 20.9% | 11.6% | 18.6% |
| Dry cleaners | 42 | 33.3% | 31.0% | 7.1% | 14.3% |
| Fencing contractors | 42 | 35.7% | 28.6% | 4.8% | 23.8% |
| Taxi firms | 38 | 36.8% | 13.2% | 0.0% | 28.9% |
| Beauty salons | 37 | 43.2% | 29.7% | 16.2% | 18.9% |
| Photographers | 33 | 33.3% | 15.2% | 12.1% | 6.1% |
| Builders | 33 | 18.2% | 15.2% | 12.1% | 12.1% |
| Gardeners | 32 | 9.4% | 9.4% | 0.0% | 3.1% |
| Aerial installers | 32 | 25.0% | 3.1% | 6.2% | 28.1% |
| Mobile mechanics | 32 | 25.0% | 18.8% | 3.1% | 9.4% |
| Locksmiths | 31 | 41.9% | 16.1% | 3.2% | 12.9% |
| Driving instructors | 31 | 29.0% | 6.5% | 6.5% | 9.7% |
| Dog groomers | 30 | 23.3% | 16.7% | 16.7% | 10.0% |
| Plumbers | 30 | 23.3% | 10.0% | 16.7% | 0.0% |
Same again: scrolls sideways on a phone, trade name pinned. Thirty nine trades had 30 or more audited sites; 36 had fewer and are not shown. Trades are not crossed with towns anywhere in this study, because those cells would be small enough to point at individual businesses.
Barbers are the outlier in both directions that matter: 56.2% have no structured data and half have no meta description, on sites selling a service people search for by neighbourhood. Takeaways are close behind at 43.8% and 42.2%. Charities have the worst meta description rate on the list at 58.2%, which is a lot of local fundraising pages that Google is writing the snippet for. At the other end, gardeners (9.4%) and dentists (13.8%) are in much better shape, dentists mostly because their sites tend to be built by specialists.
One more figure worth reporting: 25.0% of florists were not mobile friendly, the highest of any trade, against an area average of 3.8%.
What to do if this describes your website
None of the most common faults on this page need a rebuild. In rough order of payback for the time spent:
- Add a LocalBusiness block to your homepage with your name, address or service area, phone number and opening hours. This is the 51.9% problem, and on most platforms it is one plugin setting or one block of code.
- Write a meta description for your homepage and your main service pages. Roughly 155 characters, written as an advert for the page, not a summary of your company.
- Link your website to your Google Business Profile and your social accounts, and link them back. That connection is how a search engine confirms the site and the listing are the same business. Verify your Business Profile first if you have not already.
- Put a current year somewhere real on the page: a recent job, a dated post, this season's opening hours. A page whose most recent date is 2022 reads as a closed business.
- Check the padlock. If your site loads over plain HTTP, or loads images over HTTP inside a secure page, Chrome will say so to every visitor.
- Only then worry about speed. It matters, but it is the most expensive item on this list and, in Croydon, the least unusual fault.
If you would rather see your own numbers than the borough average, the free website audit runs most of these same checks on one site and sends you the list. If you want to know why the site is not showing up at all, start with the eight usual reasons.
Method
- Every business in the sample had a website recorded and had that homepage successfully requested between 11 July and 16 August 2026. Each site was fetched once.
- The homepage only. No crawl of inner pages, so a site that carries structured data on a contact page but not the homepage is counted as missing it. That is deliberate: the homepage is what usually ranks and what an assistant usually reads first.
- Sites whose homepage did not respond are reported as a separate figure and then excluded from every other percentage, because none of the other checks could run on them.
- Page speed comes from Google PageSpeed Insights mobile scores. Not every site returned one, so the speed percentages use the 735 borough sites (and 1,121 across the wider area) that did, and say so.
- Groups smaller than 30 sites are dropped, not merged and not shown. Town and trade are never crossed with each other.
- Counts and percentages only. No business name, domain, phone number, email or screenshot appears here or in the working files behind it.
What each check means
- No structured data: the page contains no JSON-LD and no microdata of any kind.
- Structured data, nothing identifying a local business: some structured data is present, but none of it describes a local business entity with the details a directory or assistant needs.
- No meta description: there is no non-empty meta description tag, so search engines compose the snippet themselves.
- No social or Maps link: the page links to no social profile and to no Google Maps listing.
- Looks abandoned: the most recent year appearing anywhere in the page source is three or more years old.
- On a builder: the page carries the fingerprint of Wix, GoDaddy or a similar drag and drop builder.
- Not mobile friendly: no viewport meta tag, or a viewport pinned to a fixed desktop width instead of the device width.
- Not on HTTPS: the homepage does not finally resolve over HTTPS.
- Mixed content: a secure page loading images, scripts or stylesheets over plain HTTP.
- AI crawlers blocked: robots.txt disallows the crawlers used by AI assistants.
- Missing or generic title: the page title is absent or is a builder default such as "Home".
- Placeholder text left in: lorem ipsum, "coming soon" or "under construction" is still on the live page.
- No favicon: no icon is declared, so tabs and bookmarks show a blank square.
Limitations, stated plainly
- This is not a census. The sample comes from a prospecting database built to find businesses that might need a better website, so it is likely to skew towards weaker sites. Read every percentage as a description of this sample, not of Croydon as a whole.
- It is a single snapshot taken over five weeks in summer 2026. Sites change; some of these will already be fixed.
- Automated checks are literal. A site can pass every check and still be a bad website, and a site can fail three of them and still win the business.
- Trade labels come from how each business is categorised in the source data, so a firm doing two things sits in one row.
- Page speed was measured once per site, and a single PageSpeed run varies. The distribution is sound; an individual score is not.
- Towns are taken from the business address, not from a boundary lookup, so the borough figures follow the postal town.
The figures on this page may be reproduced with attribution to Jones Digital and a link to this page. If you are writing about this and want a breakdown that is not here, ask and I will run it, as long as it does not narrow down to individual businesses.
Sources
Common questions.
How many websites were checked, and where?
3,241 small business homepages across the London Borough of Croydon and 20 surrounding towns, checked between 11 July and 16 August 2026. Of those, 2,054 were inside the borough itself. 210 homepages did not load at all and are excluded from the other percentages.
What does "no structured data identifying a local business" actually mean?
Structured data is a block of machine readable detail in the page source that states the business name, address or service area, phone number and opening hours. 30.7% of Croydon homepages had none at all, and another 21.1% had some structured data that did not describe a local business. Together that is 51.9%, and it means search engines and AI assistants have to infer those details from ordinary page text instead of reading them.
Are Croydon business websites worse than average?
Not across the board. On the same audit, Croydon sites were faster than the rest of the sample (30.1% scoring under 60 on mobile against 38.5%) and no less secure or mobile ready. Where Croydon is worse is upkeep: 15.8% look abandoned against 12.5%, and 25.3% link to no social profile or Google Maps listing against 19.9%.
Which trades have the weakest websites?
Of the 39 trades with at least 30 sites in the sample, barbers came out weakest, with 56.2% carrying no structured data and 50.0% no meta description, followed by takeaways at 43.8% and 42.2%. Charities had the worst meta description rate of all at 58.2%. Gardeners at 9.4% and dentists at 13.8% had the least missing structured data.
Can I quote these figures?
Yes. The figures are published under a Creative Commons Attribution 4.0 licence, so they can be reproduced with credit to Jones Digital and a link to this page. If you need a cut that is not shown, ask, as long as it does not narrow down to individual businesses.