Fewer than half of South African businesses on Google Maps link to a website of their own, audit of 4,366 listings finds

Gen1 Media, a search and web company based in Paarl, has audited 4,366 Google Business Profiles across five South African areas and four trades, finding that only 43.8% link to a website the business actually owns.

The study covers plumbers, electricians, dentists and attorneys within a 15km radius of Paarl, George, Sandton, Centurion and Umhlanga, spanning four provinces. It follows the agency's September study of AI search recommendations, which found a similar gap in a much smaller sample.

The headline numbers

Of 4,366 listings:

• 1,911 (43.8%) link to the business's own website.
• 111 (2.5%) link to something that is not their website: a Facebook page, a directory listing, a booking platform, or a page on a shared site builder.
• 2,344 (53.7%) link to nothing at all.

A simpler method, counting any linked domain as a website, would have reported 46.3%. That is the figure most audits of this kind produce, and it is wrong by two and a half points in the business's favour.

Broken down, the 111 are:

One Paarl plumbing business links to a directory profile filed under an entirely different company name. A Sandton dentist links to a third-party medical directory. A Paarl dental practice links to a salon-and-clinic booking platform.

"The Facebook ones I understand, and I would not be too hard on them," says Armin. 

Geography still does not matter

The finding that surprised the agency the first time survives the more careful count, and gets tighter.

Attorneys, not dentists, are the profession that stands apart

Grouped by profession, three of the four categories again land close together, and one falls away.

Plumbers, attorneys and electricians sit within seven tenths of a percentage point of one another. The assumption that professional services are ahead of the trades online does not hold at all.

Dentists are the weakest group, and the reason is visible in the listings themselves.

"A very large number of dental profiles are individual practitioners listed under a surname and an initial," says Armin. "No practice name, no site, just a doctor and a phone number. It is how medicine has always listed itself, and it works fine right up until the moment a machine is deciding who to recommend."

But the dentist gap has a second cause, and it points somewhere more interesting.

Among businesses actually managing their listing, attorneys pull away

Roughly two in five of the listings in the study have never been claimed by their owner. Those are entries Google generated automatically, and an empty website field on one of them may mean nothing more than that nobody has ever logged in.

Restricting the study to the 2,683 listings that have been claimed changes the picture in a way worth reporting:

And by profession, among claimed listings only:

"Two things fall out of that, and the second one I did not expect," says Armin.

"The first is that most of the dentist gap is unclaimed listings, not dentists. Once you look only at practices that have actually claimed their profile, dentists are in the middle of the pack."

"The second is attorneys. Among businesses managing their own listing, law firms are twenty points clear of every trade. That is the one genuine professional-services effect in the whole study, and it was hidden underneath the unclaimed listings the first time we looked."

Gen1 Media reports the full population as its headline figure because the customer's experience is identical either way: an unclaimed listing still appears in the results, and still has nowhere to send the person who taps it.

"Both numbers are true, and they answer different questions," says Armin. "Fifty-four per cent is what a customer runs into. Forty per cent is what businesses who are paying attention are still getting wrong. I would rather publish both than pick the one that sounds worse."

Nearly three in ten listings have never been reviewed

The audit also recorded whether each listing carries any reviews at all. Across all 4,366, 3,089 (70.8%) have at least one review, meaning 29.2% have never been reviewed by anybody.

This appears to contradict our own earlier study. It does not

In September, Gen1 Media published a study finding that when ChatGPT was asked to recommend local businesses, attorneys and dentists were linked to their own websites in eight of nine cases, while plumbers and electricians managed 15 of 37.

"On the face of it those two findings fight each other, and I would rather raise that myself than have someone else raise it for me," says Armin.

"They reconcile once you notice the two studies are counting different populations. The first one counted businesses the AI chose to recommend. This one counts everybody. So the AI is selecting out of the minority who have a real website, and the rest are simply not in the running. The claimed-listing numbers support that directly: attorneys are the best-equipped profession in the country on this measure, and they are the profession the AI linked most reliably."

What it costs the business

Gen1 Media argues the missing website matters less for ranking than for what happens after the customer has already chosen.

"You did the hard part. You came up," says Armin. "And then the platform handed your customer a menu of alternatives, because you gave it nowhere to send them. That is a lead you paid for in reputation and lost in one tap. And if your link goes to a directory, it is worse than nothing, because that page is selling your competitors on your behalf."

Method

Data was collected on 17 September 2026 via DataForSEO's business listings API, covering Google Business Profiles within a 15km radius of the centre point of each of five areas, filtered to four Google categories: plumber, electrician, dentist and lawyer. Claimed and unclaimed listings were both included; the claimed subset is reported separately above.

Each linked domain was then classified into one of three states: the business's own website, a page the business does not own, or nothing, by reducing the domain to its registrable form and matching it against a published list of social networks, South African and international directories, booking platforms and shared site builders. The classifier and the complete raw results are published with this release.

Limitations

• The radius flattens the difference between a town and a metro. A 15km circle around Paarl is Paarl. A 15km circle around Sandton takes in Germiston, Roodepoort and Kempton Park. The Gauteng and KwaZulu-Natal figures should be read as metro-wide rather than suburb-specific, which is why the study reports a rate per area rather than pooling everything into one national average.
• Sample sizes are very uneven, from 87 listings in Paarl to 2,328 in the Sandton radius, with 81% of all listings falling inside the two Gauteng radii. Rates are reported per area for that reason, and the pooled profession rates are correspondingly Gauteng-weighted. Individual category rates in the two Western Cape towns rest on small bases 11 electricians in Paarl, for instance and should not be read as precise.
• The classifier is a domain list, and domain lists are never finished. A directory not on the list is counted as the business's own website. The error therefore runs one way: the true "own website" figure is at or below 43.8%, never above it.
• A website that exists is not necessarily a good one. The study checks where the link goes, not what is on the other side of it.
• Categories are as Google assigns them, and a business listed under a different primary category will not appear.

A declared interest

Gen1 Media builds websites and sells local SEO, and the study's central finding is that South African businesses are short of both.

"We are not a neutral party here, and I am not going to pretend otherwise," says Armin. "We sell the thing this study says people are missing. That is exactly why the method, the radius, the categories, the classifier and the raw counts are all published. Anyone with a DataForSEO account can run it again tomorrow and check us. If the number comes back different, I want to hear about it."