Is schema markup worth it for a small business?

Yes, but for a smaller reason than you have probably been told. Schema markup is worth checking, because it takes 15 minutes and often turns up something wrong or missing. Fixing it is cheap. What it is not, on the current evidence, is a lever that gets a small business named in ChatGPT or Google’s AI answers on its own. Treat it as basic housekeeping, then put your real effort into what the page says.

I ran this check live with a room of business owners at an AD:VENTURE AI SEO Bootcamp in Bradford on 21 October. It was planned as a five-minute demo. It turned into the most productive 20 minutes of the afternoon, and what they found is the best answer I have to the question.

On this page:

• What is schema markup?
• Does schema get you into AI answers?
• So why bother with it at all?
• What did a room of business owners find when they ran the check?
• How do you run the 15-minute check?
• When is it worth paying someone?
• FAQ

What is schema markup?

Schema markup is hidden code on a web page that tells search engines and AI tools what the facts on the page are. Your visitors never see it. A machine reads it and learns, for example, that this business is an accountancy practice and not a piece of accounting software.

The labels come from a shared vocabulary at schema.org, run jointly by the big search engines. There is a label for almost everything, from LocalBusiness, Product, Person and Event to very specific types of business such as AccountingService, LegalService and Electrician. Most website platforms add some of it automatically. In my experience, very few add the right ones.

In the room I put it this way: we need to not only write for a human now, we need to write for a machine. Schema is the most literal version of that.

Does schema get you into AI answers?

Not on its own, based on the best study we have. In a study published in May 2026, Ahrefs tracked 1,885 pages that added schema between August 2025 and March 2026, compared them against around 4,000 similar pages that didn’t, and measured how often each was cited by Google’s AI Overviews, Google’s AI Mode and ChatGPT for 30 days either side.

The author’s own summary was that “schema had no clear positive or negative effect.” Two caveats matter for a small business. Every page in the study was already heavily cited by AI, so it says little about a page AI has never seen. And the study only looked 30 days ahead. It is the best evidence available, and it is not the final word.

The other sources line up with it.

Sources checked 27 October 2026. On a phone, scroll the table sideways.

So why bother with it at all?

Because the cost of getting it wrong is bigger than the gain from getting it right. Schema is how a machine confirms what it thinks it has read. Missing labels leave it to work things out for itself. Wrong labels it simply believes.

That matters most in three situations I see again and again with small businesses.

First, when your business name is close to someone else’s. In my experience AI tools blend similar businesses together more often than people expect. Clear labels for your name, address, phone number and the profiles that belong to you give the machine something firm to hold on to.

Second, when your type of business has its own label. A generic “LocalBusiness” tells the machine very little. “AccountingService”, “LegalService” or “Electrician” tells it exactly what you do, and in my experience most small-business sites never use the specific one.

Third, when your website platform has added labels you never chose. Some themes and plugins mark every page up as an article, or describe a shop as an organisation with no address. That is the machine being told something untrue, by you.

None of that will make an AI tool recommend you. What gets you recommended is what the page actually says. Your experience, your proof, the questions your customers ask, answered clearly. Schema just makes sure the facts underneath aren’t fighting that.

Schema won’t get you recommended. Wrong schema can get you misunderstood.

What did a room of business owners find when they ran the check?

Most of the room came away with a specific label to add or fix. They were owner-managers and a freelance marketer, all with a live website or live client work, on everyday platforms such as a free site builder, Squarespace and an online shop platform.

Everyone checked a live site, in the validator or by asking an AI tool, and several did it on their own site without help. One delegate, a tradesperson, found with me that schema.org has a label for his exact trade. That turned schema from a technical idea into something with an obvious payoff, because most of his competitors won’t know it exists.

Another found the checker returned “no items” for her page, which is itself an answer. The page had no labels at all, and her AI tool then suggested the type to start with. And one owner found that on her shop platform the easiest route to the labels she wanted is a paid app. In my experience that is common, and it is worth knowing before anyone quotes you for the work.

What struck me most was the time. The block was planned at five minutes and ran to about 20, because I let it become a hands-on exercise and people kept finding things. I have since rebuilt the day so the check is a planned 15-minute exercise.

How do you run the 15-minute check?

You need a web browser and whichever AI tool you already use. Do it on your homepage first, then your most important service or product page.

Google also has its own Rich Results Test, which checks the labels Google uses for special listings. It is useful, but narrower. For this check the schema.org validator tells you more.

If you want to go further than labels, Fleek’s guide covers how to check what AI says about your business, which is the better test of whether any of this is working.

When is it worth paying someone?

Pay for help when the check finds a problem you can’t fix in your platform’s settings. That usually means one of three things: your platform needs an add-on or code to set the right business type, the validator shows errors on your homepage, or you have found an AI tool confusing you with another business.

That is normally a small, one-off job. You pay for it once, then check it after any big change to the site. It does not need a monthly retainer.

Be wary of anyone selling schema as a way into AI answers, or as a monthly service with ranking promises attached. On the evidence above, that is selling a different product from the one you’d be buying. If I was sat where you are, I would run the check myself, fix the one or two things it finds, and spend the money on making the page itself say more about why a customer should trust you.

That last part is where AI visibility is actually won, and it is the same work I covered in are buyers shortlisting you in AI before they call? The tools only amplify what is already there, which is why AI won’t fix your marketing on its own.

FAQ

Do I need schema to appear in Google’s AI Overviews?
No. Google says there is no special schema needed for AI Overviews or AI Mode. It does use structured data to understand pages, so accurate labels help it read you correctly, but they are not a requirement or a shortcut.

Does WordPress add schema automatically?
Not on its own. WordPress relies on an SEO plugin such as Yoast or Rank Math, and those add the basics, usually your site name, pages and articles. They rarely set the specific type for your kind of business, your address or your opening hours unless someone configures it, so run the check rather than assume it is done.

Is FAQ schema still worth adding?
Only if you already have a genuine FAQ on the page. Google stopped showing FAQ rich results for all sites on 7 May 2026, so the markup no longer wins you extra space in the results. Google says it still reads structured data to understand a page, so existing markup can stay, but nobody should charge you to add it.

How often should I check my schema?
After any website rebuild, theme change or new SEO plugin, because each of those can add, remove or overwrite labels without telling you. Otherwise once a year is enough. Run the validator on your homepage and your main service page, compare it with last time, and fix anything that has changed.

I am Jonny Ross, a Fractional CMO working with scaling SMEs across the North of England. I have spent 27 years in search and visibility, starting with my own online sunglasses business in 1999, and I have been running workshops for business support programmes since 2011. More about me.