What Actually Works in SEO in 2026:
And What Doesn't Anymore

The follow-up to what an SEO and AI search ready website actually is.

Now we rank the four pillars by impact on UK trade rankings in 2026.

We also name the dead tactics that still take money from trade businesses every week.

There are exactly four things that move the needle.

In this order.

Pillar One. Technical SEO Hygiene

The least glamorous and the most decisive.

Without technical hygiene nothing else compounds.

A site Google cannot crawl properly cannot rank, no matter how good the content is.

A site that is slow on mobile loses local pack positions to faster competitors with worse content.

A site without canonical tags gets duplicate-content folded.

A site without a submitted sitemap gets one page indexed and twenty pages ignored.

What technical hygiene actually looks like

Server-side rendering or static HTML.

Not a JavaScript single-page app where the content arrives after the page loads.

Every page has a unique title tag.

Every page has a unique meta description.

Every page has a canonical URL and a valid canonical link.

The page that ranks is the page Google can confidently identify.

Sitemap.xml that lists every URL on the site, submitted to Google Search Console, refreshed when pages are added.

Most trade sites have a sitemap.

Most of them are stale.

JSON-LD structured data on every page.

LocalBusiness or Service schema with areaServed, priceRange and serviceType filled in.

FAQPage schema on every page with FAQs.

BlogPosting schema on every blog post with datePublished and dateModified set correctly.

BreadcrumbList on every page.

Mobile loading under three seconds.

Core Web Vitals in the green.

No render-blocking scripts.

No cumulative layout shift.

Google's mobile-first index does not care that the desktop version is fast.

No nopagereadaloud and no notranslate meta tags.

The Google Discover SDK research from March confirmed that either tag removes the page from Discover entirely.

A developer or a template can add them accidentally and never know.

A site that gets all of this right looks plain underneath.

That is the point.

Plain is what Google rewards.

Plain is what ChatGPT can read.

What is dead. Stuffing keywords. Buying expired domains. Renting "1,000 links for £30". Anything from an article published before 2022 that mentions "do follow / no follow" as a tactic.

Pillar Two. Non-Commodity Content With First-Hand Experience

The second-biggest lever.

The one most trades think they are doing and almost none actually are.

Content that ranks in 2026 has three signals AI engines and Google both reward.

It is specific

Not "we cover the West Midlands".

Specifically: "We cover Wolverhampton, Tettenhall, Bilston, Penn, Wednesfield, and the WV postcodes including WV1 to WV14."

Google's entity resolution loves named places.

AI engines cite content with specific named locations far more than generic content.

It demonstrates first-hand experience

Not a Wikipedia paraphrase of "how heating systems work".

An actual job with an actual problem and the actual solution.

Photos of the work.

The make of the boiler.

The reason that combi was wrong for that flat.

Real depth that no template content can replicate.

It is structured to be answerable

An H2 question with the answer in the first sentence below it.

FAQ schema applied.

Speakable schema on the FAQ blocks for voice search.

This is the structure ChatGPT, Claude and Perplexity pull from when citing sources.

A page with all three of these is rare in the UK trade sector.

The ones that have it get cited in AI search results within weeks of being published.

The ones without it get indexed and ignored.

What is dead. Listicles with no original content. Service pages that list services as bullets with no detail. "We are passionate about delivering quality" anywhere. Stock photography. Any sentence that could appear on any competitor's site without a single word changed.

Pillar Three. Google Business Profile and Merchant Center Alignment

For local trades this is most of the work and most of the result.

Roughly 44% of clicks on a local search go to the local pack at the top of the page.

That is more than positions one, two and three of the organic results combined.

Winning the local pack is winning the search.

The current state of GBP signals matters more than the historical state.

Two new reviews this week beat fifty reviews from 2023.

Three new photos this week beat a hundred photos from 2022.

Recency dominates.

What actually works

The correct primary category for your trade.

There is one correct primary category for every UK trade, and the wrong one costs 60 to 80 percent of local pack visibility.

We covered the exact picks in the GBP categories lookup post.

We covered the fix for invisible profiles in why 90% of UK tradespeople are invisible on Google Maps.

Five new photos a month, minimum.

Phone photos of completed jobs are fine.

Recency is what Google measures.

A review every week or two, minimum.

The text of the review matters less than the velocity.

A consistent trickle beats a one-off batch.

Posts via the GBP Posts feature.

Weekly, brief, written about a specific job.

Most trades never publish a single one.

Service-area-business mode set correctly, with the actual areas listed.

Categories that match the website.

Hours that match the website.

Phone number that matches the website.

Address that matches Companies House if you have a registered office.

For trades selling products (boilers, doors, kitchens, windows), Merchant Center listings linked to the website.

Google now blends Shopping results into local searches for these categories.

Done properly across all of these, GBP delivers within 8 to 12 weeks.

Done in pieces, it delivers nothing.

What is dead. Auto-generated descriptions. Stuffing the business name with keywords. Asking for reviews via SMS that drops a star-only link. Any single-action "boost" that promises to fix GBP overnight.

Pillar Four. Real Digital PR

The slowest lever and the strongest one over twelve months.

Real PR is not paid mentions.

It is not guest posts.

It is genuine third-party validation that you exist, that you do what you say, and that real publications confirm it.

What real digital PR looks like for a UK trade

Companies House registration.

Free.

Most sole traders skip it.

Limited company structure costs around £50 to set up and produces an immediate authoritative external mention.

Trade body register entries.

Gas Safe for gas engineers.

NICEIC or NAPIT for electricians.

FMB for builders.

ARB for architects.

RICS for surveyors.

These are pulled by AI engines as source-of-truth signals.

LinkedIn Company Page.

Free.

Five minutes.

Adds a heavily-indexed external entity.

Local news mentions.

Genuine ones.

Sponsoring a junior football team.

Donating to a local charity.

Completing a job for a notable property.

Taking on an apprentice.

Local papers run these stories.

The mention is real.

The link is editorial.

Both signals matter.

National trade press mentions, when warranted.

Trade magazines run case studies on completed jobs.

Most never get pitched.

The trade businesses that do pitch them get cited in AI results because the trade press is in the training data.

Virtual office or proper registered address that allows directory listings on the directories Google still trusts.

Yell.

Bing Places.

Clutch (for B2B).

Trustpilot.

These are the directories AI engines actually pull from.

The combined effect of these is what builds the brand entity.

Google's brand entity recognition is the single largest reason the same five businesses appear at the top of every local search across an entire region.

What is dead. Buying links. "Niche edits" at £30 each. PBNs. Anything sold as "high DA backlinks". The site farms that produced these have been being demoted continuously since 2022, and the links now pass exactly zero authority.

Why This Order Matters

Technical hygiene is the floor.

Without it nothing compounds.

Content is what gets cited.

GBP is what feeds the local pack.

PR is what builds the entity over months.

Fix technical first.

It is the cheapest, most boring, and most decisive lever.

A site with clean hygiene and average content outranks a site with strong content and broken hygiene almost every time.

Fix content second.

Strong content is wasted without technical hygiene.

Technical hygiene without content is invisible to AI search.

Fix GBP third.

Pure GBP work can deliver in 8 to 12 weeks, but only if the website it links to is technically clean.

Fix PR fourth, because it takes longest and pays back longest.

Twelve months in, the businesses with real digital PR are uncatchable.

They sit at the top of every search because every signal in the system tells Google and ChatGPT they are the answer.

The Practical Sequence

Most trade businesses skip pillar one entirely because it sounds boring.

They invest in pillar two with generic content.

They half-do pillar three because GBP feels like extra admin.

They never touch pillar four.

The result is the pattern we documented in our research on 135 UK tradespeople across 11 towns.

Half the market is invisible.

The other half is fighting over the few who have done the work.

If you want to know which pillars your specific business is failing, the audit takes one afternoon.

We pull your site, your GBP, your citations and your schema.

We score each of the four pillars out of ten and tell you the order to fix them.

Frequently Asked Questions