Local SEO for 80 UK cities: what actually moves the needle
Postcode schema, Google Business Profile primer, regional citations, and why we don't bother with directory spam.
Across 80+ UK cities, the same five things move local-pack rankings. Most of them are free. We have shipped city-targeted pages for trades from emergency plumbers in Glasgow to wedding photographers in Bath, and the pattern repeats: the businesses that climb the three-pack do the basics relentlessly; the businesses that buy backlink packages stay on page three.
1. Google Business Profile, properly filled
Categories, services, photos every fortnight, posts every week. Most local-pack failures come from a half-finished profile, not a half-finished website.
What "properly filled" actually means
A complete GBP listing has: a primary category that matches your single highest-revenue service (not the broadest), three to five secondary categories that match your other paid services, every service listed individually with a short description and a price-from where allowed, opening hours including special closures for bank holidays, the service area as a list of named towns and postcode districts rather than a generic radius, at least one cover photo and ten interior or process photos, the website URL with UTM parameters that let you measure GBP traffic separately in GA4, and a populated Q&A section where you answer the obvious questions yourself before a competitor does it for you. Most of the listings we audit have three out of those nine. Closing the gap typically moves a business two to four places in the local pack within 30 days.
2. LocalBusiness + PostalAddress schema
Every city page on every site we ship includes a ProfessionalService schema entity with a UK postcode. Google verifies the postcode against the GBP listing — if they match, you ship into the local pack faster.
Schema in detail — what to ship, what to skip
On a city landing page we always emit: LocalBusiness (or the most specific subtype — Electrician, Plumber, HairSalon, etc.); PostalAddress with the full street, locality, postal code, and country; GeoCoordinates with lat/lng accurate to four decimal places; OpeningHoursSpecification rather than free-text "open 9-5"; AreaServed naming the city and surrounding towns; Service entries for each individual offering with a connected Offer where pricing is fixed enough to publish; AggregateRating fed from the GBP review count and average (refresh quarterly); and Review entries for the three highest-quality customer quotes. We skip Person markup for the owner unless they appear on the page with a photo and a bio — fake Person entities are an E-E-A-T penalty waiting to happen.
3. Regional citations that already trust the postcode
Yell, Bing Places, Apple Business Connect, Bark, plus the regional chamber of commerce. Five quality citations beats fifty cheap ones.
The citations that actually move the needle in 2026
Tier one (build first, every time): Google Business Profile, Bing Places for Business, Apple Business Connect, Facebook business page, Yell.com, Yelp UK, TripAdvisor where relevant. Tier two (build next, in week two): your industry's vertical directory — Checkatrade or TrustATrader for trades, Bark for services, Open Table for restaurants, Wedding Wire for the wedding trades, Vetted for childcare. Tier three (chase in months 2-3): your regional chamber of commerce, your local business improvement district directory, the local newspaper's business listing, any local-authority procurement supplier listing if you sell to councils. NAP consistency across all three tiers is the unglamorous bit that matters: the same business name, the same telephone in the same format, the same address punctuated identically. We run a NAP audit at launch and again at month six.
4. Reviews, on Google, with real names
A 4.6 average over 60 reviews outranks a 5.0 over 8. Ask every customer, on the day of the work.
The review request flow that actually works
Same-day, branded, friction-free. Same-day: ask while the work is fresh — wait a week and the rate halves. Branded: not a generic "leave us a review" link, but a personal text from the engineer or technician with their first name and a short line about the specific job. Friction-free: a short URL (yourdomain.com/review) that 301s straight to your GBP review form so the customer is one tap from the five-star button on iOS or Android. We build that redirect into every site we ship. Conversion rates we measure: 24-35% of texted requests turn into a public review, against 4-9% for emailed follow-ups. Replying to every review — positive and negative — within 48 hours signals an active business to Google and visibly to other customers reading the listing.
5. One landing page per city you actually serve
Generic "we cover the UK" pages don't rank for "[trade] [city]". A page with the city in the H1, the postcode area in the schema, and one paragraph of locally relevant copy ranks.
What "locally relevant" looks like in copy
Not "We serve customers all over Manchester." That is doorway copy — generic, swappable, valueless. Locally relevant means: the named neighbourhoods or postcode districts you actually cover (Didsbury, Chorlton, Salford, M20, M21, M5), the local quirks that affect the job (the parking restrictions in Northern Quarter for a removals firm, the building stock age in Chorley for a plumber, the council's street-trading rules in Brighton for a food trader), the local landmarks you can be measured against (under five miles from MediaCity, between J1 and J3 of the M602), and a real testimonial from a named customer in that city. We aim for 1,700-2,500 words per city page when the trade and the market justify it; for thinner trades we still write 1,000+ with unique local statistics rather than rehashed national filler.
The traps that look like local SEO but aren't
Three things owners spend money on that we have never seen move a ranking: paid citation submission services (most of the directories they push to either no-index you or pass no authority); generic "guest post" packages on PBN-style sites (Google's 2024 link spam update eats these for breakfast and the rankings vanish in the next core update); and AI-generated "expanded city pages" that take the same 200-word template and substitute the city name eighty times. The latter triggers the Helpful Content classifier in 2026 — Google now reliably detects "wide but shallow" doorway page sets and either de-indexes or applies a domain-wide trust penalty. If you cannot defend the unique value of each page to a sceptical reader, don't ship the page.
A 30-day plan for one city
Week one: GBP audit, NAP audit, baseline measurement of current pack position with a tool that geo-emulates the postcode. Week two: GBP rebuild, top-three citation rebuild, schema deployed on the relevant landing page. Week three: review request flow live (texts to the last 30 customers), first GBP post, two photos uploaded. Week four: measure, log the delta, schedule the ongoing cadence (one post a week, two photos a fortnight, NAP recheck quarterly). Most clients see a measurable lift inside 21 days; a few stubborn niches take 60-90 days because the GBP trust signals need time to compound.
What we don't bother with
Directory submission services, paid backlink packages, comment-section spam. None of it has moved a single client past page two in three years of testing.
Measurement: how to actually tell if local SEO is working
Three rank-tracking signals matter for a UK local business and they need different tools. First, three-pack visibility for your money keyword in a geo-emulated search from the customer postcode area — Local Falcon and BrightLocal both do this well; we use a 25-point grid centred on the business address and watch the heat map shift monthly. Second, organic ranking for the "[service] [city]" long-tail query, measured from a UK IP — Ahrefs, SEMrush or the free Search Console performance report all do the job. Third, GBP discovery traffic — the impressions and clicks tab inside Google Business Profile Insights, which Google has not killed despite the threats. Watch all three. The three-pack and the GBP traffic move together when the foundational work is right; the organic ranking lags by 6-10 weeks because Google needs more crawl signals to update the regular-results algorithm.
What changes between a service area and a storefront
A business with a physical address visible to the public (a shop, a salon, a clinic, a restaurant) plays a different local-SEO game from a business that visits customers at home (a plumber, a mobile groomer, a cleaning service, a wedding photographer). Storefronts can list the address publicly, get reviews tied to the specific location, and rank in the local pack on the basis of proximity to the searcher. Service-area businesses must hide the address (GBP requires this if you don't serve customers at the address), list service areas as named towns or postcode districts (Google allows up to 20), and lean harder on category accuracy and review velocity because proximity matters less. We treat the two cases distinctly in the schema (LocalBusiness for storefronts, ServiceProvider with serviceArea blocks for SAB) and in the page architecture (storefront pages link prominently to a map and opening hours; SAB pages lead with the coverage list and the response-time guarantee).
The role of AI search in 2026 local results
Google AI Overviews and the SGE-style answer experiences have changed what shows up above the traditional local pack on roughly 18-22% of commercial-intent UK queries as of early 2026. The good news is that the citation criteria for AI Overview inclusion overlap heavily with the criteria that drive local-pack rankings — clear LocalBusiness schema, named author or business entity, citations to primary sources, content depth that resolves the user's question rather than just teasing it. Sites that have done the foundational SEO work get cited in the AI summaries without separate optimisation. The thing that hurts citation rates is thin content and AI-generated body copy; both trigger the Helpful Content classifier and quietly remove the page from the citation candidate set. Doing the basics well covers traditional ranking and AI citation simultaneously. ChatGPT, Perplexity and Claude's search experiences cite by the same rough rules — clear authorship, verifiable claims, structured data — so the work pays its way across every AI surface, not just Google's.