Stop chasing search volume: rank Glasgow keywords by actual pounds

Why search volume is a bad prioritisation metric

The Glasgow content marketer’s workflow, near-universally: pull a spreadsheet from Keyword Planner, sort by volume, filter on difficulty, write briefs for the top thirty. Ship them. Sit back. Six months on, the traffic chart is nudging up, the leads chart is flat, and the finance director wants a word. She isn’t angry. She’s just done the maths. Volume-first is a comfortable methodology because the number is huge and easy to defend. It’s also weakly correlated with revenue. “What is SEO” pulls a mountain of clicks and pays nobody’s rent. “Chartered surveyor for tenement conversion West End Glasgow” barely registers on Keyword Planner — and one client of ours signs three of those a month at £4,800 each. The graft’s in finding the second kind.

The reason is that search volume measures how many people typed a query, not how many of them will part with money. “What is SEO” has enormous volume and zero commercial intent. “Chartered surveyor for tenement conversion West End Glasgow” has a search volume so low that Keyword Planner rounds it to zero — and yet one client of ours signs three of those a month at £4,800 each.

How Semalt’s AI models revenue potential

Semalt runs a two-stage model. The first stage predicts, from a term’s surface form, its intent bucket — navigational, informational, commercial, transactional, local — and its likely conversion rate given the vertical. The second stage cross-references that intent against your own historical data (imported from Search Console, GA4, WooCommerce, Shopify, HubSpot or a webhook) to produce a GBP number: expected revenue per hundred organic sessions, at your current site’s conversion pattern.

The result is a ranked list where a term with 90 monthly searches, a 4.2% predicted conversion rate and an average order value of £340 outranks a term with 12,000 monthly searches, 0.15% conversion and £18 AOV. The scoring is transparent — every prediction ships with the four variables that produced it.

Scottish English vs. generic UK English intent

A common failure mode of generic keyword tools is treating “messages” the same in Glasgow as in London. In Scotland the word carries the sense of “grocery shopping”, and the intent on “doing the messages” is emphatically local, transactional and food-related. In London it is about SMS.

Semalt’s Scottish English pack captures roughly 180 such dialect nuances, from “piece” (packed lunch) to “close” (tenement stairwell) to the specifically Glaswegian “munter” that no B2B copywriter should ever use in a brief. When a Glasgow user searches “best close cleaning Southside”, the engine understands this is a service query about tenement stair cleaning — not a query about “shutting things down cleanly”. That single intent correction shifts a service company’s content brief entirely.

Seasonal patterns: TRNSMT, Six Nations, Old Firm, Christmas markets

Glasgow’s search calendar has spikes that generic UK data smooths away. TRNSMT festival in July pushes “hotels Glasgow” up by 340% year-round-average, but only in the four weeks leading up to the event. Six Nations rugby matches at Murrayfield push queries for “park and ride Glasgow to Edinburgh” up 180% on match weekends. The Old Firm derby (four times a year, sometimes more with cup ties) triples queries for “pubs Glasgow showing football” on match day itself, with a shorter run-up.

Semalt’s seasonal engine detects these patterns automatically from three years of Google Trends data blended with the platform’s own SERP crawl history. It surfaces “prepare now” alerts — if you are a Glasgow hotel and your TRNSMT landing page has drifted off page one by mid-May, the alert will tell you to fix it now, not in July when the spike is already burning. Even legacy queries like “T in the Park” (last held in 2016) still generate 1,900 UK searches/month, mostly nostalgic informational intent that a Glasgow music venue can capture with the right content.

Postcode-level local pack behaviour

Glasgow’s postcode geography is dense. G1 through G5 alone contain a dozen commercial neighbourhoods that behave differently in the local pack. A user in G3 (Finnieston) typing “coffee shop near me” will see a completely different map pack from a user two miles east in G31 (Dennistoun). Generic keyword tools flatten this. Semalt does not.

The keyword engine simulates Google searches from IP addresses geolocated to each Glasgow postcode district, and reports the map pack composition for every commercial query. For a Southside dental clinic considering whether to write a “dentist near me” landing page, Semalt will show that the term is unwinnable citywide but very winnable within G41 and G42 with the right Google Business Profile setup and citation strategy. Open the postcode simulator and run your own Glasgow query.

Case study: West End Bakery, from 40 to 400 orders/month

A small independent bakery on Byres Road came to us in October 2025 with a hard target: quadruple online orders from their Shopify site within six months. Their previous agency had built a content plan around volume: “bakery Glasgow” (2,900 searches/month), “bread Glasgow” (1,600), “cakes Glasgow” (4,400). Twenty-eight blog posts and eight months later, orders had grown from 40/month to 52.

Semalt’s AI took an evening to re-rank the entire keyword universe by predicted GBP revenue using the bakery’s actual AOV (£38) and CVR (2.1%). The top of the list looked very different:

• “Sourdough delivery West End Glasgow” — 70 searches/month, 5.4% predicted CVR, priority score 91.
• “Birthday cake Byres Road same day” — 110 searches/month, 7.8% predicted CVR, priority score 96.
• “Vegan brownies delivered Glasgow” — 40 searches/month, 6.1% predicted CVR, priority score 87.
• “Corporate cake order Glasgow city centre” — 30 searches/month, 3.2% predicted CVR, AOV £220, priority score 94.

Twelve new landing pages, one Google Business Profile revamp and a six-week AutoSEO cycle later, orders were at 412/month by April 2026 — a 10× jump from where the “volume-first” strategy had left them.

Semalt AI vs. Ahrefs vs. Semrush vs. Google Keyword Planner

More in this series: The Glasgow traffic GA4 never shows you, SEO changes in 3 minutes, not 6 weeks, Reading your Glasgow rivals' playbook.

FAQ

Does the AI need my sales data to work?

No, but it works much better with it. Without your data, Semalt uses vertical averages to project revenue. With your AOV, CVR by category and page-level Search Console data, projections tighten to within ~15% of realised revenue over an eight-week window. Connect your data in the dashboard.

How is the Scottish English pack maintained?

Human-in-the-loop. Our Glasgow-based editors flag new dialect variants they see appearing in Google Trends or in customer briefs. The classifier is retrained every six weeks.

Does it work outside Glasgow?

Yes. The same engine has UK-wide postcode simulation, and separate dialect packs for London, Newcastle, Cardiff, Manchester and Belfast. For non-UK markets, English (US, IE, AU) and eight European languages are supported.

Can I export the list to a brief document?

Yes. Every ranked keyword generates a one-click content brief including target intent, recommended H2s derived from People Also Ask, competitor URLs, and a suggested internal linking pattern. Briefs export to Docs, Notion, or plain markdown.

Does the seasonal engine cover B2B?

Yes. B2B verticals have their own seasonal patterns (fiscal-year-end March in the UK, budget-planning October, procurement freeze December) and Semalt models them separately from consumer verticals.

How does it connect to the rest of the platform?

The revenue-ranked keyword list is the primary input into AutoSEO, and it feeds the striking-distance dashboard in Semalt Analytics. One prioritisation feeds the whole toolchain.