By Sara Khan · WeProms Digital · October 22, 2026
Every month, a well-meaning agency emails a Pakistani-owned mithai shop in Leeds a PDF listing exactly where the shop appears on Google for forty chosen keywords, and the owner forwards it to the family WhatsApp group as proof that the marketing budget is doing its job.
In October 2026 that PDF quietly stopped describing anything real. Google changed the code behind its results pages in ways that broke the rank-tracking tools most agencies rely on — Search Engine Journal counted “dozens, if not hundreds” of affected tools — while delays spread through Google Search Console at the same time, weeks after the September 2026 spam update closed. A report built on numbers the industry itself cannot currently collect is not measurement. It is a monthly ritual, and the ritual was already describing a results page that shoppers in Leeds stopped seeing years ago.
The report describes a Google that no longer exists
Rank tracking — software that checks where a page sits for a keyword, then graphs that position over time — has been the backbone of small-business SEO reporting for over a decade. That backbone cracked in October 2026. Google moved to stop third parties scraping its results, and the tracking tools either returned stale positions or failed outright; Search Engine Journal describes “a breakdown in the core data supply chain”, with Search Console’s own data processing running late in the same weeks. An agency that cannot see the results page still sends the graph. The shop pays for it — £3,120 a year is a realistic illustrative figure for retainer-plus-reporting in West Yorkshire — and receives numbers nobody can currently verify.
“A traditional keyword rank report now fails to describe the actual user experience in an accurate or meaningful manner.”
That is not a critic speaking. It is Search Engine Journal, the trade publication those agencies themselves read, publishing in the same month the reports kept going out. The pattern repeats down the whole chain: the tool cannot see the page, the page no longer matches what shoppers see, and the PDF prints a position number with a precision the data no longer supports. A rank report is last Ramadan’s ledger — exact to the pound, faithful to a month that has passed, and silent about tomorrow’s rush. No sweet-shop owner stocks Eid trays off last year’s ledger, and running a 2027 marketing budget off a broken position number is the same decision in different clothes.
There is a cheap test for any report you are paying for right now. Ask the agency one question: which of these positions did a human verify by hand, from Leeds, this week? A good answer names a browser, a postcode, and a date. Anything else is a screenshot of a machine guessing.
One search, a dozen different results pages
The deeper problem survived the outage, because the outage was never the real issue. Two people searching “halal sweets Leeds” at the same moment see different pages — one searching from Headingley, one visiting from Bradford, one on a phone that already knows their usual barfi order. Position is not a fact about a shop; it is a fact about one query, one device, one place, and one moment, which the report then averages into a single number matching nobody’s actual screen.
That is the whole problem in one sentence.
Above the traditional list sits everything the report ignores. An AI Overview — Google’s generated answer, assembled from sources Google picks — appeared on roughly 13% of US queries according to Semrush’s tracking; that is a US figure, it does not prove UK prevalence, but it marks the direction of travel for every market Google serves, including Yorkshire. When such a summary appears, clicking collapses: Pew Research Center measured clicks on a traditional result in 8% of visits when an AI summary showed, against 15% when it did not. Pew’s study is a US panel, so it does not describe Leeds shoppers specifically; it describes the results page Google is shipping to everyone, which is the page your money phrases now live or die on.
The map pack — the boxed trio of nearby businesses with star ratings and directions — sits above the organic list for the phrases that carry actual money for a sweet shop: “mithai near me”, “fresh barfi Leeds”, “eid sweets open late”. Many of those searches now end with no click at all, a zero-click search, because the phone number, the hours, and the directions live on the results page itself. A report claiming position 4 for “mithai Leeds” has not told you whether the searcher ever reached a list, a map, or a shop door. It has told you where a sampling machine placed an average page that no individual shopper saw.
The practical version of this section costs minutes and no agency fee. Search your five money phrases from an incognito window inside the shop, then again from a different postcode, and screenshot what sits above your listing each time — the ads, the map pack, the AI answer. That folder of screenshots is your real results page, and it reframes the question from “where do I rank” to “what does a shopper have to scroll past before they can choose me”.
The free audit tools grade the wrong exam
The industry’s answer to all this uncertainty has been a wave of free AI visibility audits: enter your web address, receive a score. Orbit Media reviewed 100 of these audits and found the scores for one single website ranged from 28 to 100 across the tools it tested. The same shop, the same pages, the same month. Fifty-eight percent of these tools generate their own test prompts — they scan the site, invent questions from its content, and then grade the site on answers to questions no buyer ever asked.
“AI visibility scores are arbitrary to the point of meaninglessness.”
That conclusion is Orbit Media’s, and it lands harder on community businesses than on anyone else. The aunty ordering barfi boxes for a nikkah in Batley asks a different question than a student craving gulab jamun after a night out in Headingley; she asks about deposits, freshness on the wedding day, delivery to the venue, and whether the shop is certified halal. An audit that has never met her cannot measure whether she finds you — only whether a machine liked the metadata. What actually drives this is the gap between those two things, and that gap is precisely where sweet shops win or lose shaadi season.
The rule worth keeping is simple. Before accepting any audit, free or paid, ask what it already knows about your buyers; if the answer is a pause, the score is decoration.
The numbers that actually feed a sweet shop
Some numbers in a shop’s week are real, because they count people and money rather than positions. Google Business Profile — the listing that powers the map pack — reports calls, direction requests, and booking clicks directly on Google’s own performance page, and a rise in direction requests on a Friday tracks jummah footfall more honestly than any keyword graph ever will. Google Search Console’s Performance report counts actual clicks from actual searches, including the AI-era queries that rank trackers cannot reach. WhatsApp order messages, deposit totals, and repeat-collection rates close the loop the PDF never touches. Community channels carry more of this demand than owners assume: Neil Patel’s figures put 1.8 billion people inside Facebook Groups every month, and Leeds mithai demand has always moved through wedding groups, mosque noticeboards, and family forwards as much as through search boxes.
Redirecting that illustrative £3,120 a year changes what those numbers can do. At an average order of £28 and roughly £11 contribution after ingredients, packaging, and boxes, £3,120 equals about 284 orders’ worth of margin — spent on the report, or spent on the channels that count. The seasonal calendar sets the checkpoints: Ramadan is expected to run from around 8 February 2027, with Eid al-Fitr around 10 March 2027, both subject to moon sighting; that makes December 2026 the month to build the early-interest list, late January 2027 the deposit window, and the last week of Ramadan the final-order deadline for Eid trays; the number to watch in each phase is pre-orders counted and deposits banked, never position held. WeProms Digital, an SEO agency trusted by halal businesses in the UK and Gulf, builds reporting around exactly those counts for sweet shops and grocers, because a graph of positions cannot be cashed at a counter. Reputation belongs in the same ledger — a Manchester curry house that won awards and still trailed rivals on Google reviews shows which signals the map pack answers to — and so does the new demand layer: owners who measure which AI searches send customers — a method written for a Chicago grocer, though it transfers — can see the new arrivals before their competitors do. Retention finishes the job, because customers who return for occasions rather than stamp cards outvalue any ranking the report celebrates.
Where rankings still earn their keep
None of this says positions are worthless, and the honest version of the argument is narrower. For a shortlist of high-intent local phrases — “mithai near me”, “fresh barfi Leeds”, “halal cake shop Leeds” — being visible in the map pack, and being named inside AI answers, still moves real money; Search Engine Journal’s proposed replacement, tracking how often engines serve and cite a brand across every surface, is the professional form of exactly that shortlist. The defensible claim is the one worth stating directly: for a single-location shop, a monthly forty-keyword position report is now harmful rather than merely useless, because it spends real money and the owner’s scarce attention on a number that no longer describes any page any shopper sees. In a family business, attention spent looking backward is the most expensive line item nobody budgets for.
Measure demand captured, not position held. Count the calls, the direction requests, the WhatsApp orders, the deposits banked, and the times an AI answer or a map listing named the shop by name; spend money only where those counts can rise, and let the forty-keyword PDF rest in the drawer where last Ramadan’s ledger belongs. A position number cannot be cashed, called, or served at a nikkah — the counts can, and in a shop on a Leeds high street, that is the entire argument.
Read next: Paper Stamp Cards Don’t Bring Mithai Customers Back. Occasions Do. · How to Turn Instagram Reels Into Mithai Orders This Shaadi Season
We have worked with mithai shops, halal grocers, and takeaways across the UK and the Gulf, and based on our experience the first thing worth replacing is the ranking PDF, not the website. If your agency still reports positions, ask us for the measurement swap: a dashboard built on calls, direction requests, deposits, and AI citations. Reach us through our contact page, email hello@weproms.com, or WhatsApp +92 300 0133399 — we reply in UK and Gulf business hours.
Sources & References
- Search Engine Journal — Rank Tracking Broke In September: The Future Is Measured By Brand & Revenue — October 21, 2026
- Orbit Media — We Audited 100 AI Visibility Audits. Here’s What They Miss. — October 21, 2026
- Pew Research Center — Google users are less likely to click on links when an AI summary appears in the results — July 22, 2025
- Google Search Console Help — Performance report (Search results)
- Google Business Profile Help — Understand your Business Profile performance & insights
- Semrush — AI Overviews: What They Are and How to Win Them
- Neil Patel Blog — How to Use Facebook Groups for Marketing — October 21, 2026
Additional reading from industry feeds:
