Rankings are not enquiries: what local search measurement can count, and what it will always miss

A leaflet, two advertising platforms and one telephone

In April 2026 I paid for a leaflet drop across the BD9 postcode area around my clinic in Bradford, and ran Google and Meta advertising in the same weeks. Within days the same telephone was ringing with people who had picked up a leaflet, people who had tapped an advert, and people who had searched for a clinic and found the Google Business Profile. Neither the phone nor the form announces which. To work out what each channel had produced I needed a QR code on the leaflet leading to a tagged address, a separate tagged link in the adverts, and a member of staff willing to ask every caller how they had found us.

That is the subject of this article in miniature. Local search work is sold on rankings, but a ranking is an intermediate state; the owner wants to know whether the work produced enquiries, which means counting them in a way that traces each to its source. Most of the tools are free and documented by Google; most of the gaps are permanent and need writing down. I built the clinic's website, Business Profile, analytics, tracking and consent banner myself during 2026, without an agency, while working as a pharmacist, and what follows separates what the platforms document from what I had to work out for one clinic.

What an enquiry is for a local business

The word covers several actions, each measured differently:

• A telephone call, dialled from a number on the screen, tapped as a click-to-call link on the website, or tapped as the call button on the Business Profile.
• A form submission on the website.
• A WhatsApp conversation started from a link on the site or from the chat option on the profile.
• A booking made through a booking system linked from either.
• A message sent through the profile's chat feature.
• A direction request from the profile, which is intent rather than an enquiry, though a walk-in never fills in a form.

Google's performance report for a Business Profile defines several of these. Calls are "the number of times a customer clicked on the call button on your Business Profile", website clicks are "the number of times people clicked on the website link on your Business Profile", and messages are "the number of different conversations through messages" ^[5]. Bookings are reported only when handled through a booking provider Google recognises ^[5]. Chat gives customers "a direct way to contact your business via text or WhatsApp" ^[19]. The definition of a call counts taps on a button, not calls received; a customer who reads the number and dials it from a landline is invisible to it.

Each action ends in a different inbox, and the method of assembling them into one monthly figure is the measurement.

Why a ranking report cannot answer the question

Search Console is precise about its metrics. Impressions are "how often someone saw a link to your site on Google", clicks are "how often someone clicked a link from Google to your site", and average position is "a relative ranking of the position of your link on Google, where 1 is the topmost position, 2 is the next position, and so on" ^[21]. None of these is an enquiry. A click is a visit, and a visit can end in a call, a form or a competitor's page.

I published the clinic's own Search Console data as a case study, with the export date, the period, the file fingerprints and the limitations written out ^[18]. It records clicks and impressions rising after the work began, says in terms that "correlation is still not causation", and notes that the website was new ^[18]. It contains no enquiry count; it is a record of search visibility, not of revenue. A ranking report can be entirely accurate and still say nothing about whether the phone rang.

The measurement chain, link by link

Google Analytics 4 is built on events, which "allow you to measure a specific interaction or occurrence on your website or app" and come in four kinds: automatically collected, enhanced measurement, recommended and custom ^[2].

Key events, and why the names matter

A key event, Google's term for what was called a conversion, is "an event that measures an action that's particularly important to the success of your business", and "any event you collect can become a key event" ^[1]. A key event can be imported into Google Ads as a conversion ^[1], so one definition of an enquiry serves both the analytics reports and the advertising account.

Google's recommended event names are not sent automatically because "these events require additional context to be meaningful" ^[3]. For a lead-generating business the relevant one is generate_lead, when a customer submits a form or otherwise provides contact details, with qualify_lead and close_convert_lead for the later stages ^[3].

Forms

Enhanced measurement, a set of switches inside the Analytics interface, records form_start "the first time a user interacts with a form in a session" and form_submit "when the user submits a form" ^[4]. It is the fastest way to get a form count and the least reliable, because it fires on the browser's submit event whether or not the form was accepted. A better form event fires only on a successful submission, from the form tool's own success signal or from a thank-you page that cannot be reached any other way.

Calls and WhatsApp from the website

A click-to-call link and a WhatsApp link are ordinary links, and Tag Manager's Just Links trigger tracks "clicks only on HTML links that use the a element" ^[10]. A condition on the Click URL restricts it, and a wait-for-tags option holds the browser until the event is sent ^[10]. A condition that the Click URL begins with tel: gives a call-click event; one that it contains the WhatsApp domain gives a WhatsApp-click event. Neither records whether the call connected.

Tagging the links you control

Analytics attributes a visit to a source using campaign parameters. Google's guidance is that "when you add parameters to a URL, you should always use utm_source, utm_medium, and utm_campaign", the values then appear in the Traffic acquisition report as session source and medium, and Google provides a Campaign URL Builder for constructing them ^[6]. The rules turning source and medium into a channel are published: Direct is a saved link or a typed address, Organic Search is a search engine source or a medium of organic, and Unassigned is "the value Analytics uses when there are no other channel rules that match the event data" ^[20].

This is what makes a leaflet measurable. A printed address with utm_source set to leaflet, and a QR code resolving to it, land in a channel of their own; without the tag the scan arrives as Direct, alongside regulars who bookmarked the site. The same applies to the Business Profile's website link, the one link most local businesses forget to tag.

What the Business Profile itself can tell you

The profile's report gives calls, directions, website clicks, messages and bookings, and Google has changed how it counts direction requests "to account for multiple taps, direction request cancellation, spam" ^[5]. These figures are Google's estimates, produced by rules Google can change, not a log the business owns.

The profile can now be connected to an Analytics property, so that "performance data from your Business Profile, such as interactions, website clicks, calls, directions, messages, bookings, and menu clicks, is shared with your Google Analytics property" ^[12]. That puts the profile's counts beside the website's without merging them: a caller who tapped both buttons is in both. Tagging the profile's website link remains the way to separate a visit from the map listing from a visit from a blue link on the same page; which channel it lands in depends on the medium chosen ^[20], so choose once.

Call tracking, and what it costs

The tools above count taps; call tracking counts calls. Google Ads offers call reporting through forwarding numbers, and "you set a minimum call length, and every call that lasts at least that long is counted as a conversion" ^[7]. Third-party services extend the idea to every channel by showing a different number to each source.

The trade-off is consistency. Google's guidelines for a profile say to "use a local phone number instead of a central call center helpline number whenever possible", that "the phone number must be under the direct control of the business", and not to provide numbers "that redirect or refer users to landing pages or phone numbers other than those of the actual business" ^[17]. A tracking number that forwards to the real line sits uneasily with that, and a business whose number differs across its website, profile and directory listings has traded a cleaner report for weaker consistency signals, at a monthly fee that a few dozen calls a month may not justify.

The cheapest instrument is a person. Asking every caller how they found the business is imprecise, because people say "Google" when they mean an advert, but it captures the channels no tag can see, and it was one of the three things I relied on to sort the leaflet drop from the adverts. On the phone it belongs in the script; on a form it can be a required field.

Consent removes a share of the measurement

Every link in the chain depends on a tag being allowed to run, and in the United Kingdom that is a matter of law first. The Information Commissioner's Office guidance on storage and access technologies says that "unless the technology meets one of the exemptions, you must seek consent before you use storage and access technologies", that "there is no valid reason to pre-enable non-essential storage and access technologies", and that users must be able to reject them "as easily as they can accept them" ^[9].

The Data (Use and Access) Act's changes to the cookie rules took effect on 5 February 2026, adding an exemption for cookies that solely collect information for statistical purposes and raising the maximum penalty under PECR from £500,000 to £17.5 million or 4 per cent of global annual turnover ^[16]. The ICO's reading of the statistical exception is narrow: it requires a simple means of objecting, and "the statistical purposes exception does not allow you to monitor or track individual visitors to your service" ^[8].

Whether an Analytics installation that records key events and passes them to Google Ads falls inside that exemption is a legal question I am not qualified to settle; the cautious reading is that attributing an individual's enquiry to an advertising click is not statistics. I implemented the consent banner, Tag Manager and Analytics on the clinic's site and on highregard.co.uk, and had to decide what loads before consent and what loads only after. The consequence of a compliant banner is that some visitors decline, and their calls, forms and bookings are not in the report.

How many decline depends on the banner. A study that scraped the five most popular consent platforms across the top 10,000 UK websites found only 11.8 per cent met the authors' minimal legal requirements, and its experiment with 40 participants found that removing the opt-out button from the first page increased consent by 22 to 23 percentage points, while granular controls on the first page reduced it by 8 to 20 percentage points ^[14]. A field study of more than 80,000 users on a German website found that "the wide-spread practice of nudging has a large effect on the choices users make" ^[15]. The missing share is therefore not fixed, and cannot be borrowed from someone else's website.

Google's answer is consent mode, which "lets you communicate your users' cookie or app identifier consent status to Google" ^[11]. When consent is declined, tags send cookieless pings and Google "fills the data collection gaps with conversion modeling and behavioral modeling" ^[11]. The modelling has entry conditions a local business will not meet: at least 1,000 events per day with analytics storage denied for seven days, and at least 1,000 daily users with consent granted for seven of the previous 28 days ^[22]. Below those thresholds nothing is filled in, and the gap stays a gap.

An enquiry is not a customer

Even a perfectly counted enquiry is half an answer. Google Ads' page on offline conversion imports puts it plainly: an advert may start a customer on a path that ends at the office or over the phone, and to record that the business must save the Google Click ID with the lead and later return it with the conversion type and date ^[13]. That needs a record linking the enquiry to its outcome, which for most small businesses means a spreadsheet. Channels differ in quality, and a monthly count of new customers by source, however roughly attributed, is worth more than any dashboard.

Then there is the share no tag will ever see. Word of mouth, a referral from another professional, a leaflet kept on a fridge for three weeks: these arrive as Direct, or as a search for the business's own name that Search Console records as a click like any other ^[21]. Crediting that click to search work is a mistake in one direction; ignoring the possibility that search work made the name recognisable is a mistake in the other. The honest report says which enquiries were traced, which were reported by the customer, and which were unattributed.

The afternoon setup

None of this needs a developer, though one is faster. It is, in essence, what I built for one clinic.

1. Write down what counts as an enquiry, one line per type, and decide now whether direction requests are in or out.
2. Install Tag Manager, and Analytics through it, behind a consent banner that blocks non-essential tags until consent is given, as the ICO guidance requires ^[9].
3. Turn on enhanced measurement and check the automatic form_submit against the form's inbox for a week ^[4]; if they disagree, fire the event on the form tool's success callback instead.
4. Create Just Links triggers for links whose Click URL begins with tel: and for WhatsApp links, with wait-for-tags on ^[10], and name the events with a recommended name where one fits ^[3].
5. Mark the form, call-click, WhatsApp-click and booking events as key events ^[1] and import them into Google Ads rather than defining them twice.
6. Tag every link the business controls with utm_source, utm_medium and utm_campaign using the Campaign URL Builder ^[6]: the profile's website link, booking links, directory profiles, any printed QR code.
7. Connect the Business Profile to the Analytics property ^[12].
8. Add one question to the phone script and one field to the form: how did you hear about us.
9. Keep one sheet, one row per enquiry: date, type, source as measured, source as reported, outcome.

The last two steps survive every platform change.

What this means for a local business

The question to ask any provider of local search work is not what the rankings are but how many enquiries were counted last month, by what method, and what the method misses. A provider with real measurement answers from the data; a provider selling activity answers with impressions and a request for patience. The setup above lets the owner check.

The baseline must exist before the work starts. Search Console keeps its own history; the enquiry record does not exist until someone starts it, and a provider who begins without asking for it has made their results impossible to check.

The figure will be incomplete, and the incompleteness should be stated: consent declines remove a share of website measurement that cannot be estimated from anyone else's site ^{[14] [15]}, profile calls count taps, not conversations ^[5], and word of mouth arrives unlabelled. A report that separates enquiries traced by tag, reported by the customer and unattributed is a better document than a single total attributed to SEO, and it is the format HighRegard uses for its own clients. If enquiries are the measure, a fault that breaks the form deserves more attention than a two-place fall on a keyword report; the site's article on detecting broken conversion tracking covers the case where the counter, not the customer, stopped.

What the evidence does not settle

Google documents what each metric counts, not how accurate it is. The direction request figure was revised to remove duplicate taps and spam ^[5], which concedes that earlier figures were inflated. There is no published error rate for profile calls, website clicks or messages.

The consent research measures how banner design changes acceptance, not what a compliant banner on a small clinic's website yields. The two studies drew on the most visited UK websites and one German site, in 2019 and 2020, under the EU rules of the time ^{[14] [15]}. Whether the February 2026 statistical exemption changes UK behaviour, or whether a typical local business's Analytics configuration even qualifies for it, has not been tested in public ^{[16] [8]}.

Nor does any source settle attribution. Google's offline import and enhanced conversions tools link a sale to an advertising click when the business supplies the identifier ^[13], but say nothing about the customer who saw an advert, forgot it, was recommended by a colleague and then searched the name. And the case study I published for my own clinic shows visibility rising, with the caveats written in, but does not report enquiries ^[18]. That is a gap in the evidence for the very argument this article makes, and readers are entitled to weigh it.

What remains uncertain

Two things could change the picture. If the ICO or a court reads the statistical exemption broadly enough to cover a standard analytics installation with an opt-out, the consent gap shrinks for UK businesses; read narrowly, it stays. And if Google lowers the thresholds for modelled data, which currently begin at volumes a local business does not reach ^[22], estimated enquiry counts will appear in small accounts, and owners will need to know that an estimate is what they are looking at.

What does not change is the arithmetic. An enquiry count assembled from the phone, the form, the profile and a member of staff asking the question, with its method written down, is imperfect in ways that can be described. A ranking report is imperfect in ways that cannot, because it was never measuring the thing the business wanted to know.