Analytics and tracking that works
Tags get installed, dashboards get built, and nobody can answer the one thing the owner wants to know: which of the things we pay for brings in customers. The fix starts with the question and works back to the tag.
A shop has an advantage most businesses lack: an order record. Even so, the ad platforms, the analytics tool and the shop each report a different number, and the argument about which is right goes on every month instead of the decision it was meant to inform.
Three minutes, twelve questions, no email needed to see or copy your result.
What we set up and check
Judged against your own records, not against any tool's dashboard.
Where we come in
Tracking rarely breaks all at once. It drifts, and the first sign is an argument about whose number is right. Usually it is one of these.
Measurement for a shop goes wrong in a small number of well-known ways. Usually it is one of these.
A form submission, a phone call, a booking and a brochure download are four different things, and the reports treat them as one. Until somebody decides which of them is worth money, the totals mean very little.
For many local businesses the call is the main enquiry, and it never appears in a report. The ad that produced it looks like it failed, and gets switched off.
Last quarter's links were tagged one way, this quarter's another, and the report shows eleven versions of the same channel. Nobody can compare anything with anything, so nobody does.
A customer sees an ad on social, searches on Google and buys. Both platforms report the order, and the total of what they claim is larger than what the shop took. Budget then follows the platform that claims loudest.
Set up badly, it either hides much of your traffic or ignores what people chose. Either way the figures shift on the day it was installed, and nobody noted the date.
Forty metrics, none of them profit. A good report fits on one page and ends with a decision: spend more here, stop there, fix this before spending anything.
The question lab
Nine things are commonly missing, and each one stops particular questions being answerable. Tick what you have today. The right-hand list shows which of six questions you can answer from data and which you are answering by instinct. It is a way of seeing the gaps, not an audit of your set-up.
Tap everything that is true today.
Your measurementTap all that apply
Nothing here is stored, sent to TWV Media or used to look at your tracking. It runs entirely in your browser.
Which items answer which question is a simplification. Real set-ups vary, and so does what they can tell you.
The tracking scorecard
It scores measurement as a decision tool: whether anybody agreed what to measure, whether the tags and consent are sound, whether the data joins up to real orders and enquiries, and whether anybody uses it. Your answers stay in this browser.
What we actually do
You can bring us in for the whole set-up or the one part currently making the numbers hard to trust. Either way, one senior person stays responsible for it rather than a rota.
The three or four decisions the numbers are for, what counts as a result, and which figure will be treated as the truth. Written before any tag is touched.
Everything goes through the tag manager, named consistently, tested on real devices and documented, so the next person can see what is there and why.
A banner that records what people chose and stops tags that were refused from firing. In the UK that is a design constraint from the start, not a banner added at the end.
One naming rule for every link in every channel, kept in a shared sheet, so the report shows five channels and not eleven versions of two.
Calls, forms, bookings and orders tied back to where they came from and what they were worth, reconciled monthly against your own records.
Spend, results and profit by channel, what changed and what we would do next. Read in five minutes by the person who decides.
What wrong numbers cost
Overclaiming is common, because platforms are paid on results and measure them generously. This sum shows what it costs you when the claim is out by a known amount. It does not tell you the true figure, only what is at stake in finding it.
The last figure is the part of a year's spend that bought no extra sales beyond what the channel really produced. A platform that overclaims is not necessarily a bad channel, but the decision about its budget needs the true number.
The cheapest improvement in most accounts is learning which numbers to stop believing.
How it runs
Every stage produces something you own and could hand to somebody else. Nothing is held hostage and there is no long contract behind it.
Which decisions the numbers are for, what counts as a result, and which figure wins an argument. One page, agreed before any tag is touched.
Every tag found, named and judged. Duplicates and dead tags removed, consent wired in and everything tested on real devices.
Sources, enquiries and orders tied to the same person or sale, and reconciled against your own records each month.
A one-page report, reviewed with the person who owns the budget, ending in what to spend more on and what to stop.
For a business whose numbers do not agree, or that has never been sure what its tags are doing.
We write the plan, set the tracking up and test it, and leave documentation your team can follow.
For businesses that want the measurement kept honest and the report written, without hiring an analyst.
Before you ask
Because each tool counts differently. Platforms attribute a sale to themselves over a window that suits them, analytics tools miss people who refuse cookies or block scripts, and your own records include things the others never see, like returns. The aim is not agreement. It is knowing the size and direction of each gap, and arguing from the record you own.
Sometimes. It can recover data that browser restrictions lose and gives you more control over what leaves the site, but it adds cost and something more to maintain. For a small account the first job is usually getting the ordinary set-up right. We will say plainly if the extra layer would pay for itself.
In the UK the main rules are UK GDPR and PECR, and the regulator, the ICO, publishes guidance that is updated from time to time, so check the current version. In practice, tags that need consent should not fire until it is given, and refusing should be as easy as accepting. That is a design approach and not legal advice, so your own adviser should confirm what applies to you.
Most businesses are fine with Google Analytics 4 alongside a tag manager, because it is free and the integrations are everywhere. Privacy-focused tools suit sites that want to avoid a consent banner for analytics, with less detail in return. We work in whichever you have and will tell you if it is the wrong one.
With a call tracking number that swaps in for visitors from a given source, or a click-to-call event on mobile, depending on how many of your calls come from the website. Neither is perfect, which is why we also ask callers how they heard about you and compare the two.
Monthly for most businesses, weekly while a campaign is being tested. More often than that and people react to noise. The report should end with a decision, and if a month passes with no decision, the report is the problem.
Yes, and it is the most common job. We start by listing every tag on the site, working out which ones do something and removing the ones that do not. It often turns out that the same conversion has been counted twice for years.
Yes. Joining orders and enquiries back to their source is what turns a report about traffic into a report about money. It needs some care over personal data, so we keep what is shared to the minimum the report needs.
No. The question lab, scorecard and cost sums run in your browser. Nothing is sent to us, nothing looks at your tracking, and everything disappears when the tab closes. If you copy your score, that copy goes to your own clipboard. Our privacy notice covers the rest of the site.
Tool and rules specifics on this page were last reviewed in October 2026. Analytics tools and consent guidance change often, so if something here no longer matches what you are seeing, tell us and we will correct it.
Also worth a look
Book a call
Send us your site address before the call and we will have looked at what tags it loads and what happens to them when somebody refuses cookies. You get the first thing we would change and why, whether or not you ever hire us.
Book a free 20-minute call