Google Ads conversion tracking is the one number most business owners take on trust. The account says forty leads, so there were forty leads. In 2026 that assumption is quietly costing people money, because the figure you see in the interface is very often lower than the number of enquiries that actually arrived, and sometimes a great deal lower.
The reasons are technical but the effect is commercial. Browsers have spent years tightening what they allow, consent banners now block a slice of every audience, and people research on a phone before buying on a laptop. Every one of those gaps removes a conversion from the record while the sale still lands safely in your inbox.
That matters more than it used to, because Google's bidding is automated. Smart Bidding learns from the conversions it can see, so a broken setup does not only mislead you, it actively teaches the algorithm to chase the wrong clicks. In this guide we explain where the losses happen and the fixes we apply for our own clients.
The first leak is third-party cookies. Safari and Firefox block them outright, and between them they carry a meaningful share of UK browsing, particularly on iPhones. When the cookie that connects a click to a purchase never survives the journey, the conversion is entirely real but completely invisible, and your reported cost per lead looks far worse than it truly is.
The second leak is consent. Every visitor who declines tracking on your cookie banner takes their data with them, and on a typical UK site that is a large minority of your traffic. The third is the gap between devices and channels. Someone taps an ad on the train, thinks it over, then rings you two days later from the office.
Put those together and it is common to find accounts recording only two thirds of what genuinely happened. We audit this at the start of every Google Ads management engagement, and the pattern repeats. Nobody had done anything wrong. The setup was simply installed years ago and never revisited as the web changed around it.
Enhanced conversions are the highest return fix available to most small businesses, and they take about an afternoon to set up. When someone submits a form or completes a checkout, your site sends Google a hashed version of the email address or phone number they typed. Hashing scrambles it irreversibly before it leaves the browser, so no readable personal data ever travels.
Google then matches that hashed identifier against signed-in users and recovers conversions the cookie alone would have lost. In real accounts we typically see reported conversions rise by somewhere between five and thirty per cent within a fortnight. Nothing changed in the business itself. The measurement simply caught up with what was already happening.
There are two routes in. Enhanced conversions for web run through Google Tag Manager or the Google tag, and suit most service and e-commerce sites. Enhanced conversions for leads work better when your sales happen over the phone or inside a CRM. Either way, make sure your privacy policy describes the practice properly before you switch anything on.
Consent mode is the piece most UK businesses have only half implemented. It is a signal layer that tells Google whether a given visitor agreed to analytics and advertising storage. When it is configured properly, visitors who decline still trigger anonymous pings, and Google uses those to model the conversions those people most likely completed.
Advanced consent mode is the version worth having. Basic mode withholds the tags entirely until someone clicks accept, which leaves Google nothing at all to model from. Advanced mode fires cookieless pings from the first page view, so the modelling has enough raw material to work with. Most certified consent platforms offer both, and the default is not always the better choice.
Two cautions are worth knowing. Modelling needs volume, so a site with a handful of conversions a month will not get reliable estimates and should treat modelled figures as directional only. Enhanced conversions also depend on the same consent signals, so if the banner is misconfigured your hashed identifiers are discarded at Google's end while everything appears to be working.
If your sales close on the phone, in a showroom or several weeks after the first enquiry, the click that started it all needs to travel with the lead. Offline conversion imports do exactly that. You store the click identifier alongside the enquiry in your CRM, then upload the outcome once the deal is won, so Google learns which clicks became revenue rather than which became form fills.
One change is worth noting this year. Google now applies a tighter window on offline uploads, so the habit of sending a spreadsheet once a month will see rows rejected. Move to a weekly upload, or better still automate it straight from your CRM. The gap between a lead and a paying customer is exactly where most advertising budgets are wasted.
Server-side tagging is the next step up. Instead of the browser reporting to Google, your own server does, which sidesteps ad blockers and browser restrictions. It costs a little to run and is overkill for a small local service business, but for an online store with real order volume the recovered data pays for it quickly.
You can sanity check all of this yourself. Start by counting reality for a fortnight. Every enquiry email, every phone call, every order confirmation. Then compare that total with what Google Ads reports for the same dates. A gap of ten per cent is fairly normal. A gap of forty per cent means something is broken and deserves an afternoon of attention.
Next, open your conversion actions and look for duplicates. Accounts that have run for years often count a thank you page and a form submission as two separate conversions, or still include a legacy action nobody remembers creating. Check the attribution setting too, and make sure only the actions representing genuine business value are marked as primary.
Finally, test the whole path with Google Tag Assistant. Submit your own form, complete a test order, and watch whether the tag fires and whether the consent signals look correct. If the numbers still refuse to add up after all that, talk to our team and we will run a full tracking audit on the account.
Getting Google Ads conversion tracking right is not glamorous work, but it changes everything downstream. Accurate data means Smart Bidding optimises towards the customers you actually want, your reported cost per acquisition becomes a figure you can plan around, and the decision to raise or cut budget stops being guesswork dressed up as strategy.
Our order of work is the same every time. Fix the conversion actions first, then add enhanced conversions, then sort out consent mode, then add offline imports if the sales cycle calls for them. Each step is independent, each one recovers data you were already earning, and none of them requires rebuilding your website.
If you are spending money on Google Ads and quietly suspect the reporting is not telling the whole story, you are probably right. At Primedia we audit tracking as part of every account we take on, and we are always happy to look at yours. Get in touch and we will tell you exactly where the data is leaking.
Because browsers block third-party cookies, visitors decline consent banners, and people switch between devices before buying. Each of those breaks the link between the ad click and the sale, so the conversion happens in real life but never reaches your Google Ads account. Gaps of thirty to fifty per cent are common in accounts that have not been updated recently.
Enhanced conversions send Google a hashed version of the email address or phone number a customer entered on your site. The data is scrambled irreversibly before it leaves the browser, and Google matches it against signed-in users to recover conversions the cookie missed. Most businesses see reported conversions rise by five to thirty per cent after switching them on.
If you advertise to people in the UK or the European Economic Area, you need a compliant consent banner and Google expects consent signals to be passed through. Advanced consent mode is the better option because it sends anonymous pings for visitors who decline, which lets Google model the conversions those visitors probably completed.
They are statistical estimates of conversions from visitors who declined tracking, not individually verified sales. They are genuinely useful for spotting trends and for feeding automated bidding, but treat them as directional rather than exact. Modelling also needs a decent volume of consented traffic, so very small accounts should rely on them cautiously.
Do a full review every six months, and always after a website redesign, a change of consent platform, or a new form or checkout system. A quick monthly sense check is also worth doing. Compare the enquiries you genuinely received with what the account reports, and investigate any gap that keeps widening.
Google Ads conversion tracking under-reports in 2026. Here is why your numbers are wrong, and the consent mode and enhanced conversions fixes that work.
Website accessibility for small businesses explained. What UK law requires in 2026, what WCAG Level AA means in practice, and the fixes that matter most.
AI now summarises your ratings and review gating is banned. Read our 2026 guide to Google reviews for small businesses, including how to ask properly.
