How a tracking audit works
An audit starts with an inventory, not with a tool. You list what the business considers a result — a form, a phone call, a WhatsApp message, a booking, a purchase — and then you list what is actually installed: containers, analytics properties, advertising pixels, plugins that inject their own tags, and any code a previous developer left behind. The two lists rarely match, and the gap between them is most of the finding.
Then each item is tested rather than assumed. You walk the site as a customer would, in preview or debug mode, and watch what fires: whether the event triggers at all, whether it triggers once, whether it carries the right value and currency, whether it reaches every platform that needs it, and whether it survives on mobile, where most Nepali traffic arrives. Finally the totals are reconciled against something real — the inbox, the call log, the CRM — because a tag that fires correctly can still be counting the wrong thing.
Why a tracking audit matters
Tracking degrades quietly. Nobody breaks it deliberately; a theme is updated, a form plugin replaced, a thank-you page renamed, a consent banner installed, and the events stop or double without any warning appearing anywhere. Because the charts keep drawing, the fault can run for months before somebody notices that the reported leads do not resemble the enquiries in the inbox.
It matters more once budget is automated. Smart bidding and Advantage campaigns optimise towards the conversion signal they receive, so a broken or duplicated signal directs real money at the wrong people quickly. Auditing before scaling a budget is far cheaper than auditing after a quarter of misdirected spend.
Common mistakes with tracking audits
The commonest is testing only the happy path. Auditors submit the main contact form on a laptop, see the event fire and stop. Real customers use phones, abandon and resubmit, refuse cookies, arrive from an in-app browser, and use the secondary forms nobody remembers — the newsletter box, the brochure download, the click-to-call button in the header.
The second is auditing the tag but not the definition. An event can fire perfectly and still be counted as a conversion when it should not be, or imported into an ad platform twice under two names. The third is treating it as a one-off project: an audit describes a moment, and the next website release can undo it. The fourth is producing a list of technical faults with no indication of which ones actually distort the numbers a decision depends on.
How to act on it
Run one when something changes and once in a while when nothing does — after a redesign, a platform migration, a plugin change, before increasing spend meaningfully, and periodically as maintenance. Fix in order of financial consequence: signals feeding automated bidding first, then reported conversions, then diagnostic events.
Write the result down as a measurement document that names each conversion, where it fires, what value it carries and which platforms receive it, so the next audit has something to compare against. Individual events are then re-checked with ordinary tag validation after each release, and if you would rather have someone else look first, my free tracking audit covers the same ground.
