Question 1

Why do GA4 and Google Ads report different numbers for the same purchase?

Accepted Answer

Walk one purchase through both rulebooks. A buyer clicks your Google ad on Monday, returns through an organic search on Thursday, and buys. Google Ads counts 1 conversion and books it to Monday - the date of the click - because Ads claims any conversion inside its click-through window and reports against click date. GA4 records the same purchase on Thursday and assigns credit under its own attribution model, where the organic visit can take some or all of it - so the paid-search row may show a fraction, or nothing. One purchase, counted once by each tool, filed under different dates and different channels. Both are correct under their own documented counting, and a month of purchases multiplies the gap into totals that cannot match. This is why forcing the tools to agree is a fool's errand, and why the Reconciliation Rule exists instead: platform numbers steer bidding inside that platform, GA4 reads cross-channel behaviour, and the CRM is the truth the business reports. We reconcile monthly and annotate the known variances, so nobody wastes a quarter chasing a discrepancy that is working as documented.

Question 2

We already have GA4 installed. What would you change?

Accepted Answer

Usually the event layer. A default install gives you automatically collected and enhanced-measurement events - pageviews, scrolls, outbound clicks - which describe traffic but not your funnel. The gap is custom events with clean parameters, key events mapped to revenue stages, and naming conventions that stop the property silting up. The audit also routinely finds missing cross-domain linking, key events firing on thank-you pages that reload, and UTM chaos burying half of paid traffic in unassigned. Each one quietly corrupts a different report.

Question 3

So which platform's numbers should the business trust?

Accepted Answer

Each one, for its own job - and none of them as the verdict. Platform figures steer bidding inside that platform: Google Ads and Meta each claim any conversion their ads touched inside their own attribution window, Meta including view-through conversions from people who saw but never clicked. GA4 reads behaviour across channels under its own model. Three tools, three rulebooks, three legitimate answers to different questions. The verdict belongs to the CRM - that is the Reconciliation Rule. Pipeline and revenue get reported from the deal record, the platforms get reconciled against it monthly, and known variances are annotated instead of argued about. Once each number has a job, the discrepancies stop being a crisis and start being documentation.

Question 4

What is server-side tagging, and do we need it?

Accepted Answer

A tagging server - typically on cloud infrastructure you control - receives one event stream from the browser and forwards it to GA4, Google Ads and Meta. You get a first-party endpoint, fewer third-party scripts in the page, and control over exactly which fields each platform receives. The honest answer on need: it earns its keep when ad spend is meaningful, when browser-side signal loss is visibly hurting match rates, or when data control matters. Below that, a clean browser-side build with enhanced conversions covers most of the gap. We will tell you which side of the line you are on.

Question 5

What are key events, and how do they relate to conversions?

Accepted Answer

In GA4, everything is an event, and the events that matter get marked as key events. Share a key event to Google Ads and it becomes a conversion the bidding can optimise toward - same underlying event, two names, two jobs, both documented. The discipline is choosing them well. Mark too many and bidding optimises to noise; count a thank-you page that reloads and one lead counts three times. We define key events against CRM stages and audit the counting method on every one.

Question 6

Do we need BigQuery?

Accepted Answer

GA4 includes a native BigQuery export at no licence cost - reason enough to switch it on from day one, because it banks raw event-level data you own permanently and cannot backfill later. Whether you query it yet is a different question. It matters when you need analysis the GA4 interface cannot do - custom attribution, joining web events to CRM records, long-window cohorts. Until then it costs little to let it accumulate, and the history is there when you need it.

Question 7

Can you show us why a page underperforms, not just that it does?

Accepted Answer

Yes - that is the qualitative layer. Microsoft Clarity heatmaps, scroll maps and session recordings show where attention lands, where forms get abandoned and where rage clicks cluster, at no licence cost. Paired with GA4 it turns "this page leaks" into "this field is the leak". We use it surgically: GA4 flags the pages worth watching, Clarity explains them, and the finding feeds the CRO or development backlog with evidence attached.

Question 8

Will all this tracking slow our site down?

Accepted Answer

Done properly, no - and server-side tagging usually makes pages lighter, because the browser sends one stream to your endpoint instead of loading a script per platform. Tags load asynchronously, fire after content, and get audited so dead ones are removed rather than accumulating. Tag bloat is real on neglected sites - years of pixels nobody owns. Part of every build is an inventory: keep, consolidate, or delete.

Question 9

How do you keep UTM tagging consistent across teams and agencies?

Accepted Answer

With the Naming Constitution - a register, not a memo. One documented taxonomy - allowed sources, mediums and naming patterns, all lowercase - plus a shared link builder, so the convention is easier to follow than to break. Then we audit. A quarterly sweep catches rogue tags and the channels they pollute, because one camelCase campaign name from a freelancer can split a channel into three reporting rows for a quarter.