Agency Pre-Audit Scorecard for GA4 Cleanup Calls

A cleanup call goes sideways when the agency and the client are still arguing about whether the data is "actually broken." A short agency pre-audit scorecard fixes that. Before the call, you grade the export for source/medium drift, paid-social naming QA, missing campaign discipline, and attribution mismatches across email or ecommerce handoffs. The goal is not a giant audit deck. It is a one-page proof object that shows where trust is leaking before anyone burns another hour in GA4.

Why agencies need a pre-audit scorecard before the cleanup call#

A cleanup sale usually dies in one of two ways. Either the call starts too broad, so the buyer hears a generic analytics pitch and delays the project. Or the call starts too deep, so you jump into tags, consent mode, and checkout flows before proving the visible business problem. A scorecard keeps the first conversation at the right altitude: clear enough to show the damage, narrow enough to justify the next step.

That matters because broken UTM hygiene rarely looks dramatic at first glance. The dashboard still loads. The acquisition report still has channels. But GA4 is case-sensitive, so email, Email, and EMAIL are different mediums, and 10–20% of sessions commonly land in Unassigned. When source, medium, and campaign values fragment like that, as much as 26% of conversions can be credited to the wrong channel. The scorecard is how you make that hidden mess visible in five minutes instead of forty-five.

The 5 sections every agency pre-audit scorecard should include#

1. 1

   1) Overall grade

   Open with one A–F health grade and a one-sentence reason. Example: "D because paid-social source values are fragmented, Unassigned is elevated, and lifecycle traffic is splitting across multiple medium labels." This is the line buyers remember and forward.

2. 2

   2) Source / medium drift summary

   Show the top duplicate or conflicting source/medium values ranked by impact. This is the fastest way to prove that the report is not stable enough for channel decisions. Use the same logic described in how to audit broken UTM data before a QBR, just condensed for pre-sales.

3. 3

   3) Paid-social naming QA

   Audit whether paid social is using one canonical source and one canonical medium policy. Buyers recognize this pain immediately because Meta naming drift tends to poison performance reporting long before anyone notices the taxonomy is drifting.

4. 4

   4) Attribution mismatch spot checks

   Pick one or two exact mismatch lanes such as Klaviyo to GA4, Shopify checkout to GA4, or HubSpot email tracking to GA4. You do not need a full root-cause report yet. You need enough evidence to show the mismatch is structural, not random.

5. 5

   5) Cleanup recommendation

   End with the next action: approve a deeper cleanup, run a done-with-you audit, or route the account through a one-time cleanup service. The scorecard should point to the engagement, not stop at diagnosis.

What to score inside source/medium drift#

This section is the backbone of the scorecard because it catches the highest-trust failures first. You are checking whether channel labels are stable enough to read over time, not whether every campaign is perfectly architected. If the same paid-social source appears as facebook, Facebook, fb, and meta, the account already failed the trust test.

If you want a tighter governance baseline after the sale, move the client into a locked naming policy using a UTM naming convention for teams. But before the sale, the simpler question is enough: are the current labels stable enough to trust? The scorecard answers that with evidence instead of theory.

How to handle paid-social UTM naming QA inside the scorecard#

Paid social deserves its own section because it is where agency buyers feel the pain fastest. A campaign can be performing fine while reporting looks unstable purely because naming drift split one source across several rows. The account team then wastes time debugging bids or creative when the real issue is taxonomy drift.

Add one mismatch lane that matches the client stack#

This is where the scorecard feels custom instead of generic. Pick the mismatch lane the client will recognize instantly. For ecommerce and lifecycle teams, that is often Klaviyo GA4 attribution mismatch or Shopify direct traffic spikes in GA4. For RevOps and CRM-heavy teams, it may be HubSpot email or campaign tracking that lands in unexpected source/medium rows. You are not trying to solve every tool in the pre-audit. You are showing one concrete proof that the reporting break crosses systems.

A simple scoring model agencies can reuse#

1. Give source consistency a 0 to 25 score based on how many canonical channel values are fragmented.
2. Give medium consistency a 0 to 25 score based on whether the same lane is split across medium synonyms or case drift.
3. Give campaign discipline a 0 to 25 score based on campaign formatting, token stability, and naming reuse.
4. Give cross-tool integrity a 0 to 25 score based on whether one stack mismatch is visible and explainable.
5. Convert the total into a grade and keep the scoring language stable from client to client, so the scorecard becomes a recognizable proof asset.

You do not need fake precision here. The goal is consistency, not actuarial math. If you grade ten prospects over a month, the score should help you compare who needs a light governance fix versus a real cleanup project. That is why an audit-first workflow beats generic discovery calls. The proof object gets sharper every time you reuse it.