Use when campaign reporting is full of near-duplicate sources and you need to know which rows to trust.
utm-campaign-tagging-audit.md
You are auditing campaign tagging. You are not judging campaign performance, only working out which rows in the report misstate where traffic came from.
Tagged URLs or a source and medium export: {{TAGGING_EXPORT}}
The convention we are meant to follow: {{TAGGING_CONVENTION}}
Channels and platforms actually in use: {{CHANNELS_IN_USE}}
Channel grouping rules in the reporting tool: {{CHANNEL_GROUPING_RULES}}
Output:
1. A table: Value seen | Parameter | Which part of {{TAGGING_CONVENTION}} it breaks | Where it lands under {{CHANNEL_GROUPING_RULES}} | Corrected value.
2. Groups of values that are one campaign spelled several ways, with the single spelling to keep. Include differences that are only capitalisation or spacing.
3. Misfiled traffic: paid arriving as organic, internal links carrying campaign tags, email with no tag at all, and any platform in {{CHANNELS_IN_USE}} that does not appear in {{TAGGING_EXPORT}}.
4. The reporting consequence of each group, stated as which channel is over counted and which is under counted.
5. Fixes split into three lists: change at source going forward, rewrite inside the reporting tool, and cannot be corrected retrospectively.
Rules:
- Do not propose rewriting historic data without saying which saved reports and audiences break.
- If {{CHANNEL_GROUPING_RULES}} does not cover a value you found, say the rules need extending rather than inventing one.
- Sort section 1 by sessions affected where the export carries them, and say plainly if it does not.
Replace each placeholder with your own detail. The more specific you are, the less the model invents.
When do I audit tagging rather than tidy the campaign report?
When the source and medium report is full of near duplicates and you cannot tell which rows to trust. Tidying the report treats the symptom. This works out which rows misstate where traffic came from, which is a prerequisite for any channel comparison you were planning to build on top of it.
What do I need in front of me?
A tagging export or a source and medium list, the convention you are meant to follow, the platforms actually in use, and the channel grouping rules from your reporting tool. Export a full year. The worst tagging usually comes from a campaign run by somebody who has since left the team.
What comes back?
A table of broken values with the rule each breaks and where it lands under your grouping rules, duplicate spellings grouped with the one to keep, misfiled traffic such as paid arriving as organic, the over and under counting consequence per group, and fixes split three ways. The cannot be corrected list is the honest part.
What is the mistake that costs me here?
Fixing the tags before the grouping rules. Grouping rules repair history as well as future traffic, while new tags only start from today, leaving two eras in one report. Check which saved reports and audiences break before rewriting anything historic, which is exactly what the prompt asks you to state.