Four data sources, one report, and a clear rule for each number.
An agency’s content work, search performance and AI referrals lived in different tools and spreadsheets. I built a reporting framework that brought them together without blurring what each one measures.
- Client
- A business compliance services provider, the same client as the Direct traffic project
- Service
- Reporting and analysis
- Systems
- GA4, Google Search Console, an AI visibility tool, Google Sheets
- When
- October 2026
What each source can show
Each source answers a different question, so each keeps its own definition in the report.
Four sources, four separate measures.
Source and what it measures
Visits and conversions from people first acquired through organic search, plus sessions that came from AI services.
Visits and conversionsImpressions, clicks and position in Google search.
Search performanceHow often the brand appears and is cited in AI answers.
AI presenceWhich pages the agency worked on, and when. Used to find visits that started on a worked page.
Worked pagesThe rule that matters most: AI referral visits are sessions on the site. AI visibility is how often the brand is mentioned. The report never adds them together.
The question
The team wanted to see whether the agency’s content and search work was paying off. The answer was spread across GA4, Search Console, an AI visibility tool and four spreadsheets of content work. Each source counts something different, so one chart without rules would mislead.
What I did
- Combined four records of content work (content audit, editorial calendar, on-page optimization and link building) into one list of pages and dates.
- Cleaned and deduplicated the page addresses so each page appears once, with a clear rule for which work date counts.
- Defined how the report connects organic search, pages the agency worked on, and conversions.
- Kept AI referral visits, which are real sessions, separate from AI visibility, which is how often the brand shows up in AI answers.
- Added a period-over-period view of top landing pages and a filter to include or exclude pages, then reviewed it with the team.
What it showed, and what stayed open
What it showed
- One reporting structure, with the definition of each measure written down.
- AI referral visits and AI visibility reported as two different things.
- A clear rule for when a page counts as worked on.
What stayed open
- Visits that start on a worked page don’t show that the content work caused a conversion, and the report doesn’t claim that.
- Sign-in, account and “(not set)” pages stayed in the view until the team decides whether to exclude them.
- The GA4 limits found in the Direct traffic project still apply to this data.
What should your report be able to answer?
Tell me who reads it, what they decide with it, and where the numbers come from.