Mountain View
Analytics

The signup count was right. Where the signups came from was the gap.

A SaaS platform wanted to know whether its signup reporting could be trusted. I checked the count against internal reporting and separated what was accurate from what was still missing.

Tell me what you're working onSee what it showed
Client
A SaaS platform
Service
Analytics audit
Systems
GA4 and the client’s internal reporting
When
August to October 2026

Two questions, two different answers

The count and the source of each signup are separate questions. They had to be checked separately.

How many signed up?

GA4’s total lined up with internal reporting. The signup events weren’t overcounting.

Can be trusted
Where did they come from?

Landing page and channel were missing for a large share of signups. Channel reports needed a caveat.

Still incomplete

Why it matters: an accurate count can sit next to incomplete source data. Reporting them as one number hides the gap.

The question

Signups start on a marketing website and finish in an app, on the web or on a phone. When channel reports looked off, the team needed to know whether signups were being miscounted, or whether the count was fine and the source data was the problem.

What I did

  1. Investigated the signup discrepancies the team had flagged.
  2. Checked GA4’s overall signup totals against the client’s internal reporting.
  3. Kept clicks on signup buttons separate from completed signups, and compared rates against the right base.
  4. Worked with the client’s team on telling web and mobile signups apart in internal reporting.
  5. Followed a change the client’s engineer made to improve attribution, and reported plainly that it didn’t help much.

What it showed, and what stayed open

What it showed

  • GA4’s overall signup count lined up with internal reporting, so the count could be trusted.
  • The real issue was attribution. The signup events weren’t firing falsely.
  • The team knew which part of the reporting to rely on and which part to treat with care.

What stayed open

  • Landing page and channel information stayed incomplete for a large share of signups.
  • The attribution change tested in late August didn’t make a material improvement.
  • Channel analysis needed more work before decisions could rest on it.
A dirt trail climbing through wildflowers toward a line of pines.

Which part of your reporting do you trust?

Tell me what you count, where it comes from, and what doesn’t add up.

Tell me what you're working on