On one project, we were stuck with three recurring questions:

Why is Meta taking credit for half the orders?

Why does GA4 show everything as direct, even when we’re clearly running paid traffic?

And why does Google Ads show conversions that mysteriously never made it into the CRM?

Every time we dug into it, we hit the same wall: “We don’t have raw data. Just reports.”

Translation: We’re guessing. With graphs.

No one had a clue.

It was like having a pile of money, but no idea where it came from. Like getting a bonus and forgetting what job you did for it.

If you’re working only with GA4, Ads Manager, and your CRM, here’s what’s really happening:

You’re looking at a picture carefully curated by someone else.

You can’t connect a session to a sale.

You lose people every time they switch devices or sneeze near a cookie prompt.

You can’t build your own attribution logic.

And your insights? You’re working off what little data survived the platform’s internal filters — like trying to reconstruct a wedding from confetti.

So what happens?

You either undervalue good channels or overspend on fake performance. Congrats.

Here’s how I build analytics that doesn’t lie to me:

1. I export raw events from GA4 into BigQuery.

2. I capture every click ID — gclid, fbclid, utm, client_id — into the CRM. Those are your fingerprints.

3. I centralize everything into one warehouse. No spreadsheets of doom. No “we’ll merge that later.”

4. I store the user’s first touch. Because 9 times out of 10, that’s what goes missing. I track it with cookies, and carry it all the way to checkout like it’s precious cargo (because it is).

5. I combine auto-tagging with strict UTM discipline. Google Ads gets its auto-tagging, sure. Everything else follows a precise UTM scheme.

6. I build cohorts — not just by visit, but by key milestones. First visit, first registration, first purchase. That’s how you see the difference.

Raw data is like the black box in a plane.

When things go sideways, it’s the only way to know what actually happened.

It lets you trace the full journey, build your own attribution, prove which channels actually work — and most importantly, make decisions based on reality, not what a platform thinks you want to hear.

If you’re unsure where to start — start small. Track what matters. Connect the dots. Build your truth.

Because as long as you’re only looking at pretty dashboards, you’re not an analyst. You’re just a spectator — watching someone else’s version of your story.

And spoiler: they’re not rooting for you.

If you work with GA4 to BigQuery exports, be sure to check out my SQL cheat sheet.