Your attribution dashboard is probably wrong. Here's how to fix it in 30 days.

Most marketing teams make budget decisions from a dashboard they quietly don't trust. Paid social claims one number, Google Ads claims another, the CRM shows a third, and finance only believes the bank statement. Everyone is "data-driven", but nobody agrees on what the data says.

This isn't a tooling problem you can buy your way out of. It's an engineering problem, and it has a fairly predictable fix.

Why the numbers disagree

Browser tracking leaks. Ad blockers, cookie-consent rejections, Safari's tracking prevention and users switching devices all mean a meaningful share of conversions never reach your analytics tags. Each ad platform then fills the gap with its own modelling, which naturally favours itself.

Every platform grades its own homework. Meta, Google and TikTok each attribute the same sale to themselves using different lookback windows and rules. Add them up and you'll often "sell" more than you actually did.

The revenue lives somewhere else. Your analytics tool sees a sign-up. Your billing or back-office system sees the deposit, the refund, the churn three months later. If the two are never joined, you're optimising campaigns for sign-ups, not for profit.

The 30-day fix

Week 1: Audit what's actually firing. List every tag, pixel and event. Check them against real transactions for a sample period. You'll usually find duplicated events, missing parameters and conversions that fire before payment actually succeeds.

Week 2: Move key events server-side. Send purchase and deposit events from your backend, not the browser. Server-side tracking (for example via GA4's Measurement Protocol or the ad platforms' conversion APIs) is far more complete and isn't affected by ad blockers.

Week 3: Build one source of truth in your warehouse. Land ad spend, analytics events and backend revenue in one place, such as BigQuery. Join them on a stable user or transaction ID. This is the table everyone, including finance, agrees to use.

Week 4: Decide on an attribution model you can explain. Start simple: position-based or data-driven within your own warehouse, plus periodic holdout tests to check which channels genuinely add incremental revenue. A model your team understands beats a black box nobody trusts.

What changes

Once spend and real revenue sit in the same table, the conversation shifts from "which platform do we believe?" to "which campaigns produce customers who are still paying in 90 days?" That's the question that actually moves margins.

If your dashboards disagree with each other, that's normal. Leaving it that way is the expensive part.