First-Party Measurement Foundation
Google Analytics 4 — free, with Consent Mode built in
Cross-channel multi-touch reporting · Server-side tagging support
First-touch vs multi-touch vs incrementality testing. What survives iOS 17+ and GDPR, and the analytics stack that actually answers "what's working?"
Every major ad platform's attribution dashboard has a structural conflict of interest: it gets to grade its own homework. Meta's reported conversions and Google's reported conversions, summed together, routinely exceed total actual conversions — both platforms are claiming credit for the same customer, and neither has an incentive to fix that.
This isn't a reason to give up on measurement. It's a reason to triangulate across methods rather than trusting any single dashboard at face value — first-touch and multi-touch models for day-to-day signal, incrementality testing for the real causal answer when it matters most.
The post-cookie shift (iOS 17+ tracking limits, GDPR consent requirements, third-party cookie deprecation) hasn't made attribution impossible — it's made individual-level tracking less reliable, which is exactly why methods that don't depend on it have become more valuable, not less.
Three approaches with very different relationships to causation.
| Model | What It Measures | Best For | Limitation |
|---|---|---|---|
| First-Touch | The very first interaction in the journey | Understanding top-of-funnel discovery | Ignores everything after the first touch |
| Multi-Touch | Credit spread across multiple touchpoints | Long, multi-channel buying journeys | Still correlational, not causal |
| Incrementality Testing | Causal lift via holdout/geo experiments | Proving a channel's true causal impact | Slower, requires real experiment design |
The table tells you what each model claims to measure. This is how much you should actually trust each claim.
No single model is the whole answer — these steps build a layered approach.
Some signals held up. Others quietly stopped meaning what they used to.
| Method / Signal | Survives iOS17+/GDPR? | Notes |
|---|---|---|
| Third-party cookies | No | Largely deprecated across major browsers |
| First-party data (your CRM/email) | Yes | Increasingly the most reliable signal available |
| Server-side / Conversions API | Partially | Reduces but doesn't eliminate the data gap |
| Incrementality / geo experiments | Yes | Doesn't depend on individual-level tracking at all |
| Platform "view-through" attribution | Mostly No | Heavily inflated, treat with real skepticism |
For day-to-day channel optimization, multi-touch is a workable, imperfect model that's good enough for routine decisions. For big annual budget allocation calls, incrementality testing on your largest channels is the only method that actually answers the causal question.
Either way, first-party data is the foundation that survives regardless of which model you build on top of it.
View Our Full Attribution Tool RankingsYes, but individual-level tracking is meaningfully less reliable than it used to be. Methods that don't depend on individual tracking — incrementality testing and media mix models — have become more valuable precisely because of this shift.
Multi-touch models show which channels were present in a journey (correlation); incrementality testing shows whether a channel actually caused additional conversions that wouldn't have happened otherwise (causation). A channel can show up frequently in multi-touch reports while contributing little true incremental value.
At least annually on your largest budget channels, and any time you make a major strategic shift in spend allocation. More frequent testing on smaller channels is reasonable if you have the resources to run it well.
Smaller budgets can run simplified geo or holdout tests, though detecting a statistically meaningful effect gets harder as sample size shrinks. It's more accessible than commonly assumed, but the precision of the result scales with your spend and traffic volume.
An MMM estimates each channel's contribution using aggregate historical spend and outcome data rather than individual tracking. It's most valuable for larger, multi-channel budgets making high-level allocation decisions — likely overkill for a single-channel, small-budget operation.
As one input among several, yes — GA4's multi-touch reporting is a reasonable baseline. Treat it the same way as any platform-reported number: useful for relative trends, not the final causal word on what's actually driving revenue.
Yes for most teams — even an imperfect, correlational signal beats no signal at all for day-to-day decisions. The key is knowing its limitations and not treating it as the definitive answer for major budget reallocations.
This guide covers the methodology — our category page covers current feature comparisons across every attribution and analytics platform we've reviewed.