Planning PPC budgets was never straightforward for me, especially when facing differing data from Google Ads, Meta Ads, GA4, and my CRM/CMS. I often ask myself, what numbers should I actually report, and how can I ensure I’m optimizing for a genuine impact?
Like many, I believed better tracking, cleaner UTMs, or a refined analytics setup might solve the problem. But often, it’s something else entirely—the attribution trap.
We’ve been taught to rely on data-driven marketing. Ideally, analytics tools clarify what’s effective if configured right. But is it enough to just follow the data?
Attribution can be misleading. Without a solid framework, I find myself making budget decisions based on incomplete insights, potentially damaging the business.
Let’s consider: Attribution assigns conversion credit to channels, which is useful, but it doesn’t reveal which channels actually drove those conversions.
This may sound academic, but understanding it is crucial for solving the measurement puzzle. I’ll explore why attribution fails, how to use existing data effectively, and if incrementality testing is necessary.
Why ads, analytics, and CRM numbers never match
Aligning ad networks, GA4, and CRM data seems impossible. These systems serve different purposes, follow different methodologies, and measure distinct moments in the customer journey.
Your customer journey as a framework
Picture someone clicks on a Meta ad, sees retargeting on YouTube, then Googles the brand before buying—all in a week.
With standard attribution windows, both Meta and Google Ads report one conversion. GA4 and my CRM also show one, likely crediting Google Ads paid search.
Did Meta create a “duplicate” conversion? No. Meta can’t see Google Ads interactions, so it can’t detect duplicates.
GA4 and CRM probably ignore Meta Ads. Should I move Meta Ads budget to Google Ads branded search based on that? Probably not.
Structural differences as diagnosis enhancers
It doesn’t end there:
- Attribution date: Ad platforms credit conversions on the click day, whereas GA4 and CRMs report based on conversion day, leading to discrepancies with long customer journeys.
- Cross-device behavior: Different devices for interactions lead to CRM discrepancies if users aren’t merged correctly.
- Privacy restrictions: Ad blockers and cookie consents prevent some conversion tracking, and ad networks use modeled conversions to fill these gaps, unlike CRMs.
Some issues are fixable with better configuration, such as server-side tagging, offline conversion imports, and consistent UTMs. However, structural differences mean expecting full correlation is unrealistic.
Your single source of truth: The attribution trap
Once I accepted the number disparities, my next temptation was choosing a single source of truth, often GA4 or CRM, and relying on it. That’s where the attribution trap snaps shut.
Every tool uses an attribution model. Regardless of model—be it first-click, last-click, linear, time decay, or data-driven—they all have limitations.
Every attribution model has blind spots
- Last-click. Although easy to understand, it’s easy to exploit by rewarding the final touchpoint and undervaluing demand generation.
- First-click. It rewards discovery but ignores what convinces a customer to convert.
- Linear and time-decay. While they seem balanced, they’re quite arbitrary, as customer journeys don’t follow strict rules.
- Data-driven. Despite its sophistication, its mechanisms remain opaque, perpetuating a “black box” issue.
What happens depending on your source of truth
Hopefully, you now grasp the deeper issue: attribution addresses which touchpoints deserve credit once a conversion occurs. Relying solely on one tool means you can’t escape the attribution model’s blind spots.
If I depend solely on my CRM, I fall into the last-click attribution pit, often focusing on branded search. Over time, I might see demand decline despite strong results from my chosen source of truth.
Conversely, depending only on ad platform data means inflated results reporting, showing 2x to 4x more revenue than finance actually sees, resulting in increased marketing budgets while finance calls for cuts.
GA4 seems mature, but it only captures on-site customer journeys, missing awareness campaigns that might not result in website visits.
Realizing each tool’s fundamental flaws will lead someone to suggest incrementality testing — Did this campaign drive otherwise impossible conversions?
Incrementality tests: The perfect solution?
Incrementality measures results from your campaign — conversions that wouldn’t have existed without it.
Think of two worlds: one where the ad ran, the other where it didn’t. The difference between these worlds is your incremental impact. Everything else is baseline activity.
Attribution vs. incrementality
This distinction is crucial. Many reported conversions, especially from retargeting and branded search, are from individuals who would have converted anyway.
An ad followed by a conversion doesn’t guarantee the ad caused it. Incrementality testing measures the real credit.
For budgeting, distinguishing between true conversion drivers and illusions is vital.
A retargeting campaign showing strong ROAS might deliver little incremental value. If I cut it, conversions barely change; keeping it means paying for illusory performance.
How to test for incrementality
Testing incrementality involves experiments with two groups: one exposed to the ad and one that isn’t. Here are some typical methods:
- Geo holdout. Compare regions where campaigns run versus those where they don’t and observe conversion differences.
- Audience holdout. Platforms like Google and Meta allow excluding portions of the target audience from ad exposure, then measuring outcome differences.
- Time-based testing. Temporarily halt campaigns to assess changes in conversion volumes, though this method carries risks like seasonal effects blurring results.
Is incrementality right for you?
For those managing large budgets — say €1 million per month — you’re likely familiar with these tests. But what if you’re running a smaller operation?
At this scale, incrementality can be impractical as reliable tests demand meaningful test and control group distinctions, necessitating significant data and spend.
Nonetheless, I can use shortcuts, particularly around branded search, to spot potential problem areas.
Triangulation: The actionable decision-making process
Considering attribution limitations and incrementality tests for big advertisers only, I rely on triangulation.
Utilize existing tools, acknowledging their imperfections, and educate clients or leaders on not sticking to a “single source of truth.”
Start with your CRM/CMS
These systems track genuine deals and revenue. Treat all other figures as explanatory attempts.
If the ad platforms together show $50K revenue while Shopify reports $35K, trust Shopify as it reflects reality.
It can even differentiate conversions from new versus returning customers, crucial for measuring nCAC.
Overlay my customer journey onto ad platform results to understand campaign impacts along the journey, using available incrementality tests to decide budget allocation better.
Improve on triangulation
Attribution windows: Long customer journeys challenge interpretation. Segment campaigns by customer journey stages, and shrink attribution windows to improve outcomes.
Track ratios: Keep the gap between ad platform conversions and CRM data consistent. Sudden changes might reveal an incrementality insight.
Triangulation won’t provide clean numbers. But it will deliver a consistent decision-making framework, far superior to false precision.
Welcome to the real world
The teams that struggle the most force three systems into one report or search for the ultimate, fair attribution model.
Teams making informed decisions embrace complexity over a single truth, fostering data skills to match reality’s complexities.
Ensuring our decision-making stays realistic and accommodating of uncertainties makes all the difference.
Inspired by this post on Search Engine Land.
