Every e-commerce brand running paid media hits the same moment. Google Ads reports £50,000 in revenue. Meta reports £30,000. GA4 shows £45,000. Shopify says £35,000. The finance team is looking at £35,000 and wants to know why marketing is claiming £80,000.
The standard response is to assume something is broken: the tracking, the UTMs, the tag setup. Fix those, and the numbers will align. They won’t. These systems were built for different purposes, use different methodologies, and measure different moments in the customer journey. Structural misalignment is not a bug. It’s how these tools work.
The damage comes from choosing one of them as a single source of truth and making budget decisions from it alone.
Why The Numbers Differ: The Structural Causes
Take a straightforward customer journey: someone clicks a Meta ad, gets retargeted on YouTube, then searches for your brand on Google before converting, all within seven days. Using default attribution windows, both Meta and Google Ads will report one conversion each. GA4 and your CRM will record one conversion, most likely crediting Google paid search.
Example customer journey — 7 days
Meta ad click → YouTube retarget → Google brand search → Conversion
- Google Ads: 1 conversion credited.
- Meta Ads: 1 conversion credited.
- GA4 / CRM: 1 conversion, attributed to Google paid search.
- Total reported across platforms: 3 conversions.
- Actual conversions: 1.
Meta hasn’t invented a duplicate conversion. It has no visibility into what happened on Google. It correctly recorded a click that preceded a conversion within its attribution window. The disagreement between platforms isn’t dishonesty. It’s each system reporting accurately from its own limited view.
Beyond multi-touch duplication, three further structural causes create the gaps:
Attribution date differences. Ad platforms attribute conversions to the day the click occurred. GA4 and CRMs report on the day the conversion happened. For accounts with long customer journeys: considered purchases, B2B leads, and high-ticket e-commerce. This creates meaningful date-range discrepancies that look like data loss but are actually timing differences.
Cross-device behaviour. A user who clicks a Google Ads ad on mobile, then returns on desktop via organic search and converts, will appear as one conversion in your CRM. But the CRM won't merge the mobile and desktop sessions, so the source it records may not match the source Google Ads is crediting.
Privacy restrictions. Ad blockers, browser-level tracking prevention, and consent banner rejections mean a measurable share of conversions isn’t tracked at all. Some platforms fill the gap with modelled conversions. Your CRM records nothing. The gap this creates is structural and grows over time as privacy adoption increases.
Server-side tagging, offline conversion imports, and consistent UTM parameters address some of this, particularly the cross-device and privacy gaps. They don’t resolve multi-touch attribution duplication.
That’s a methodological problem, not a technical one.
The Attribution Trap
Once teams accept that the numbers differ, the standard response is to nominate a single source of truth. Usually, GA4 or the CRM. Then make every decision from that one number.
That’s the attribution trap. Every tool follows an attribution model, and every attribution model has structural blind spots that favour certain channels and systematically undervalue others.
Last-click attribution
- Rewards: The final touchpoint before conversion (typically branded search).
- Systematically undervalues: All demand-generation activities that created the intent to convert.
First-click attribution
- Rewards: The discovery channel that first introduced the customer.
- Systematically undervalues: Nurture and conversion touchpoints that helped close the sale.
Linear attribution
- Rewards: All touchpoints equally across the customer journey.
- Systematically undervalues: The actual relative influence of each touchpoint, since credit is distributed arbitrarily.
Time-decay attribution
- Rewards: Touchpoints closest to the conversion event.
- Systematically undervalues: Upper-funnel activities that built awareness and intent earlier in the journey.
Data-driven attribution
- Rewards: Touchpoints weighted according to statistical contribution to conversion.
- Systematically undervalues: Transparency and interpretability, as the weighting methodology is often a black box with limited visibility.
The practical consequence: whichever model your single source of truth follows, budget decisions made from it alone will consistently over-invest in the channels it rewards and defund the channels it can't see.
The most common version of this problem: a brand relies on CRM data, which follows last-click, and consistently attributes performance to branded search. Demand generation channels (Demand Gen, Meta prospecting and YouTube) look inefficient by comparison. Budget shifts toward brand. Three years later, branded search volume has declined because the activity that created brand awareness was systematically defunded.
What Each System Can and Cannot See
Google Ads
Can see
Clicks, impressions, conversions attributed within its window. Cross-campaign data within Google's ecosystem.
Cannot see
Meta interactions. Offline behaviour. The full multi-channel journey. What would have happened without the ad.
GA4
Can see
On-site behaviour. Session sources where tracking fires. Cross-channel last-click by default.
Cannot see
Awareness activity that didn't generate a site visit. Consented users only. Not a neutral arbiter.
CRM / Shopify
Can see
Actual orders and revenue. New vs returning customer distinction. The business reality.
Cannot see
Which marketing activity caused the order. Multi-touch journeys. Upper-funnel contribution.
The CRM is the only system that records what actually happened commercially. Shopify showing £35,000 is the truth. Google Ads and Meta are showing a combined £80,000, which is each platform’s interpretation of its contribution to that £35,000. Those are different questions.
Incrementality: The Right Question, Often the Wrong Tool
Attribution answers: given that a conversion happened, which touchpoints should get credit? Incrementality answers a different question: did this campaign cause conversions that wouldn't have happened otherwise?
The distinction matters for budget decisions. A retargeting campaign with a strong attributed ROAS may be capturing conversions that would have happened anyway. Users already in-market, already familiar with the brand, already close to buying. Showing them an ad and claiming credit is what attribution does. Incrementality tests whether removing the ad changes the outcome.
Three approaches to incrementality testing:
Geo holdout
Divide your market into comparable geographic regions. Run campaigns in some, go dark in others. Measure the difference in conversions. Practical and relatively straightforward to set up for most e-commerce brands.
Audience holdout
Exclude a defined percentage of your target audience from seeing ads, then measure the conversion difference between exposed and unexposed groups. Available in Google and Meta. Important caveat: only valid for comparing campaigns within the same platform; cross-platform holdout comparisons are not methodologically sound.
Time-based test
Pause a campaign for a defined period and measure what happens to overall conversion volume. High risk: seasonality, competitor activity, and external events can distort results. If the campaign was incremental, performance is damaged during the test. Best used for lower-spend, stable-demand scenarios.
Incrementality testing at the scale required to produce statistically reliable results typically requires significant monthly spend—in the region of £80,000–£100,000 or more per month to run controlled experiments with meaningful differences between test and control groups. For most e-commerce brands, full incrementality testing isn’t operationally feasible. Shortcuts exist for specific decisions, particularly branded search, where auction insights can indicate whether competitor bidding makes brand campaigns necessary.
Triangulation: The Practical Alternative
For brands that can’t run formal incrementality tests, triangulation is the operational answer. Use all three systems simultaneously, understand what each one can and cannot see, and make decisions from the pattern rather than any single number.
In practice:
- Start with the CRM. Shopify or your CRM records actual revenue. Every other number is an attempt to explain it, not to replace it. When Google Ads and Meta report combined revenue that exceeds what the CRM shows, the CRM is right. The gap is platform attribution overlap, not additional revenue.
- Identify where duplication is occurring. If you're running Demand Gen and Meta retargeting simultaneously to the same audience, both will claim credit for conversions in the overlap. Map the customer journey across both platforms before concluding that either is underperforming.
- Track the ratios over time. The gap between ad platform-reported conversions and CRM conversions should remain relatively stable. Build a simple report that tracks the ratio. Google Ads reported revenue as a multiple of Shopify revenue, for example. If the ratio holds, the measurement framework is stable. If it changes materially, investigate. It may indicate a tracking change, a campaign mix shift, or a genuine incrementality signal.
- Use GA4 for on-site behaviour, not for channel attribution. GA4 is strongest on what happens after the click: landing page performance, funnel drop-off, on-site conversion rate. Using it as the arbiter of which channels deserve budget is where its limitations create the most damage.
- Segment campaigns by journey stage. Awareness campaigns, such as YouTube and Demand Gen, at the top of the funnel should not be evaluated on last-click attributed revenue. Set micro-conversions appropriate to the stage: engaged sessions, product page views, and email captures. Evaluate them against those metrics, not against conversion volume; they were never designed to own.
The Decision-Making Shift
The teams that consistently misallocate budget are the ones trying to force three systems to produce the same number or searching for the attribution model that finally feels fair.
The correct position is to accept that these systems measure different things from different angles and to build a decision-making framework that uses all three rather than elevating one.
Your CRM tells you what happened. Your ad platforms tell you their interpretation of why. Your GA4 tells you what happened on-site between the click and the conversion. None of those is the full picture. Together, they’re workable.
360 OM View
The measurement problem in paid media isn’t a tracking problem. It’s a framework problem. Fix your tracking through server-side tags, offline conversions, and clean UTMs, but don’t expect alignment. Expect three different views of the same commercial reality and build the discipline to make decisions from all three. The CRM is the ground truth on revenue. Everything else is context.
Need a fresh perspective? Let’s talk.
At 360 OM, we specialise in helping businesses take their marketing efforts to the next level. Our team stays on top of industry trends, uses data-informed decisions to maximise your ROI, and provides full transparency through comprehensive reports.
.png)
