You Are Probably Measuring Your Performance Marketing Wrong

Why Platform Reported ROAS Is Misleading Your Budget Decisions

The attribution problem that costs businesses millions.

Every advertising platform attributes conversions to itself. Google counts conversions from people who clicked Google ads. Meta counts conversions from people who saw or clicked Meta ads. Both count the same conversion when the same person interacted with both.

Add up platform reported conversions and you get a number substantially higher than actual conversions. This is not fraud. It is the structural consequence of each platform reporting its own attribution window independently.

The scale of the problem grows with media mix complexity. A business running Google Search, Google Display, Meta, LinkedIn, and programmatic might see platform reported conversions 40 to 60 percent higher than actual conversions. Decisions based on these numbers systematically overinvest in marketing.

How platforms take credit for sales that would have happened anyway

Brand search is the classic example. Someone searches for your brand name, clicks your paid ad, and converts. Google reports this as a paid search conversion. But that customer was looking for you specifically. They would have found you organically. The ad captured rather than created the conversion.

Retargeting has the same problem. You show ads to people who visited your site, some of whom were already planning to purchase. They convert, and the retargeting platform claims credit. But some percentage would have converted without the reminder. Retargeting takes credit for intent that already existed.

The question that matters is not 'did this channel touch the conversion' but 'would this conversion have happened without this channel'. Platforms cannot answer this question because answering it honestly would reduce their reported performance.

Building measurement frameworks that reveal true performance

True performance measurement requires looking beyond platform reports. This means connecting advertising data to actual business outcomes through your own systems. Revenue in your accounting system. Customers in your CRM. Orders in your ecommerce plat

The gap between platform reported conversions and verified conversions reveals the scale of attribution inflation. This gap varies by channel, campaign type, and attribution window. Understanding it enables more accurate budget allocation than trusting any single platform's numbers.

Incrementality Testing: The Only Way to Know What Actually Works

Incrementality testing answers the question that attribution cannot: what happens when we turn something off? By deliberately withholding advertising from a test group and comparing their behaviour to a control group, you can measure true incremental impact.

The results often surprise. Channels that appear highly efficient in platform reporting sometimes show minimal incrementality. The conversions they claim were largely going to happen anyway. Other channels that appear modest in platform metrics show strong incrementality. They are creating demand rather than just capturing it.

Incrementality testing requires statistical rigour and sufficient scale. Small tests produce noisy results. But for businesses with adequate conversion volume, incrementality measurement transforms budget allocation from guesswork to evidence.

Cross Channel Attribution Without the Complexity

Attribution models distribute credit for conversions across touchpoints. Last click gives all credit to the final interaction. First click gives all credit to the initial discovery. Linear distributes credit equally. Time decay weights recent interactions more heavily. Position based credits first and last touches most.

No model is correct. Each is a simplification of complex customer journeys. The question is which simplification is least misleading for your business.

For businesses with short purchase cycles and direct response objectives, last click often provides useful signal despite its limitations. For businesses with long consideration phases and multiple touchpoints, position based or data driven models better reflect reality.

Practical approaches for businesses without data science teams

Not every business needs complex attribution modelling. Practical approaches include comparing platform reported conversions to actual conversions, testing channel removal to measure impact, tracking customer source through surveys and CRM data, and using Google Analytics multi channel funnels to understand path patterns.

The goal is not perfect attribution. Perfect attribution is impossible. The goal is attribution good enough to make better decisions than competitors using worse attribution. Even simple improvements over platform defaults create advantage.

When to trust last click and when to ignore it entirely

Last click attribution works reasonably well when the purchase decision is simple and immediate. Someone searches for a product, clicks an ad, and buys. The last click was the relevant click.

Last click fails when purchase decisions involve research, comparison, and consideration. The last click might be a brand search from someone who discovered you through display advertising weeks earlier. Crediting only the search click ignores the display campaign that created awareness.

Use last click for direct response campaigns with short cycles. Use more sophisticated models for brand building, high consideration purchases, and long sales cycles. Match measurement complexity to purchase complexity.

The Metrics That Predict Future Growth vs Report Past Activity

Vanity metrics feel good but predict nothing. Impressions, reach, and clicks tell you what happened but not what it means. Commercial metrics connect activity to business outcomes.

Customer acquisition cost matters because it determines whether growth is profitable. Customer lifetime value matters because it determines how much acquisition cost is sustainable. Return on ad spend matters because it determines whether investment creates or destroys value.

But even these metrics can mislead without context. Low customer acquisition cost means nothing if acquired customers never repurchase. High lifetime value projections mean nothing if based on insufficient data. Return on ad spend in platform reports means nothing if the attribution is inflated.

The metrics that matter most are those connecting marketing activity to verified business outcomes. Revenue from customers acquired through marketing. Profit margin after marketing costs. Market share changes attributable to marketing investment. These are harder to measure but more meaningful than platform vanity metrics.

Build measurement around what you need to know to make better decisions. If you need to know whether to increase Google Ads budget, measure whether Google Ads creates incremental value at current spend levels. If you need to know whether brand investment is working, measure brand awareness and consideration changes over time. Match measurement to decision.