Right now, somewhere in your marketing team, someone is making a six-figure budget decision based on data that is fundamentally wrong.
I know that sounds dramatic. But after auditing dozens of 7 and 8-figure brands over the past few years, I can tell you this with absolute confidence... most attribution data is fiction. Comfortable, convincing fiction that tells you exactly what you want to hear. But fiction nonetheless.
Here's the uncomfortable truth: Meta says it drove the sale. Google says it drove the sale. Your email platform says it drove the sale. They're all looking at the same customer and all taking credit.
And you're making budget decisions based on that mess.
Let's fix that.
Jump ahead:
- The Triple-Counting Problem
- Let's Do The Maths
- The Three Layers of Attribution You Actually Need
- How to Run Your First Incrementality Test
- The 5-Step Attribution Audit Framework
- What To Do With This Information
The Triple-Counting Problem (And Why Your ROAS Is Probably Inflated)
I was speaking to a DTC founder last month. Scaling fast. Spending about £80k a month across Meta, Google, and email. Their blended ROAS across all platforms? A healthy-looking 5.2x.
But something didn't add up. When we looked at actual revenue growth versus total ad spend growth... the numbers were off. Way off. Revenue had grown 30% year-on-year, but their total attributed revenue across all platforms had apparently grown 65%.
Where was the extra 35% coming from?
Nowhere. It was phantom revenue. Every platform was claiming credit for the same conversions.
This is what I call the triple-counting problem. And according to recent research, the gap between platform-reported ROAS and true incremental ROAS often reaches 2-3x. For some channels like branded search and retargeting, that inflation can be 5-10x.
Read that again.
Your retargeting campaigns might be taking credit for 5-10x more revenue than they're actually generating.
Let's Do The Maths (Because Nobody Else Will)
Let me walk you through a real scenario I see constantly.
Say you're spending £50,000 a month on paid channels:
- £25,000 on Meta prospecting
- £10,000 on Google Shopping
- £10,000 on branded search
- £5,000 on Meta retargeting
Your platform dashboards tell you:
- Meta prospecting: £100,000 revenue (4x ROAS)
- Google Shopping: £50,000 revenue (5x ROAS)
- Branded search: £80,000 revenue (8x ROAS)
- Meta retargeting: £40,000 revenue (8x ROAS)
Total attributed revenue: £270,000. On £50k spend. Looks incredible, right?
But your actual total revenue for the month is £180,000.
There's £90,000 in phantom attributed revenue floating around. That's a 50% inflation. And if you're making budget decisions based on those numbers... you're flying blind.
Here's what's really happening. A customer sees your Meta ad on Monday. Searches your brand name on Wednesday (Google takes credit). Gets a retargeting ad on Thursday (Meta retargeting takes credit). Opens an email on Friday and buys (email takes credit). The original Meta prospecting ad also takes credit through its attribution window.
One customer. One purchase. Four platforms claiming the win.
Now here's the question that matters... which of those touchpoints actually caused the purchase? Which ones were this customer going to do anyway? That's where most brands completely fall apart.
The Three Layers of Attribution You Actually Need
Most brands operate on just one layer. The smart ones use all three. Here's the framework we use at Elevate when we're trying to get a genuine picture of what's working.
Layer 1: Platform Attribution (What You're Probably Using Now)
This is your Meta Ads Manager, Google Ads dashboard, Klaviyo reporting. It's useful for optimising within a channel. Which ad creative is working? Which audience segment converts best? Which email subject line gets clicks?
What it's good for: Tactical, within-channel decisions.
What it's terrible for: Cross-channel budget allocation. Understanding actual business impact.
Don't throw it away. Just stop treating it as gospel.
Layer 2: Blended Metrics (The Reality Check)
This is where you zoom out and look at the business, not the channels. The metrics that matter here:
- Blended CAC: Total marketing spend divided by total new customers. No platform attribution involved. Just cold, hard maths.
- MER (Marketing Efficiency Ratio): Total revenue divided by total marketing spend. Again, no attribution gymnastics needed.
- Contribution margin after marketing: What's actually left after you've paid for the ads and the product?
I've seen brands with "amazing" platform ROAS that are actually losing money on a contribution margin basis. The attribution data told them everything was working. The bank account told a different story.
If your platform ROAS is 5x but your MER is 2x... you have a £90,000 honesty gap. And that gap is where bad decisions live.
Layer 3: Incrementality Testing (The Gold Standard)
This is the one that changes everything. Incrementality testing answers the only question that actually matters... "Would this sale have happened anyway without this marketing activity?"
According to eMarketer's latest research, over 52% of brand and agency marketers now use incrementality testing. And the brands that implement it properly report 15-30% improvement in marketing efficiency by cutting non-incremental spend.
Let that sink in. 15-30% more efficient. Not by spending more. By spending smarter.
How to Run Your First Incrementality Test (Step by Step)
This doesn't have to be complicated. Here's the exact process I'd recommend if you've never done this before.
Step 1: Pick Your Biggest "Sacred Cow" Channel
You know the one. The channel everyone in your team believes is untouchable. The one with the amazing ROAS that nobody dares question. For most brands, it's branded search or retargeting.
Start there. Because that's where the biggest surprises hide.
Step 2: Design the Holdout
The simplest version is a geo-holdout test. Pick a region where you turn off (or significantly reduce) spend on that channel. Keep everything else the same. Run it for 2-4 weeks minimum.
For example... if you're running branded search nationally, turn it off in one region (say, the Midlands) and keep it running everywhere else. Then compare.
Step 3: Measure What Actually Changes
Don't just look at the channel metrics. Look at:
- Total revenue in test region vs. control region
- Total new customer acquisition in both regions
- Direct traffic changes
- Organic search changes (did organic pick up the slack?)
Step 4: Calculate Incrementality
The formula is straightforward:
Incremental lift = (Revenue with ads - Revenue without ads) / Revenue with ads
If you were spending £10,000 a month on branded search claiming £80,000 in revenue, and when you turned it off the test region only lost £12,000 in revenue... your true incremental ROAS isn't 8x. It's 1.2x.
That's not a rounding error. That's a fundamental difference in how you should be allocating budget.
Step 5: Reallocate With Confidence
Take the savings and reinvest them into channels that show genuine incrementality. Usually, that's top-of-funnel prospecting, content, and building organic presence. The boring stuff that doesn't look flashy in platform dashboards but actually drives real business growth.
A CPG brand that ran this exact process discovered their heavy TV-influenced digital attribution showed 35% of conversions, but incrementality testing revealed only 8% incremental lift. They reallocated that budget and saw overall efficiency improve by 22% in the following quarter.
The 5-Step Attribution Audit Framework (Steal This)
Here's the exact framework we walk clients through at Elevate. You can run through this yourself in a day.
1. Sum your platform-reported revenue across all channels. Compare it to your actual revenue. If attributed revenue is more than 130% of actual revenue, you have a serious triple-counting problem. (Most brands sit at 150-200%. Some are worse.)
2. Calculate your MER and blended CAC. These are your ground-truth metrics. If they tell a different story than your platform dashboards, trust the blended numbers.
3. Map your customer journey properly. Not what you think it looks like. What it actually looks like. Use post-purchase surveys ("How did you first hear about us?"), GA4 path analysis, and proper customer journey mapping. You'll be surprised. The journey is almost always messier and longer than your attribution model suggests.
4. Identify your highest-risk channels. Any channel with a ROAS that looks too good to be true probably is. Branded search and retargeting are almost always inflated. Flag them for incrementality testing first.
5. Run incrementality tests on your top 2-3 channels. Start with the sacred cows. Work your way through. Build a genuine understanding of what each channel is actually contributing.
This is the kind of holistic analysis that most agencies simply don't do. They'll optimise your Meta campaigns in isolation, or tweak your Google Shopping bids, but nobody asks the fundamental question... "Is this channel actually driving growth, or just taking credit for it?"
That question requires looking at paid, organic, CRO, email, and direct traffic together. Not in silos. Together. And that's where most teams and agencies fall short.
What To Do With This Information (Starting Monday)
Look. I get it. This is uncomfortable. Nobody wants to find out that the channel they've been championing for the past 18 months is less effective than they thought.
But here's the thing...
Brands that waste an average of 23% of marketing spend on non-incremental activities are essentially burning cash to make their dashboards look pretty. As of 2026, with rising CPMs and increased competition across every platform, you simply cannot afford to keep doing that.
Here's your action plan for next week:
- Monday: Pull your total attributed revenue across all platforms for last month. Compare it to your actual revenue. Calculate the gap.
- Tuesday: Calculate your blended CAC and MER. Write them on a sticky note and put them next to your monitor. These are now your north star metrics.
- Wednesday: Identify your top 2 "sacred cow" channels. The ones with suspiciously high ROAS that nobody questions.
- Thursday: Design a simple geo-holdout test for one of those channels. Pick your test region. Set a start date.
- Friday: Brief your team. Get buy-in. Start the test the following Monday.
That's it. Five days. No expensive attribution platform needed. No complex statistical models. Just a willingness to ask uncomfortable questions and follow the data wherever it leads.
If you're running a DTC brand at scale and your attributed revenue is significantly higher than your actual revenue... you don't have an attribution problem. You have a decision-making problem. And the sooner you fix the data, the sooner you can fix the decisions.
The brands that will win in the next 12 months aren't the ones spending the most. They're the ones who know, with genuine confidence, where their money is actually working.
Be one of those brands.
