How to Measure Paid Media: Why Attribution Falls Short and Incrementality Wins

Marketers have never had more data at their fingertips, yet one of the toughest questions still remains: “How do I measure paid media?”

For years, attribution models have been the go-to answer. Multi-touch attribution (MTA), last-click, first-touch—you name it. These models promised clarity: assign credit for a sale to the right channel and optimize accordingly.

But here’s the problem: attribution doesn’t tell you the whole story. In fact, it often tells you the wrong story.

The better way forward? Measuring paid media through incrementality. Let’s break down why attribution is misleading, why incrementality matters, and how to put it into practice.

Why Attribution Sounds Good but Doesn’t Work

Attribution models attempt to answer: Which ad gets credit for this conversion?

Take a simple example:

• A customer sees a Facebook ad on Monday.
  
• They click a Google ad on Wednesday.
  
• They buy directly from your website on Friday.
  

Now—who gets credit? Facebook? Google? Both?

Different models give you different answers:

• Last-click attribution: Google gets 100%.
  
• First-touch attribution: Facebook gets 100%.
  
• Multi-touch attribution: They split the credit.
  

At first glance, this feels neat and organized. But here’s the flaw: attribution assumes that without the ad, the purchase wouldn’t have happened.

That’s a huge assumption.

In reality, many people who see or click ads would have converted anyway. Attribution doesn’t distinguish between “caused by the ad” and “would have happened no matter what.”

Example: Branded Search Ads

Imagine someone searches for your brand name—“Nike shoes”—clicks your Google ad, and buys. Attribution will tell you the ad drove the sale.

But would that person really not have bought Nike shoes without the ad? Highly unlikely. That’s a waste of ad spend disguised as performance, and as a result, you’re likely over-invested in branded search.

The Incrementality Approach

Instead of asking “Which ad gets credit?” incrementality asks:

👉 “What extra value did my ads create that would not have happened otherwise?”

Incrementality measurement compares a test group (exposed to ads) with a control group (not exposed) to see the lift in conversions.

It tells you:

• If people who saw your ads bought more than people who didn’t.
  
• By how much.
  
• Which campaigns or channels actually move the needle.
  

Example: Facebook Campaign

Let’s say you spend $10,000 on Facebook ads.

• Attribution says: 1,000 conversions came from Facebook, so your CPA is $10.
  
• Incrementality test says: Only 400 of those conversions wouldn’t have happened without ads. Your true incremental CPA is $25.
  

Which number is more useful? The second one. Because it tells you what’s actually being added by your spend, not just what’s being claimed.

Why Incrementality Beats Attribution

Here are three big reasons incrementality is the superior way to measure paid media:

1. It cuts through the noise. Attribution models inflate numbers because they count every touchpoint as a driver. Incrementality strips it down to what’s real.
   
2. It stops rewarding waste. Channels like branded search or retargeting often look amazing in attribution reports. Incrementality testing exposes when they’re just harvesting conversions that were going to happen anyway.
   
3. It guides smarter budget allocation. If you know which campaigns actually drive net-new growth, you can scale those and cut the rest.
   

How to Start Measuring Incrementality

Incrementality doesn’t have to be complicated. Here are three practical ways to get started:

1. Geo-testing. Run ads in one region, pause them in another, and compare results.
   
   • Example: Run Facebook ads in Texas but not Florida for two weeks. Measure sales lift.
     
2. Holdout groups. Exclude a percentage of your audience from seeing ads and compare outcomes.
   
   • Example: Withhold 10% of your retargeting audience from ads and see if they buy at the same rate.
     
3. Platform lift studies. Platforms like Meta and Google offer built-in incrementality tests. While they’re not perfect (since they’re grading their own homework), they’re better than attribution alone.
   

Bringing It All Together

If you’re still relying on attribution to measure paid media, you’re probably overestimating the value of your spend. Attribution can’t answer the most important question: “Did my ads actually drive incremental growth?”

Incrementality, on the other hand, shows you the truth. It’s not about assigning credit, it’s about proving value.

So the next time you’re evaluating your paid media performance, remember:

• Attribution will make you feel good.
  
• Incrementality will make you smarter.
  

And if you want to know how to measure paid media in a way that actually grows your business, incrementality is the answer.