Connected TV Trends & Blogs | Insight | tvScientific

CTV Incrementality 101: How to Run a Lift Analysis | tvScientific

Written by tvScientific | Sep 25, 2025 1:00:01 PM

TL;DR: Your Cheat Sheet to Incrementality Testing

Incrementality testing measures the true lift your CTV ads generate, separating real impact from background noise. With Prime Day around the corner, running a test now helps you prove your ads are driving incremental conversions — not just capturing purchases that would have happened anyway. Platforms like tvScientific make it simple to set up, measure, and optimize with built-in, always-on incrementality testing.

 

A Step-by-Step Guide to Incrementality Testing for Performance TV Advertisers

Amazon Prime Day has become one of the largest retail moments of the year, with billions in sales happening in just a couple days. In fact, July’s Prime Day matched Cyber 5 revenue for the first time, pulling in a whopping $14 billion on Amazon alone. 

This October’s Prime Day is expected to follow this trend, with shoppers primed (pun intended) to spend, and advertisers competing harder than ever to capture attention.

But with ad spend surging, how do you know if your CTV campaigns are truly driving incremental conversions — or just riding the wave of increased consumer demand?

That’s where incrementality testing comes in. By isolating the real lift your ads generate, you can:

  • Prove your ads influenced Prime Day purchases, not just organic shopping behavior
  • Optimize spend in real time to push more sales
  • Build confidence in your CTV investment ahead of other seasonal peaks like Black Friday and Cyber Monday

Let’s break down exactly how to run an incrementality test for your CTV campaigns.

What is Incrementality Testing?

Incrementality testing in advertising is an approach to performance measurement that determines the exact level of contribution — or “lift” — a channel’s advertising has on audience behavior. It relies on two groups:

  • The test group: Exposed to your ad
  • The control group: Not exposed to your ad

If the exposed group performs better, the test shows a positive incremental lift. If the control outperforms the exposed group, it’s negative incremental lift. And if there’s no difference, the test is neutral.

A difference between the results is only consequential if that gap is statistically significant.  

Think of it as a specialized form of A/B testing. Instead of asking “Which ad works better?”, incrementality testing asks “Does advertising here work at all, and by how much?

How to Run Incrementality Testing for CTV Advertising

Step 1: Create a Hypothesis and Establish Performance Metrics

It’s surprising how often advertisers can jump into any testing process before they’ve clearly established what it is they’re trying to test or how they’re going to assess success. So, start by clarifying what you want to learn.  

With incrementality testing, you’re going to uncover whether an advertisement works better than no advertisement. This means you could run a test using messaging and creative you believe is effective, but you could also run a test with placebo or dummy ads, which would measure the fundamental value of the channel.

On top of generating your hypothesis, you also want to choose the performance metrics you’ll measure against. Common metrics include:

  • Purchases
  • Website visits
  • Average order value (AOV)
  • Conversion rate 

Step 2: Define Your Testing Audience

Next, you need to build a testing audience derived from your overall target audience. Create two groups, just as we mentioned above:

  • Test group: Will see your ads
  • Control group: Won’t see your ads

When you’re defining this subset, you want to make sure both groups are as similar as possible and account for any variables that could have an impact on the outcome of your test. For example, if you’re selling luxury products, keep income levels consistent. Otherwise, your results could be skewed.

For advertisers looking for precision audience control during testing, tvScientific’s CTV advertising and attribution platform provides numerous targeting parameters. These include age, gender, education, income, household, zip code, brand affinity, online purchases, offline purchases, viewing habits, device, business vertical, and 15,000 more segments. This level of granularity provides advertisers with complete control over their incrementality testing process.

Step 3: Determine Timeframe and Ad Content, and Run Your Test

Once your audience is defined, the next step is deciding when to run your test and what type of creative to use. Here are a few best practices to keep in mind.

  • Duration: While the timeframe of the test depends on the volume of interactions (not any specific amount of time), it’s a good idea to run your test for at least a week, or until you have enough conversions to be statistically confident.
  • Timing: Consider running your test during a relatively normal time of the year. Avoid periods with unusual buyer behavior (i.e. holidays, elections, etc.) unless you’re specifically testing for them. These moments could impact your audience’s state of mind in unpredictable ways and undermine the validity of your results.
  • Creative: If you’re looking to measure the raw, incremental value of CTV advertising, you need to run placebo creative during your test. This creative needs to be irrelevant to your business, whether it’s a placebo or a public service announcement (PSA) ad. By running this kind of test campaign, you’ll be able to get an accurate handle on the baseline value of CTV advertising.

With these done, it’s go time. Flight your ad to your test group while suppressing the control group’s exposure.

Step 4: Analyze Test Data and Calculate Incrementality

When the test ends, compare the performance of your test versus control groups side-by-side. Look for differences in your key metrics (purchases, AOV, etc.).

From there, calculate:

  • Incremental lift (%): The performance gap between test and control
  • Incremental cost per acquisition (iCPA): The cost of each additional conversion driven by your ads
  • Incremental return on ad spend (iROAS): The revenue generated from those additional conversions, compared to your ad spend

With these figures in hand, you now have a clear, data-backed view of the real impact your CTV campaigns are driving.

Finding a Platform that Supports Incrementality Testing

While the impact of traditional TV ads has always been difficult for advertisers to pin down, the arrival of CTV advertising means those days are behind us. Now, advertisers can measure the value of their ads with the kind of precision that had only been possible on digital channels. 

Brands no longer have to wonder if their TV ad campaigns are working. All they need to do is run incrementality testing on their CTV advertising platform to find out.

tvScientific is the only CTV platform with a built-in, always-on incrementality testing feature that continuously compares your campaign’s performance to a statistical baseline. Unlike walled gardens, it provides full transparency into incremental lift and ROI while ensuring your campaigns deliver fraud-free results. tvScientific’s AI driven optimization ensures you’re not just getting attributed conversions, but true incremental outcomes. 

Ready to prove your CTV campaigns are driving incremental sales this Prime Day? Reach out to our team and get started.

Performance TV Incrementality Testing FAQ

Why is incrementality testing better than last-touch attribution? Last-touch can over-credit other channels or organic activity. Incrementality testing isolates the true effect of your ads.

How long should an incrementality test run? It should last at least one week, but the exact duration depends on traffic and conversion volume needed for statistical significance.

Can I run an incrementality test during Prime Day itself? Yes, but it’s even more valuable to test in the weeks leading up to Prime Day, so you can optimize and allocate spend confidently before the surge.

Do I need a large budget to run incrementality testing? Not necessarily. With a platform like tvScientific, advertisers of all sizes can run tests with flexible budgets and still get meaningful results.