· By Nick Fernandez
The Incrementality Test Every Agency Should Run Before Budget Season

Key Takeaways
- Understanding the incrementality of your paid media will allow you to understand what conversions are truly new incremental sales.
- Understanding incrementality vs. attribution in paid media will help avoid overpaying in highly competitive channels for simply retargeting existing customers.
- Using IAB Project Eidos standards allows you to maintain consistency in cross-channel measurement in open web display and programmatic.
- Adhering to ANA benchmarks means you can avoid paying for poor-performing AI slop and non-converting viewable inventory.
- By using Ad.com Display, you can scale your inventory reach. With 30 billion monthly searches and 1.4 trillion bid requests across premium publishers.
As agency executives prepare for budget cycles, it is key to prove that spending in Paid Media is generating incremental sales for clients. Historically, Multi-touch and platform dashboards have been used in media to justify budgets. However, due to complexities in today’s programmatic supply chains, high viewable rates and low CPM’s may not equate to business growth. To win in the next budget cycle, you should implement strict incrementality testing in your paid media.
In this article, we review differences between incrementality and attribution, recent findings in 2026 from the IAB and ANA quality benchmarks, and provide a practical example of a holdout test. Finally, we explore how Ad.com Display allows brands to reach high-quality and transparent inventory across thousands of verified websites.
What is a Paid Media Holdout?
Paid Media holdouts allow you to understand the true incremental revenue lift a media channel is providing. A control set is randomly created, and the test group is shown your ads, while the control group is either not shown any ads or shown PSA placeholder ads.
Once the test period has finished, you can compare conversion rates in each group. This lets you understand the incremental sales lift you received; in other words, exactly how many extra sales were made purely from media exposure.
Running a paid media holdout test before the budget cycle means you can avoid double-counting, identify if any organic conversions have been lost, and gain evidence to move the budget to high-performing channels.
Media Incrementality Versus Attribution: The 2026 Industry Shift
Historically, multi-touch and last-click attribution models have been used to measure value across multi-touch points in paid media. However, as seen in updates from the industry in late 2026, using standard attribution models in paid media can mean inefficient budgets.
The IAB State of Data 2026 report shows that up to 75% of buy-side advertisers believe that current leading advanced measurement models lack rigor, timeliness, and credibility. To resolve this, IAB launched Project Eidos, which aims to develop consistent measurement taxonomies and frameworks across multiple channels so we can accurately attribute business results to media exposure.
At the same time, the ANA Programmatic Transparency Benchmark (Q1-Q2 2026 Update) found that high-performing media buyers convert 54% of their programmatic spend into quality impressions; however, those with lower performance are wasting budget on poor-quality inventory. 1.3%-2.4% of open web programmatic spend is wasted on low-quality AI slop content. This means that even though these sites have high traditional viewability rates, they still have very inefficient inventory.
According to guidance from the Media Rating Council (MRC), display ads are considered viewable if at least 50% of pixels in an ad are visible for 1 second. However, having high viewability does not mean that the ad has driven qualified leads.
"As marketing budgets face increased scrutiny, measurement frameworks must provide timely, credible insights to justify spend and guide optimization... Modernizing measurement standards allows buyers to separate true incremental outcomes from inventory noise." – IAB Measurement Advisory Committee Panelist. (IAB State of Data 2026).
Measurement Attribute | Standard Attribution | Incrementality (Holdout) |
Purpose | Attribute Value | Net New Revenue |
Bias | Biased towards last-click/retargeting | No organic conversion overlap |
Data Used | Third-party cookies and pixels | Matched market tests |
Impact on Budget Decisions | Encourage overpaying for existing customers. | Encourage investment in new customers. |
Incrementality Test Plan (for Agency Executives)
To protect media spend, executing a holdout test in your Open Web Display and Programmatic channels will provide valuable insight for budget planning. The following 4 steps can be implemented:
Within the incrementality test, split your total target audience into 2 groups. Within the ‘test region’ (80% of your target audience), campaign ads will be shown. From this, a total conversion volume can be calculated. Within the ‘control region’ (20% of your holdout audience), no ads will be shown, meaning a base conversion rate can be measured. Multiplying the conversion rate in the control region by 4.0 and taking this away from the conversions in the test region will give agency analysts the true incremental revenue lift generated by the campaign.
- Match Market Regions: Select 2 separate regions. Ensure they match in conversion rates. (For example, matching DMAs—Charlotte & Raleigh). Assign 80% of target audience reach to the test region and 20% to the control region.
- Run campaign: Within the test region only, run your active display campaign. Do not show any ads within the control region. Run for 30 days.
- Calculate Conversions: After 30 days, calculate conversions in each region (per 100,000 people):
Incremental % = (Conversion Rate Test - (Conversion Rate Control / Conversion Rate Control)) x 100
Reallocate Spend based on CPIC: Calculate the CPIC of your campaign by dividing total spend by net new conversions. Move spend away from channels with high CPIC values and into proven incremental inventory.
Counterpoint and Strategic Limitations
Although geo-matched paid media holdouts give clear evidence of a campaign’s impact, there are some limitations to consider.
Limitation
Matched-market paid media holdouts require a significant amount of base conversions and strict separation of signals across channels. If tests run for less than 30 days, external factors within the market, such as specific local promotions, competitive spending, or supply chain issues may cause inaccuracies in the control conversion rates. It is important to verify statistical significance before making large changes to budgets.
5 Top Questions to Consider Before Budgeting
Before finalising media plans, agency executives should verify that assumptions being made in measurement are correct. To ensure budgets are invested in actions that bring tangible business results, consider the following questions:
What % of conversions are organic?
Consider whether the performance you are seeing in your campaigns is actually incremental or if the network is attributing conversions from existing loyal customers.
How are we removing AI Slop and poor inventory from our programmatic spend?
Ensure you are running supply path optimisation within your campaigns to only show ads on curated publisher domains. Consider if you may be showing inventory on click farms.
Are we measuring CPIC in addition to CPA?
Update metrics in your company to focus on incremental acquisition cost to avoid overinvesting in saturated retargeting.
How is our attribution model accounting for ZC search and AI Overviews?
Verify that you are effectively measuring brand awareness in AI Overview channels. Even though conversions in these channels may be lower, they should be included in your measurement strategy.
Have we validated the baseline of our media?
Ensure that you have validated historical data, using a matched market holdout method. Do not just rely on platform dashboards.
Increase Transparency in Display with Ad.com
To improve incremental ROI, it is important to reach high-intent audiences on brand-safe publishers. Ad.com Display provides an enterprise solution to deliver verified performance.
- Global Reach: Access 30 Billion Monthly Searches and 1.4 Trillion Bid Requests across the open web.
- Trusted Publishers: Integrate directly with 1 million trusted publishers and 150,000+ premium websites. Provide 100M+ daily impressions.
- Engaging Formats: Drive quality engagement with engaging display, native, and contextual units.
- Proven Performance: Compared to legacy display networks, performance and ROI are 31-57% higher on Ad.com.
Using contextual supply with transparent programmatic execution, Ad.com Display helps agencies remove waste in media spend and gain incremental customer intent.
Conclusion
Moving from attribution to incrementality testing will greatly increase efficiency in your media budgets. Within this article, we have explored how IAB Project Eidos and ANA quality benchmarks highlight issues with relying on last-click metrics. We have provided a practical example of a geo-holdout test to implement in your media and explored ways to avoid spending on low-quality inventory. By using holdout tests to verify campaign performance and increasing inventory reach via platforms like Ad.com, agency executives can enter budget cycles with proven strategies to grow client businesses.
Want to start measuring real incremental sales on premium publishers? Try Ad.com Display solutions to get access to 30 billion monthly search interactions and drive growth for clients.
Frequently Asked Questions
How does IAB Project Eidos impact measurement in agencies?
Project Eidos aims to produce consistent taxonomies and data standards across programmatic display, CTV, and digital video to enable consistent analysis of incrementality, attribution, and marketing mix modeling across various publisher environments.
Why is CPIC a more accurate metric than CPA?
CPIC only counts conversions that have been directly influenced by the advertisement, removing any organic conversions. CPA means that media spend is credited for all conversions attributed to a campaign. This can inflate performance figures by attributing conversions from existing loyal customers and organic search traffic.
How does AI Slop affect quality in programmatic display?
‘AI Slop’ is a mass-produced, very low-quality content site designed to make money from advertising. According to ANA benchmarks, AI Slop makes up up to 2.4% of programmatic open web spend. These sites may pass basic MRC viewability requirements but have a very low conversion rate. If not filtered out in supply path management, agencies are wasting media spend on these sites.
What size sample is required for a reliable matched market holdout?
To ensure statistically significant results, a minimum of 15-20% of total target audience reach is required in markets over 30 days. This ensures that small variations in sales between regions won’t impact the accuracy of incremental lift calculations.
Can Viewability be optimized without impacting page performance?
Yes. Using viewable lazy loading, responsive ad placement, and fixed sticky placements means that ads can be shown on the page without affecting page performance. The ad will only appear once the user has scrolled far enough to see it. This can provide high viewability figures (MRC standards) with no degradation to page performance metrics (LCP).


