Overview
Exclusion Signals identify users with low or negative predicted incrementality. These users either do not convert at all or convert regardless of ad exposure — in both cases, excluding them from campaigns prevents spend on users where advertising has no measurable effect. This guide covers the setup of Exclusion Signals on Meta, including configuration, testing, and scaling.Prerequisites
Technical requirements:- An active connection to your Meta Ads account
- Sufficient data history for signal generation (30 days minimum recommended)
How Exclusion Signals Work
Innkeepr ranks anonymous visitors by predicted incremental lift for the objective you select — conversion probability, customer lifetime value, or average order value. Visitors at the bottom of that ranking show low or negative treatment effects: advertising has no measurable impact on whether they reach the objective. What sets this apart from traditional customer exclusions is where the ranking starts. Innkeepr scores visitors on the anonymous level — no login, no email, and no CRM match required. Behavioral first-party signals are enough to identify someone as non-incremental long before they identify themselves, which is where the vast majority of paid traffic sits. The same logic extends to identified customers: a visitor who has already fired a purchase event may be very likely to buy again, but that likelihood is often organic rather than ad-driven. Anonymous non-incremental traffic and existing customers end up in the same place — excluded — for the same reason: the ad does not change the outcome.Setup Instructions
Step 1: Create Exclusion Signals Exclusion Signals are built from different user segments, each with its own lookback window. The segments identify groups of users where ad exposure is unlikely to have an incremental impact. There are different ways to configure Exclusion Signals, and not every setup works equally well across all accounts. The optimal configuration depends on factors such as conversion volume, product category, and overlap with your existing audiences. As of March 2026, we recommend the following segments as a starting point. Recommended segments: Start with three visitor segments at different lookback windows, plus one purchaser segment. Splitting visitors across multiple lookback windows lets you see which recency band drives the most incremental value and adjust each one independently.
Each segment is created in Innkeepr and uploaded to Meta. After the first upload, Innkeepr sends
innkeepr_targeting events to your connected Meta pixel. You need to approve these events in Meta Events Manager before Meta processes the signal data. Without this approval, Exclusion Signals will not populate.
Step 2: Set Up the Campaign Structure
Test Exclusion Signals within a single campaign using separate ad sets. Both the control and the test ad set use the same base targeting and your existing default exclusions. The test ad set adds the Exclusion Signals on top.
Campaign structure:
Important: Keep all other variables identical across ad sets:
- The same creatives (images, videos, copy) in every ad set
- The same optimization event (e.g. Purchase, Add to Cart)
- The same bid strategy
- The same placements (or all set to Advantage+)
- The same base targeting — the only difference is the added Exclusion Signals
Monitor & Optimize
Weeks 1–2: Observation phase- Do not adjust budgets, bids, or creatives during the initial learning phase
- Check that both ad sets are spending their allocated budget and exiting the learning phase
- Verify that conversions are tracked correctly across both ad sets
- Compare ROAS, new customer share, and AOV between the control and the test ad set
- The impact of Exclusion Signals depends on the objective and the business model. Exclusion Signals can improve ROAS by removing low-incrementality spend, but they can also shift the new customer ratio — for example, excluding past purchasers concentrates budget on acquisition
- If the test ad set reduces conversion volume significantly without improving any of these metrics, the exclusion segments may be too broad — consider adjusting the lookback windows
- Apply the Exclusion Signals to your other active campaigns on Meta
- Combine multiple segments — you can apply all exclusion segments to the same ad set simultaneously
- Monitor performance after rollout — allow one to two weeks for each campaign to stabilize
- Exclusion Signals can be combined with Seed Signals in the same campaign — apply seed-based LALs as targeting and Exclusion Signals as exclusions on the same ad sets
Best Practices
Creative isolation The test is only valid if the creatives are identical across all ad sets. If you want to test new creatives or add additional creatives during the test, add them to all ad sets at the same time. Never add creatives to only some ad sets — this makes performance differences between ad sets unattributable. Combining with Seed Signals Exclusion Signals and Seed Signals serve complementary purposes: Seed Signals direct spend toward high-incrementality users, while Exclusion Signals remove low-incrementality users from the targeting pool. Using both at the same time is the recommended configuration once each has been validated individually. Lookback window tuning The recommended lookback windows (30 days, 30–90 days, and 90–180 days for visitors; 360 days for purchasers) are starting points. Adjust them based on your account:- Shorter lookback windows produce smaller, more conservative Exclusion Signals
- Longer lookback windows exclude more users but may include users whose behavior is no longer relevant
- Set the purchaser lookback window to match your new-customer definition — this determines how long a past buyer stays excluded from acquisition campaigns
Troubleshooting
Exclusion Signals are not populating
Exclusion Signals reduce conversion volume without improving efficiency
No measurable difference between test and control
Performance drops after sales phases
Performance drops can coincide with sales events such as Black Friday or seasonal promotions. During and after these phases, traffic composition and conversion patterns shift temporarily. Innkeepr typically detects these phases automatically and adjusts the underlying models accordingly.
If a performance drop persists for more than one to two weeks after a sales event has ended, reach out to the Innkeepr team to verify that the model has recalibrated correctly.