Limited Ads Explained: Earning From Visitors Who Decline Cookies

The Traffic Nobody Talks About

Publishers spend a lot of energy on consent rates, and rightly so. But even a well-designed banner leaves a share of EU and UK visitors who decline device storage or never answer. What happens to the ad slots shown to those visitors is often left to defaults nobody has checked. On sites using Google’s ad products, the answer is frequently limited ads — or, when the setup is wrong, no ads at all.

This article explains what limited ads are, how they differ from personalised and non-personalised ads, and how to make sure your unconsented traffic is still earning what it can.

Three Serving Modes, Three Sets of Conditions

Google’s publisher products distinguish between three modes:

The important point: non-personalised ads are not the fallback for users who reject cookies. If Purpose 1 is not consented, limited ads are the mode intended to keep serving.

When Limited Ads Can Serve

Google’s help documentation sets out the exact conditions for limited ads, and they depend on what the TC string contains for Google as a vendor and for the purposes Google relies on. In broad terms, limited ads become possible when there is no consent for Purpose 1, a valid TC string is present, and the other conditions Google lists are met. Read the current version of that documentation before relying on any summary, including this one, because the conditions are specific and change over time.

Two practical implications follow:

  1. You need a valid TC string. If your CMP fails to load, or the string is malformed, Google may not be able to serve anything for that impression. A broken CMP is costlier than a rejected banner.
  2. Vendor configuration matters. If Google is missing from your vendor list, or the user has objected to the processing Google needs, limited ads may not be eligible.

What Limited Ads Do to Revenue

Limited ads earn less than personalised ads, often substantially less, because they cannot use frequency capping, audience data or cookie-based measurement. They also draw from a narrower set of demand: buyers who need identifiers to bid will not compete for these impressions. Exact effects depend on your audience, formats and demand mix, so measure on your own traffic rather than assuming any benchmark.

Even so, a low-value impression is better than an empty slot. For sites with large EU audiences, the difference between “limited ads served” and “nothing served” on unconsented traffic can be meaningful.

How to Check Your Setup

1. Measure the unconsented share

Use your CMP analytics to find the share of EU and UK pageviews without Purpose 1 consent, including users who never answered. That is the inventory at stake.

2. Compare fill by consent state

Break down ad requests and fill rate by consent state in your reporting. If fill for unconsented traffic is near zero, something is blocking limited ads.

3. Inspect a real ad request

Reject all in a clean browser profile, load a page and inspect the ad request in DevTools. Confirm a TC string is attached and that the request reaches the ad server. Then check whether a creative renders.

4. Check your vendor list

Confirm that Google is included in your CMP vendor configuration and that you have not removed it by accident while trimming vendors.

5. Review other demand

Header bidding partners and other exchanges have their own rules for impressions without consent. Some will not bid; some may still bid with contextual signals. Ask each partner what they do, and avoid sending requests that will always come back empty, since they add latency without revenue.

Making Unconsented Inventory Work Harder

Key Takeaways

← Blog Read All →