Buyers Now Rank Targeting Over Content Quality: A First-Party Data Plan

Most publishers assume buyers pay for quality first: good journalism, safe context, a real audience. A new IAB survey suggests the order has changed. In its 2026 Digital Video Ad Spend and Strategy Report, released this month, targeting capabilities were the top criterion buyers use to decide where TV and video dollars go, cited by 49% of respondents. Content quality came second at 46%.

That gap is small, but the direction is the story. IAB says targeting rose 10 points year over year, and it links the jump to signal loss from IP address degradation and rising AI-driven traffic, which make reliable targeting harder. Small and mid-size advertisers, those spending under $50 million a year, drove the shift. IAB notes they are also the buyers most exposed to open-market buying, where identity problems are sharpest.

The survey is about video, but the logic applies to every format. When identifiers get weaker, buyers pay more for inventory that tells them who is likely on the other end. Publishers who know their readers well have an asset they are probably not sending into the auction. Finance publishers are a good example.

The headline numbers

For context, IAB and its research partners estimate U.S. digital video ad spend will reach $81.9 billion in 2026, up 11%, and that digital video will pass 60% of total TV and video spend for the first time. Money is still growing. What is changing is what buyers screen for when they allocate it.

What finance publishers already know about their readers

A personal finance or markets site learns a lot from behavior alone, without a login. Someone who has read four retirement-planning articles this month, used a mortgage calculator and opened a guide to refinancing is a fairly clear signal. So is a reader who checks market coverage every morning, or one who reads small-business banking and payroll content.

Today, most of that insight stays inside the analytics tool. The bid request carries a URL, maybe some page keywords, and whatever identifiers the browser allows. On Safari and Firefox, which limit third-party cookies, that is often not much. The buyer sees a finance page but not the fact that this reader is a repeat retirement-planning visitor.

Seller Defined Audiences, in plain terms

IAB Tech Lab built a specification for exactly this. Seller Defined Audiences (SDA) was released in February 2022 under Project Rearc, and Tech Lab followed with an implementation guide in October 2023. The idea is simple: the publisher groups its readers into audience segments, labels them with a shared taxonomy, and sends the segment IDs in the bid request instead of exposing raw user data.

The guide's key points for publishers:

• Build audiences from what you observe. Tech Lab's own example is content-based inference: someone who keeps reading soccer coverage is probably a soccer fan. Sign-up answers and on-site interactions work too, and a login is not required.
• Label them with IAB Tech Lab's Audience Taxonomy 1.1, which is recommended. If you use another taxonomy, you must declare which one and make sure buyers can translate the IDs before any campaign goes live.
• Declare what is in the segment. Tech Lab points to its Data Transparency Standard to describe how an audience was built, and strongly recommends its compliance program.
• Protect against leakage. The spec says platforms should avoid sending user or device identifiers and SDA segments in the same request, so buyers cannot attach your audience definitions to their own IDs and reuse them elsewhere. Cookieless browsers are where Tech Lab suggests testing first, because there is little identifier value to give up there.

How it travels through Prebid

If you run Prebid.js, the plumbing already exists. The Prebid first-party data feature lets you set ortb2.user.data with a list of segments and a segtax value that names the taxonomy. The OpenRTB segtax registry assigns 4 to Audience Taxonomy 1.1 and 7 to Content Taxonomy 3.0, so you can send both an audience signal on the user object and a contextual signal on site.content.data.

Prebid also supports setBidderConfig, which lets you send first-party data only to the bidders you choose. That matters. You do not have to hand your best segments to every partner on day one.

A four-week plan for a finance site

Week 1: define a small set of segments

Start with five to eight audiences that a finance advertiser would recognize and want: retirement planners, home buyers and refinancers, active investors, small-business owners, credit card comparison shoppers. Write the rule for each one in plain language, for example "read three or more retirement articles in 30 days." Map each to the closest Audience Taxonomy 1.1 node.

Keep away from segments that describe someone's financial situation rather than interests, such as debt trouble or low credit. Several state privacy laws treat financial information carefully, and buyers in regulated categories are cautious too. Have counsel review the list before launch.

Week 2: wire it up and check the bid requests

Compute segment membership on your side, store it in first-party storage, and write it into the Prebid config before auctions run. Then check real bid requests in your browser's developer tools or with your SSP to confirm the segments arrive with the right segtax. Honor consent signals: if a reader has opted out of targeted advertising, do not attach audience segments.

Week 3: test with two or three partners

Use setBidderConfig to send segments to a few SSPs that confirm they pass SDA through to DSPs. Keep a control group of traffic without segments. Compare bid rate, CPM and revenue per thousand sessions, split by browser. The cleanest read is usually on Safari, where there is no cookie to muddy the result.

Week 4: package it for direct and PMP sales

If a segment moves CPMs, turn it into a deal. A private marketplace built on "active investors on a finance site" is an easier sale than run-of-site, and it gives smaller advertisers, the same group driving the targeting shift in IAB's survey, a way to buy your audience without their own identity graph.

What not to expect

SDA is not a switch that raises every CPM. Some DSPs read seller segments well and some barely look at them. Frequency capping and measurement still rely on other signals, which the Tech Lab guide acknowledges. Treat this as a test with a control group, not a launch.

But the direction in IAB's data is clear. Buyers are paying for signal, and publishers who understand their readers are the natural source of it. A managed partner like HBDR can help with the Prebid wiring and the bidder tests; the audience insight has to come from you.