Programmatic Media Quality Explained: A Guide for Healthcare Marketers

Most pharma media teams get a quality report every month. It usually shows a viewability rate, an invalid traffic rate, maybe a brand safety block rate, and a line that says everything is within benchmark. Then someone pulls the domain list and finds a few hundred sites nobody has heard of taking a real slice of the budget. That gap between the summary and the spend is what this series is about.

This guide explains the main pieces of programmatic media quality in plain language and points to a shorter explainer on each one. The goal is that a brand lead, an analytics lead, or a new planner can sit in a quality review and know which questions matter.

What programmatic media quality actually means

An impression passes through several hands before it reaches a screen. A publisher makes ad space available, one or more SSPs and resellers pass it along, your DSP bids, a verification vendor watches, and a measurement partner may later tie the exposure to an outcome. Quality can break at each step. So it helps to think of quality as a stack of separate questions rather than one score.

A campaign can be strong on one layer and weak on another. A tight NPI list delivered on poor inventory is still a poor buy. Clean premium inventory shown to the wrong people is too. The existing piece on inventory quality vs. audience quality covers that split in more depth.

Where the money goes before it buys media

Before talking about the impression itself, look at the dollar. Some share of every programmatic budget pays for things other than ad space: DSP fees, data fees, verification, measurement, agency trading fees. People call this non-working media. The term gets used loosely, and the short definition page what is non-working media lays out what falls on each side of the line.

Non-working does not mean wasted. A verification fee that keeps you off fraud is arguably the best money in the plan. The waste worth hunting is working media that buys nothing useful, like impressions on sites built only to serve ads. That is a quality problem disguised as an efficiency win.

Supply paths and inventory: where the impression came from

Supply paths

The same impression on the same page can often be bought through several routes. Each reseller in the chain may take a cut, and each extra hop adds a chance for spoofing or misrepresentation. Supply path optimization, or SPO, is the work of cutting those routes down to the ones you trust.

Classic SPO tended to favor the cheapest or most efficient path. That makes sense on cost, but it can push you toward paths that are cheap for a reason. Quality path optimization adds quality signals to the decision: authorization in ads.txt, a clear seller identity in sellers.json, the number of schain hops, and the IVT and viewability seen on each path. The existing supply path optimization guide for pharma covers what to remove, and the piece on ads.txt, sellers.json, and schain explains how to read the files.

Real sites versus sites built to sell ads

Made-for-advertising, or MFA, sites are built to capture cheap traffic and resell it at a higher rate through dense ad layouts. They often look fine on standard metrics. Viewability can be high because ads are packed into the viewport. IVT can be low because the visitors are real people who arrived from a paid link and left quickly.

The ANA's 2023 programmatic transparency study reported that MFA made up about 21 percent of impressions in its sample. Your own number may be higher or lower, and the only way to know is to check the domain list. The article on made-for-advertising sites in pharma explains the signs and the exclusion options.

Viewability, attention, and invalid traffic

These three are the metrics most teams already receive. They are also the ones most often misread.

• Viewability uses the MRC standard: for display, at least 50 percent of pixels in view for one continuous second; for video, 50 percent for two continuous seconds. It is a floor, not a guarantee anyone looked.
• Invalid traffic splits into general IVT (known bots, data center traffic) and sophisticated IVT (harder fraud that needs deeper analysis). A low total rate can hide one bad path.
• Attention metrics try to estimate whether a person actually engaged with the ad. They are useful for comparing placements, but methods differ by vendor and there is no single accredited standard to lean on.

The explainer on viewability, attention, and IVT goes through each one, including what to ask a vendor.

Brand suitability and audience accuracy

Brand safety and suitability without losing scale

Pharma brands have good reasons to be careful about adjacency. The common mistake is overcorrecting. Broad keyword blocklists that include words like "cancer," "death," or "overdose" can block the exact health news and condition content where patients and HCPs spend time. Oncology brands sometimes block the word that describes their own category.

Suitability frameworks, such as the categories and risk tiers that came out of the GARM work (GARM itself shut down in 2024, but its framework is still widely referenced), let you separate content that is truly unacceptable from content that is sensitive but appropriate. The guide to brand safety and suitability for pharma covers how to set tiers and test them.

Audience accuracy: reaching the right person

For HCP campaigns, quality also means the person who saw the ad is the prescriber on your list. That depends on the identity graph that links an NPI to devices, emails, or cookies. Match rate tells you how many records were linked. It does not tell you how many links were correct. A partner can report a high match rate on a graph that is stale or loosely built.

The article on how to evaluate a third-party business identity graph covers truth-set testing, recency checks, and pilot design. The same logic applies to consumer health audiences.

Auctions, price, and pacing

Most open programmatic auctions now run as first-price auctions. You pay what you bid, so DSPs use bid shading to estimate the lowest price likely to win. Publishers set floors. Narrow HCP audiences mean many buyers chase the same small set of people, which pushes clearing prices up. A pharma buyer who sees high CPMs on a tight specialty list is often seeing real competition, not a mistake. The explainer on bid floors, bid shading, and auction dynamics explains what you can control.

Pacing connects here too. When a campaign overpaces, DSPs tend to win more, often on cheaper and lower quality inventory. When it underpaces, the fix people reach for is loosening quality controls. The piece on pacing in programmatic explains how to read a pacing report before you touch anything.

Quality outside the bid stream: AI answers

One newer quality question sits outside programmatic buying but lands on the same teams. AI search tools now answer health questions directly, and they sometimes get a brand wrong: outdated safety information, a dose from an old label, or a competitor's indication. The article on when AI answers misrepresent a healthcare brand covers how to monitor and correct the sources those answers draw on.

How to run a basic quality review

You do not need a new tool to start. A useful quarterly review can be built from reports you already have access to.

1. Pull spend by domain and app for the quarter. Sort by spend, not impressions.
2. Mark each of the top 200 domains as known publisher, unknown, or likely MFA. Look at the share of spend in each group.
3. Pull spend by SSP and by schain length. Check whether a few resellers carry an unusual share.
4. Compare viewability and IVT by path and by domain group as well as in total.
5. Review the blocklist and the block rate. If blocks are high on health news, revisit the keywords.
6. Check the audience side: when was the identity graph last refreshed, and when was it last tested against a truth set?
7. Write down the two or three changes you will make and how you will tell if they worked.

That last step matters most. Quality reviews often end with a list of concerns and no owner. Pick the changes, give them dates, and check the domain list again next quarter.

Practical takeaway

Ask your agency or trading desk for one report this month: last quarter's spend by domain and app, sorted by spend, with each row's viewability and IVT rate. Read the top 100 rows yourself. Whatever you find there will tell you which article in this series to read next.

Editorial note. Analysis and frameworks are the author's own and do not represent Acxiom or any current or former employer, client, or named platform. Examples labeled hypothetical or illustrative are not results from real campaigns. Nothing here is legal, regulatory, or medical advice.