Method
Brand safety and placement quality
Two problems usually treated as one. The reputational risk is the smaller of them — the larger is spend on inventory where nobody could have seen anything.
Brand safety is generally framed as the risk of appearing next to objectionable content. That risk is real and rare. The adjacent problem is neither, and it is where the money goes.
The two problems
Reputational adjacency. Your advertisement next to content that damages you. Serious when it happens, and largely addressed by category exclusions and inclusion lists.
Placement quality. Impressions served where no person plausibly saw them: below the fold and never scrolled to, on pages that auto-refresh every few seconds, in apps bought for volume, or to non-human traffic. No reputational damage — just spend converted into a number on a report.
The second is far more common, far less discussed, and consumes a great deal more budget.
Viewability, and what the standard actually requires
The industry's baseline definition of a viewable display impression is modest: a portion of the advertisement in the viewport for a short, continuous period. It means the advertisement had an opportunity to be seen — not that anyone looked.
A campaign can therefore have entirely respectable viewability and still be mostly unseen. Viewability is a floor, not a measure of attention, and buying purely to raise the figure optimises for the wrong thing.
What we actually do
- Pull the placement report weekly, at property level. Not the channel summary. The properties.
- Exclude aggressively and keep excluding. New inventory appears constantly; a list built once and never revisited stops working within a quarter.
- Build inclusion lists where the category warrants it. Regulated and reputation-sensitive categories are better served by a known list than by an ever-growing exclusion list.
- Buy on viewable impressions where the platform supports it. A higher stated unit price for inventory that was actually in a viewport is usually cheaper per outcome.
- Check the anomalies. A property delivering enormous volume with near-zero engagement is worth a look before it is worth a payment.
The exercise that changes most clients' minds: sort the placement report by spend and read the top fifty properties out loud. It takes ten minutes. Almost every inherited account contains names in that list that nobody would have chosen deliberately — and it is the cheapest audit in media buying.
Where automated tools help and where they do not
Verification vendors detect invalid traffic, measure viewability and block categories, and they are worth their fee above a certain spend. What they cannot do is decide what is appropriate for your brand — a category block cannot distinguish serious journalism about a difficult subject from content that trades on it, and blunt keyword blocking routinely defunds exactly the quality journalism a brand would want to be seen in.
That judgement is a buying decision. It is the part we do rather than automate.
What we report
The placement report itself, not a summary of it — top properties by spend, what was excluded this period and why, viewability, and invalid traffic where it is measured.
Clients are frequently surprised the first time. That surprise is the point: a report you cannot be surprised by is not telling you anything.
Last reviewed: 11 September 2026