How we evaluate inventory quality
Every exchange says it values quality. Far fewer can tell you what they actually check before a publisher goes live, or what gets a publisher removed afterward. We think that omission is the whole problem, so this is the bar oxavane holds inventory to — concretely, and the same way for everyone. Curation is not a marketing posture here. It is the work of deciding which screens belong on the exchange and which do not, and then enforcing that decision continuously.
The one thing we are buying: a real screen with a real audience
Strip everything else away and inventory quality reduces to a single claim that has to be true: a real person had a real opportunity to see this, on a real screen, in the context described. Most fraud and most low-quality supply is a violation of one clause in that sentence — the person is not real, the screen is not real, the opportunity did not exist, or the context was misrepresented. Our entire process is built to verify each clause before we connect a publisher and to keep verifying it after.
What we check before a publisher goes live
Real screens, real placements
We confirm that the inventory corresponds to physical or connected screens that genuinely exist and genuinely run the placement described. For in-venue and connected surfaces that means understanding where the screen is, what it is, and how an ad renders on it. A loop that never actually displays, a "screen" that is a server invention, or a placement buried where no audience can see it does not pass.
Accurate declarations
Everything a publisher tells us has to match what we can observe. Declared dimensions, formats,
positions, supported creative, environment, and the owner of the inventory all have to reconcile
with reality and with the publisher's own transparency files. A bid request that claims one thing
while the surface is another is the precise mechanism of spoofing, and it is disqualifying. This
is also where supply-chain transparency is non-negotiable: a publisher that will not declare its
authorized sellers and owner domain in ads.txt and app-ads.txt does not
get connected, full stop.
Correct IAB taxonomy mapping
Every surface and audience is mapped to the IAB Tech Lab taxonomy, and the mapping has to be honest. Content category, context, and audience signals are what let a buyer decide a placement is suitable and brand-safe before bidding. Mislabeling a context to make it look more premium or more targetable than it is corrupts that decision for every buyer downstream, so we treat taxonomy accuracy as a quality gate, not a formality.
Audience validation
We validate that the audience represented for a surface is the audience actually present at it. Claims about who sees a screen — how many, in what context — have to stand up to scrutiny rather than rest on an operator's say-so. Impressions we cannot validate, we do not sell as validated.
What gets inventory rejected or removed
The same standards run continuously, not just at onboarding. The categories of abuse we screen for and remove are specific:
- Incentivized exposure dressed up as organic attention — impressions a viewer was paid, rewarded, or coerced into generating, sold as if they were genuine interest.
- Made-for-advertising (MFA) surfaces — properties that exist only to manufacture impressions, with no real content, audience, or reason to be there beyond serving ads.
- Auto-refresh abuse — placements that churn impressions on an aggressive timer to inflate counts, so that the "opportunity to see" is fictional even when the screen is real.
- Misrepresented context or audience — any gap between what is declared and what is delivered, whether in taxonomy, placement, or who is actually watching.
- Undeclared or unauthorized supply paths — inventory whose seller or intermediary cannot be reconciled against the relevant transparency files.
A surface that crosses these lines after going live is removed, not quietly tolerated because it adds volume. Scale that is built on impressions a buyer would refuse if they could see them is not scale we want.
Why curation protects honest publishers most
It is easy to read a quality bar as a buyer-side protection, and it is one — buyers get supply whose claims hold up, with fewer dollars lost to fraud and misrepresentation, which is exactly what raises their confidence and their bids. But the publishers who benefit most are the honest ones. When an exchange lets incentivized, made-for-advertising, and auto-refresh inventory in, that supply does not sit in a corner; it drags down the price and trust of everything around it, and the operator running real screens with real audiences pays for someone else's abuse.
Curation reverses that. By keeping the bad supply out, we protect the rate and reputation of the good supply, so a publisher doing the right thing is not punished for the marketplace's tolerance of those who are not. A smaller, cleaner pool that buyers trust is worth more — to every honest participant in it — than a larger one they have to discount for risk. That is the whole bet of an exchange that curates, and it is the one we are making.
Operate real screens and want them evaluated for the exchange? Write to publishers@oxavane.com — a human replies within one business day.