Floor Pricing Myths: Why Your Price Floors Are Probably Wrong
Price floors are one of the most powerful yield tools available to publishers — and one of the most commonly misconfigured. Static, intuition-based floors are leaving money on the table in both directions. Here's the signal-driven approach to floor pricing that actually works.
The Two Ways Floors Fail
Most discussions of floor pricing focus on one type of failure: floors set too high, causing fill rate to collapse and revenue to fall. But there's an equally costly failure in the opposite direction: floors set too low, clearing at prices that dramatically undervalue the inventory.
Both failures have the same root cause: floors set without adequate data. A floor that doesn't reflect the true market value of a specific impression on a specific placement at a specific time is either underpriced or overpriced — and either way, the publisher loses.
Why Static Floors Don't Work
Imagine a publisher setting a $2.00 CPM floor on their homepage leaderboard. On a Tuesday afternoon in January, that floor is aggressive — it might block significant fill. On a Thursday evening in Q4, when CPMs for that placement are running $4–6, the floor is leaving $2–4 on every cleared impression.
Static floors assume that impression value is constant. It isn't. Impression value varies by time of day, day of week, audience composition, content context, device, and dozens of other factors. A floor that doesn't reflect this variation will be wrong for most impressions.
Signal-Driven Dynamic Floors
The solution is floors that respond to the signals that determine impression value. A floor pricing system that knows the time, the audience signals, the content context, and the historical clearing price for similar impressions can set an optimal floor for each individual impression — not a single number for an entire ad unit.
This is only possible with rich signal data. A floor pricing algorithm with access to full signal context can predict the clearing price distribution for each impression and set a floor that maximizes expected revenue. Without signals, it's guessing.
Metrux's intelligent routing includes signal-aware floor optimization — using enriched bid request data to guide floor decisions in real time, ensuring every impression is priced to match its actual market value rather than a static estimate.
The Practical Approach
Publishers who want to move to signal-driven floors should start by analyzing their clearing price data by signal dimension: time, device, content category, and placement. This analysis typically reveals that CPM distributions vary by 300–400% across these dimensions. Setting a single floor for a placement that has a 10th percentile CPM of $1.20 and a 90th percentile CPM of $6.40 will always be wrong for most impressions. The data is there — it just needs to inform the floor.
Ready to enrich your bid stream?
Metrux delivers 20–40% yield improvement through signal enrichment — no dev work, no tag tax, no risk.
Request Early Access →