TL;DR: In auction buying, CPMs are set by competition: the price you pay is determined by what others are willing to pay for the same impression. In negotiated buying, CPMs are set by agreement: the price reflects what both parties accept as fair given the inventory and the mandate. The CPM patterns that result are different in shape, predictability, and relationship to delivery quality. Buyers running both models need separate evaluation frameworks for each.

Programmatic buyers have spent fifteen years learning to interpret CPM patterns in auction environments. They know what efficient CPMs look like for different inventory tiers, how seasonal competition affects clearing prices, and how floor price changes by publishers flow through to their cost reports. This expertise does not transfer directly to negotiated buying, because the dynamics that produce CPMs in a bilateral negotiation are different from the dynamics that produce them in an auction.

Understanding those differences is not academic: it affects how buyers set mandate parameters, evaluate deal efficiency, and compare bilateral spend to open exchange alternatives.

How CPMs are set in auction buying

In a real-time bidding auction, the clearing price for an impression is determined by the bids of competing buyers. In a second-price auction, the winner pays the second-highest bid plus one cent; in a first-price auction, the winner pays their bid. In either format, the price is a function of demand: when many buyers want the same impression, prices rise. When demand falls, prices fall.

The buyer's CPM in an auction environment is an outcome, not a commitment. The buyer sets a bid or a target CPM, but the actual clearing price is determined by the market at the moment of the auction. A buyer cannot predict the exact CPM they will pay for a given impression opportunity; they can only set parameters that define the range they are willing to pay.

The implication is that auction CPMs are volatile in a way that reflects demand-side dynamics the buyer does not control. A competitor increasing their budget, a publisher adjusting their floor, or a change in auction dynamics on the exchange can all produce CPM movements that have nothing to do with the buyer's campaign performance.

How CPMs are set in bilateral negotiation

In a bilateral negotiation, the CPM is agreed between the buy-side agent and the sell-side agent. The buy-side agent proposes a CPM within its mandate range; the sell-side agent evaluates the proposal against the publisher's floor and inventory mandate; and the two agents negotiate until they reach an agreed price or determine that no agreement is possible.

The agreed CPM is a commitment, not an outcome. When the DealSheet is executed, both parties have agreed to the stated CPM. The buyer will not pay more than the agreed rate for the inventory covered by the deal; the publisher will not offer the inventory at a lower rate to another buyer while the deal is active.

The implication is that negotiated CPMs are stable in a way that auction CPMs are not. A buyer who has agreed a CPM of £5.50 for a defined inventory block over a four-week period knows the cost structure of that deal at the point of execution. The CPM will not change because a competitor increased their budget.

What the CPM pattern looks like in practice

Buyers comparing their bilateral CPMs to their open exchange CPMs often find that the bilateral CPMs appear higher on a simple cost-per-impression basis. This comparison is misleading if it does not account for the differences in inventory quality, delivery guarantee, and reconciliation accuracy.

In open exchange, the lowest CPM impressions are typically the lowest quality: high MFA exposure, low viewability, poor contextual relevance. The average CPM for a well-optimised open exchange campaign reflects a mix of quality tiers that the buyer cannot always disaggregate precisely.

In bilateral negotiation, the inventory is specified in the DealSheet. The buyer knows exactly what inventory they are getting. The CPM reflects the agreed value of that specific inventory, not an average across quality tiers. A bilateral CPM that appears higher than the open exchange average may represent better inventory at a premium that is lower than the quality premium would imply.

The more useful comparison is cost per verified impression against brand safety and viewability standards, not raw CPM. Buyers who have run this comparison consistently report that the apparent premium in bilateral CPMs reduces or disappears when quality-adjusted.

What this means for mandate parameter setting

Because CPM dynamics in bilateral negotiation are different from auction dynamics, the CPM range in a media buying mandate should not simply replicate the CPM targets from an equivalent open exchange campaign.

The mandate CPM range should reflect the buyer's assessment of fair value for the specific inventory tier being negotiated, not the competitive clearing price for comparable inventory in open exchange. For premium publisher inventory with strong viewability and brand safety performance, the mandate CPM range should reflect that premium. For mid-tier inventory on defined domains, it should reflect the negotiated value of guaranteed delivery rather than the competitive cost of winning the same impressions in an auction.

Mandates with CPM ranges set from auction data without adjustment tend to produce agents that decline a disproportionate number of proposals because the sell-side mandate floors are set above the buyer's auction-derived targets. The adjustment is not large, but it requires buyers to think about CPM as a negotiation anchor rather than as an auction ceiling.