Programmatic display largely abandoned second-price auctions in favour of paying what was bid. The change altered the incentives for every buyer in the market.
The second-price mechanism made honesty easy
Under second-price rules, the winner paid slightly more than the runner-up rather than their own bid. Bidding one's true valuation was therefore the sensible strategy.
Overbidding gained nothing, because the price was determined by someone else's number, and underbidding only risked losing an impression worth having.
That property made the system easy to participate in, which is why it dominated the early years of automated buying.
Header bidding broke the arrangement
As publishers began soliciting bids from multiple sources simultaneously and comparing them, the sequence of nested second-price auctions produced inconsistent outcomes.
Prices could be reduced at more than one stage, and neither buyers nor sellers could reliably reconstruct why a given impression cleared where it did.
Moving to a single first-price auction restored a comprehensible rule at the cost of removing the buyer's protection against overpaying.
Bid shading emerged to fill the gap
Buyers who continued submitting true valuations paid them in full and immediately looked expensive relative to competitors.
The response was shading, where the platform estimates the lowest bid likely to win and submits that instead of the full valuation.
This is a prediction problem rather than a rule, and the quality of the estimate is now a genuine differentiator between buying platforms.
Transparency moved from price to policy
First-price auctions made the clearing price obvious, but they also increased the importance of floor prices, which publishers set and adjust continuously.
A dynamic floor that moves in response to observed bidding can capture much of the value shading was intended to preserve.
The two systems now adjust against each other, which is why clearing prices for identical inventory drift over time without any change in advertiser demand.
Volatility became normal
Because both sides are estimating rather than following a fixed rule, prices for the same inventory move more than they did under the older mechanism.
Buyers see win rates change without having changed anything, driven by adjustments made by sellers or by rival shading models.
Managing a campaign in this environment means watching win rate and clearing price together, since a rising win rate at a rising price is usually a sign of overpaying rather than of success.