The Mental Model

If you’re building analytics for programmatic advertising, you’ll quickly drown in metrics. Ad requests, bids, fills, wins, timeouts, revenues from both sides. It’s a lot. The trick is to stop thinking about them as isolated numbers and start seeing them as supply and demand dynamics.

Publishers (SSPs) are the supply side. They have ad inventory to sell. Advertisers (DSPs) are the demand side. They want to buy impressions. Everything else is just measuring how well these two sides are matching up.

Supply side metrics

These tell you what publishers are offering and what they’re getting for it.

Inventory metrics:

  • Total endpoint requests - raw traffic hitting publisher sites
  • Total ad requests - actual inventory being offered for sale
  • Floor price (avg/median) - minimum price publishers will accept
  • Supply revenue - what publishers actually earn

Publishers control volume and set floor prices. Lower floors mean higher fill rates but potentially less revenue per impression. It’s a balancing act.

Quality metrics:

  • Supply chain presence/completeness - transparency in the ad supply chain
  • Invalid traffic detection - catching fraud before it damages inventory value
  • Render rate - ads that actually displayed successfully

Demand side metrics

These show how advertisers are responding to available inventory.

Bidding metrics:

  • Bid requests - opportunities shown to demand side
  • Total bids - how many times demand actually participated
  • Bid price (avg/median) - what advertisers are willing to pay
  • Demand revenue - total advertiser spend
  • Eligibility ratio - percentage of demand that qualifies to bid

Performance metrics:

  • Bids won - successful purchases
  • Win rate - competitiveness of bidding
  • Impressions - delivered ads

Where supply meets demand

This is where it gets interesting. The auction is the intersection point:

Supply (Floor Price) <-- AUCTION --> Demand (Bid Price)

Three key ratios tell you how well the marketplace is functioning:

  • Bid rate - demand participation in available supply
  • Fill rate - supply successfully monetized by demand
  • Win rate - demand success in competitive environment

Economic balance

Revenue metrics show the health of the transaction:

  • eCPM/eRPM - revenue efficiency for both sides
  • Gross profit = demand revenue - supply revenue (platform take)
  • Margin - platform efficiency in matching supply and demand

Friction points

These metrics expose where the marketplace is breaking down:

  • Bid timeouts - technical barriers to matching
  • Bids below floor - demand-supply price misalignment
  • No bids - demand rejecting supply opportunities
  • No-bid rate - overall mismatch rate

Reading market conditions

Once you understand the framework, you can diagnose market state quickly.

Balanced market:

  • High fill rate + high win rate = efficient matching
  • Avg bid price > avg floor price = healthy price discovery
  • Low timeout rate = infrastructure is keeping up

Supply-heavy market:

  • High no-bid rate = excess supply, weak demand
  • Bid prices approaching floor prices = price pressure on publishers

Demand-heavy market:

  • Low fill rate despite high bid rate = supply scarcity
  • High bid prices = competition driving prices up

This framing helps you figure out whether performance issues come from supply-side constraints (inventory quality, floor pricing), demand-side factors (bid competitiveness, eligibility), or marketplace efficiency (technical performance, fraud detection).