Real-Time Bidding (RTB)

Definition

Real-Time Bidding (RTB) is a programmatic advertising method where individual ad impressions are bought and sold via an automated auction that occurs in near real time—typically while a webpage or app is loading. RTB is most commonly associated with open exchange buying, where many advertisers can bid on available inventory based on targeting and pricing rules.

From a marketing perspective, RTB enables audience-based buying at impression level, letting marketers optimize bids and targeting dynamically to balance reach, efficiency, and outcomes (such as conversions) across a large set of publishers and placements.

How to calculate (the term)

RTB itself is a mechanism, but campaigns running via RTB are evaluated using auction and delivery economics plus outcome metrics:

• Bid rate = Bids submitted / Bid requests
• Win rate = Impressions won / Bids submitted
• Clear price CPM = (Media cost / Impressions won) * 1000
• Effective CPM (eCPM) = (Total spend / Impressions served) * 1000 (includes fees, if reported that way)
• Fill rate (publisher-side) = Impressions sold / Impressions available
• Average bid CPM = (Sum of bid prices / Number of bids) * 1000 (if bids are tracked)
• Frequency = Total impressions / Unique users reached (as defined by the identity method)

Outcome metrics (common in RTB optimization):

• CTR = Clicks / Impressions
• CVR = Conversions / Clicks (or Conversions / Impressions, depending on definition)
• CPA = Total spend / Conversions
• ROAS = Revenue attributed / Total spend
• Viewability rate = Viewable impressions / Measurable impressions

How to utilize (the term)

RTB is typically used inside a DSP (buyer) and accessed through exchanges and SSPs (seller). Common use cases include:

• Prospecting and scale
  • Rapidly reach relevant audiences across many sites/apps without negotiating individual publisher contracts.
• Retargeting
  • Bid more aggressively for users with recent high-intent behaviors (site visits, product views), within privacy and consent constraints.
• Dynamic bid optimization
  • Adjust bids based on predicted conversion probability, value, device, time of day, geography, and placement quality signals.
• Incremental reach
  • Add reach beyond walled gardens, especially for display and video, with frequency management controls.
• Test-and-learn
  • Quickly test new creative, audiences, and contextual strategies due to fast feedback loops and flexible targeting.

Operationally, RTB works best when you:

• Define bidding goals (CPM efficiency vs CPA/ROAS)
• Apply brand suitability and fraud controls
• Establish frequency and recency rules
• Monitor auction health (win rate, bid rate, pacing)
• Optimize across supply, audiences, and creative

Compare to similar approaches, tactics, etc.

Topic	RTB (Open Auction)	Private Marketplace (PMP)	Programmatic Guaranteed (PG)	Direct IO (Traditional)
Pricing	Auction-based, dynamic	Typically fixed floor or negotiated	Fixed price, reserved	Negotiated
Access	Broad (many buyers)	Invite-only / curated	One-to-one agreement	One-to-one agreement
Delivery certainty	Variable	Higher than open auction	Highest (reserved)	High (reserved)
Transparency & controls	Varies by platform; can be high	Often higher quality signals/controls	Strong controls	Strong controls
Best for	Scaled reach and performance	Quality supply + audience access	Predictable delivery, premium placements	Premium placements and sponsorships

Best practices

• Treat auctions like markets, not vending machines
  • Monitor win rate, floor price pressure, and bid shading (where supported) to avoid overpaying or under-delivering.
• Invest in supply quality controls
  • Use allowlists, blocklists, app bundle controls, and supply-path optimization to reduce low-quality inventory and duplication.
• Use brand suitability standards consistently
  • Set category exclusions, keyword/context filters, and publisher controls aligned to your brand and risk tolerance.
• Harden your fraud and invalid traffic defenses
  • Use pre-bid fraud filters and post-bid monitoring; validate suspicious spikes with log-level checks where possible.
• Separate audience targeting from measurement assumptions
  • Identity is messy; document what “unique reach” and “frequency” mean in your environment.
• Optimize with the right success metric
  • For performance: focus on CPA/ROAS and incrementality methods where feasible. For brand: prioritize reach, frequency, viewability, and attention proxies.
• Keep creative testing continuous
  • RTB performance can decay fast; rotate variants and refresh assets to avoid fatigue.
• Validate attribution and incrementality
  • RTB can look “great” in last-touch attribution; use holdouts, geo tests, or platform incrementality options when available.

Future trends

• Greater reliance on contextual and first-party signals
  • Continued shift away from broad third-party identifiers toward contextual targeting and privacy-safe activation.
• More curated supply via marketplaces
  • Increased use of curated auctions and packaged inventory to balance scale with quality.
• Auction mechanics and price optimization
  • Wider adoption of bid shading and algorithmic floor management to reduce inefficiency.
• Privacy-preserving measurement
  • More modeling, clean rooms, and aggregated reporting approaches to measure outcomes under tighter privacy constraints.
• Cross-channel RTB growth
  • Expansion and maturation of RTB patterns in channels like CTV and audio (with channel-specific constraints and standards).
• Tighter governance expectations
  • More demand for transparency around fees, intermediaries, and “working media” calculations.

Related Terms

• Demand-Side Platform (DSP)
• Supply-Side Platform (SSP)
• Ad Exchange
• Open Auction
• Private Marketplace (PMP)
• Programmatic Guaranteed (PG)
• Header Bidding
• Supply-Path Optimization (SPO)
• Viewability
• Invalid Traffic (IVT) / Ad Fraud