Unit
- Real-time bidding: An automated auction system where ad impressions are bought and sold on a per-impression basis in milliseconds. When a user visits a webpage, the publisher's ad server sends a bid request to an ad exchange, which then auctions the impression to multiple demand-side platforms (DSPs) simultaneously.
- Bid request components: Each bid request contains data about the user (such as cookies or device IDs), the page content, the ad slot size, and other contextual signals. DSPs use this information to decide whether to bid and at what price.
- Auction timeline: The entire process from user page load to winning ad display takes under 100 milliseconds. The ad exchange collects bids, determines the winner, and relays the ad markup back to the publisher's page within that window.
- Key participants: The ecosystem includes publishers (sell inventory), ad exchanges (marketplaces), DSPs (buyers), supply-side platforms (SSPs represent publishers), and data management platforms (DMPs provide audience data).
- First-price vs second-price: In a first-price auction, the winner pays their bid price. In a second-price auction, the winner pays one cent above the second-highest bid. The industry has shifted largely to first-price auctions for transparency.
- Programmatic direct: Some impressions are sold via programmatic guaranteed, where a fixed price is agreed in advance, bypassing the auction. This is common for premium reserved inventory.
- Header bidding: An advanced technique where publishers run an auction on their page before calling their ad server, allowing multiple demand sources to bid simultaneously, increasing yield.
- Importance for marketers: Understanding the mechanics helps optimizers control bid prices, frequency, and targeting to achieve campaign KPIs without overspending.
- OpenRTB protocol: The industry standard for real-time bidding communication between ad exchanges and DSPs. It defines the structure of bid requests and responses, including standardized fields for user data, inventory details, and bid parameters. Adherence to OpenRTB ensures interoperability across platforms, enabling seamless integration for buyers and sellers.
- Bid price optimization: DSPs use machine learning models to adjust bids based on historical performance, user behavior, and campaign goals. Factors like click-through rate probability, conversion likelihood, and time of day influence the bid price. Optimizers set floor prices and bid caps to balance cost efficiency with winning rates.
- Frequency capping: A control mechanism that limits the number of times a specific user sees an ad within a given time period. DSPs enforce frequency caps to prevent ad fatigue and reduce wasted spend. Standard caps range from 3 to 5 impressions per user per day, though configurations vary by campaign objective and audience.
- Viewability and fraud detection: Advertisers require that impressions are viewable by humans. Tools like viewability measurement and invalid traffic detection help identify non-viewable or fraudulent impressions. DSPs integrate with third-party verification vendors to filter out low-quality inventory and ensure campaign performance metrics are accurate.
Funnel Stage
AWARENESS
Objective: Introduce the concept and technical flow of real-time bidding to new operations practitioners.
Activities:
- Review a typical bid request schema (JSON) to understand data fields.
- Compare first-price and second-price auction models using simulated scenarios.
- Identify the key latency constraints in the bid-response cycle.
Metrics: Comprehension assessed through quiz on auction mechanics and participant roles.
Quick Revision
- RTB auctions every impression individually in under 100 milliseconds.
- Bid requests include user, page, and slot data that DSPs evaluate.
- First-price auctions dominate, changing bidding strategies.
- Header bidding and programmatic direct are alternative transaction methods.