Real-Time Bidding (RTB) is a programmatic transaction model where a single digital ad impression is auctioned and sold in the milliseconds between a user clicking a link and the webpage loading. A Supply-Side Platform (SSP) sends a bid request containing anonymized user data to multiple Demand-Side Platforms (DSPs), which algorithmically evaluate the impression's value and submit a bid on behalf of advertisers.
Glossary
Real-Time Bidding (RTB)

What is Real-Time Bidding (RTB)?
Real-Time Bidding is the automated, per-impression auction mechanism that powers programmatic advertising, enabling instantaneous ad placement based on user data.
The highest bidder wins the auction, and their ad creative is instantly served to the user. This entire process, governed by the OpenRTB protocol, relies on low-latency infrastructure and predictive click-through rate (CTR) models to value each impression, making it the foundational economic engine of modern digital advertising.
Core Characteristics of Real-Time Bidding
Real-Time Bidding (RTB) is defined by a set of distinct technical characteristics that enable the automated trading of ad impressions within milliseconds. These core components govern how bids are requested, evaluated, and resolved before a webpage renders.
Per-Impression Auction
Unlike traditional bulk ad buying, RTB operates on a per-impression basis. Each single ad slot on a webpage triggers an independent auction. This granularity allows advertisers to value every user uniquely based on their data profile, ensuring they only pay for the specific individuals they want to reach rather than broad, undefined audiences.
Sub-100ms Latency Window
The entire RTB transaction must complete in under 100 milliseconds from the ad request to ad serving. This includes:
- User data passing to a Supply-Side Platform (SSP)
- Bid request broadcast to multiple Demand-Side Platforms (DSPs)
- DSP evaluation against advertiser criteria
- Bid response and auction winner selection
- Ad creative retrieval and rendering Any delay beyond this window results in a default or blank ad, losing revenue.
Second-Price Auction Mechanics
Most RTB exchanges use a second-price auction model. The highest bidder wins but pays $0.01 more than the second-highest bid. This encourages truthful bidding, as the winner pays the market-clearing price rather than their maximum willingness to pay. Some exchanges have shifted to first-price auctions, where the winner pays exactly what they bid, requiring bid-shading algorithms to avoid overpayment.
OpenRTB Protocol Standardization
Communication between SSPs and DSPs is standardized by the OpenRTB protocol, maintained by the IAB Tech Lab. This specification defines the JSON-based request and response objects, including:
- BidRequest: Contains site, user, device, and impression details
- BidResponse: Returns bid price, ad markup, and tracking URLs This standardization enables interoperability across hundreds of ad tech vendors without custom integrations.
Real-Time User Valuation
DSPs evaluate each bid request by applying proprietary CTR prediction models and conversion rate models to calculate an expected value per mille (eCPM). The bid price is derived from:
- Predicted probability of a click or conversion
- Advertiser's target cost-per-action (CPA) or return on ad spend (ROAS)
- User segmentation data (demographics, browsing history, purchase intent) This valuation must occur in single-digit milliseconds to stay within the total latency budget.
Cookie Sync and Identity Matching
Before an auction can occur, the SSP and DSP must map their respective user identifiers through a cookie sync process. When a user visits a publisher site, the SSP redirects to the DSP's sync URL, passing its user ID. The DSP maps this to its own ID and stores the match table. This identity resolution is critical for applying audience data and frequency capping during the bid evaluation.
Frequently Asked Questions
Clear, technical answers to the most common questions about the programmatic auction mechanics, protocols, and strategies that power real-time bidding.
Real-Time Bidding (RTB) is a server-to-server programmatic auction protocol where a single digital ad impression is bought and sold in the milliseconds between a user loading a webpage and the ad rendering. When a user visits a publisher's site, a Supply-Side Platform (SSP) sends a bid request containing anonymized user data, page context, and minimum floor price to multiple Demand-Side Platforms (DSPs). Each DSP, acting on behalf of an advertiser, evaluates the impression against its targeting criteria and Click-Through Rate (CTR) prediction models to calculate a bid in under 100ms. The SSP collects all bids, declares the highest bidder the winner, and returns the winning ad markup to the browser. The entire process, governed by the OpenRTB protocol, completes before the page finishes loading, ensuring a seamless user experience.
Enabling Efficiency, Speed & Accuracy
Intelligent Analysis, Decision & Execution
We build AI systems for teams that need search across company data, workflow automation across tools, or AI features inside products and internal software.
Talk to Us
Search across company data
Give teams answers from docs, tickets, runbooks, and product data with sources and permissions.
Useful when people spend too long searching or get different answers from different systems.

Automate internal workflows
Use AI to route work, draft outputs, trigger actions, and keep approvals and logs in place.
Useful when repetitive work moves across multiple tools and teams.

Add AI to products and internal tools
Build assistants, guided actions, or decision support into the software your team or customers already use.
Useful when AI needs to be part of the product, not a separate tool.
Related Terms
Explore the core mechanisms and infrastructure that power programmatic advertising auctions, from bid requests to ad serving.
Bid Request
A structured data packet broadcast by a Supply-Side Platform (SSP) to multiple Demand-Side Platforms (DSPs) when an ad impression becomes available. It contains critical contextual signals for valuation:
- User Data: Cookie ID, device type, IP-derived geolocation, browser language.
- Context Data: Page URL, content category, IAB taxonomy classification.
- Inventory Data: Ad slot position, accepted creative sizes, minimum bid floor.
- Privacy Signals: GDPR consent string, CCPA opt-out status. The DSP must parse this payload and return a bid response within a strict timeout window, typically 100-150 milliseconds.
Supply-Side Platform (SSP)
A technology platform used by digital publishers to manage, sell, and optimize their available ad inventory. The SSP acts as a yield-maximization engine:
- Header Bidding Wrapper: Orchestrates a pre-auction among multiple demand sources before calling the primary ad server.
- Floor Price Optimization: Dynamically adjusts the minimum acceptable bid based on historical demand patterns and current fill rates.
- Channel Management: Segments inventory into direct deals, private marketplaces (PMPs), and the open exchange.
- Ad Quality Screening: Filters malicious creatives and enforces competitive separation rules before rendering.
Ad Exchange
A digital marketplace that facilitates the real-time transactional connection between DSPs and SSPs. It operates as a neutral auctioneer:
- Auction Mechanics: Typically runs a second-price auction (the winner pays $0.01 more than the second-highest bid), though many have shifted to first-price auctions.
- Traffic Shaping: Filters low-quality bid requests to reduce infrastructure load on DSPs.
- Fee Structure: Extracts a transparent "take rate" from the clearing price, usually 5-20%.
- Identity Resolution: Maps various user identifiers (cookies, MAIDs, UID2.0 tokens) to a common namespace for cross-platform targeting.
Win Notice & Ad Serving
The post-auction workflow that finalizes the transaction and delivers the creative. After the exchange declares a winner:
- Win Notice URL: The DSP receives a ping confirming the auction win and the final clearing price.
- Creative CDN Retrieval: The user's browser is redirected to the advertiser's Content Delivery Network (CDN) to fetch the HTML5 or video creative asset.
- Client-Side Rendering: The creative executes in-browser, firing an impression tracker to confirm viewability.
- Bid Loss Feedback: Some exchanges provide a loss reason code (e.g.,
Deal ID & Private Marketplaces
A unique identifier that facilitates non-open-auction transactions between a specific buyer and seller. This bypasses the fully competitive RTB environment for guaranteed access:
- Preferred Deals: The seller offers inventory at a fixed price before it goes to open auction; the buyer gets a "first look."
- Private Auctions: A closed group of invited buyers compete in a real-time auction for premium inventory.
- Programmatic Guaranteed: A direct reservation buy executed programmatically, guaranteeing a fixed volume of impressions at a fixed CPM.
- Deal ID Transparency: Both parties use the ID to reconcile billing and delivery reports in their respective platforms.

About the author
Prasad Kumkar
CEO & MD, Inference Systems
Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.
His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.
Partnered with leading AI, data, and software stack.
How We Work
Custom AI workflows for your Business
One-fit-all AI don't work for modern businesses. At Inferensys, we aim to understand your business & custom requirements; which we use to define most efficient agentic workflows, the data, and the tools for your business.
01
Review the use case
We understand the task, the users, and where AI can actually help.
Read more02
Pick the right approach
We define what needs search, automation, or product integration.
Read more03
Build the first useful version
We implement the part that proves the value first.
Read more04
Improve from there
We add the checks and visibility needed to keep it useful.
Read moreThe first call is a practical review of your use case and the right next step.
Talk to Us