Real-time bidding (RTB) is the process by which digital ad impressions are bought and sold through instantaneous automated auctions. Each time a user loads a webpage or opens an app, an auction takes place in under 100 milliseconds — advertisers submit bids for that specific impression, and the highest bidder's ad is served to the user.
RTB is the transactional engine that powers programmatic advertising. Before RTB, ad inventory was sold in bulk through direct deals — an advertiser would purchase 100,000 impressions on a specific website at a fixed price. RTB replaced this with a per-impression model where each ad view is valued and sold individually based on the specific user, context, and timing.
How real-time bidding works
Bid request generation begins when a user visits a page with ad space. The publisher's ad server sends a bid request to one or more ad exchanges. This request contains information about the available impression — ad size and format, page URL, content category — and anonymized information about the user — device type, geographic location, browsing history signals, and audience segment memberships.
Bid evaluation happens at the demand-side platform (DSP) level. Each connected DSP evaluates the impression opportunity against the campaigns it manages. The DSP considers the advertiser's targeting criteria, budget, bid strategy, and predicted value of the impression to determine whether to bid and how much to offer.
Auction execution follows one of two models. In a second-price auction (historically dominant), the winner pays $0.01 more than the second-highest bid. In a first-price auction (now the industry standard), the winner pays exactly what they bid. The shift to first-price auctions has made bid optimization more important, as overbidding directly increases costs.
Ad serving delivers the winning creative to the user's browser or app. The entire process — from page load to ad display — completes in under 100 milliseconds. The user experiences no perceivable delay.
Post-auction reporting records the outcome for all participants. The winning advertiser's DSP records the impression, the publisher's SSP records the revenue, and tracking pixels begin monitoring for post-impression engagement and conversions.
Why RTB matters
Impression-level valuation is the fundamental advantage. In a bulk purchase, an advertiser pays the same price for every impression regardless of the user viewing it. In RTB, each impression is priced based on its specific value — a high-intent user on a relevant page commands a higher bid than a random user on an unrelated page. This efficiency benefits both advertisers (better targeting) and publishers (higher revenue for valuable impressions).
Audience-first buying replaces site-first buying. With RTB, advertisers buy access to specific users wherever they appear online, rather than buying placement on specific websites. This shift fundamentally changed media planning — the question moved from "which websites should we advertise on?" to "which users should we reach?"
Dynamic pricing ensures the market sets fair prices. Impressions are worth what advertisers are willing to pay, which varies by time of day, day of week, seasonality, user value, and competitive dynamics. RTB's auction mechanism captures this dynamic pricing naturally. Platforms like Soku AI help advertisers navigate RTB auctions across multiple exchanges and platforms, optimizing bids based on real-time market conditions.
Challenges and considerations
Latency constraints limit the complexity of bid decisions. DSPs must evaluate impressions and submit bids in under 10 milliseconds to meet exchange deadlines. This constrains the sophistication of real-time bid models, though advances in edge computing and pre-computed bidding models are expanding what is possible.
Bid shading has become essential in first-price auctions. Without bid shading — algorithmically reducing bids below the maximum willingness-to-pay — advertisers consistently overpay. Sophisticated DSPs use machine learning to predict the minimum winning bid and adjust accordingly.
Supply path optimization (SPO) addresses the problem of the same impression being offered through multiple exchanges and SSPs simultaneously. An advertiser's DSP might see the same impression through three different paths, each with different fees. SPO identifies the most efficient path to each impression, reducing costs and improving transparency.
Ad fraud remains a persistent challenge. Sophisticated bots can generate fake bid requests, mimicking real user behavior to sell nonexistent impressions. While industry standards like ads.txt and sellers.json help verify legitimate supply, fraud detection remains an ongoing arms race.
Privacy-driven signal reduction is changing RTB mechanics. With third-party cookies being deprecated, the user signals available in bid requests are diminishing. This reduces the granularity of per-impression valuation and is driving the industry toward contextual targeting and cohort-based bidding approaches.
