Supply-Side Platforms · Header Bidding, Yield & Publisher Revenue

An SSP connects a publisher's ad inventory to multiple programmatic demand sources — DSPs, ad networks, and direct advertisers — simultaneously, running auctions to find the highest-paying buyer for each impression. Before SSPs, publishers using programmatic had to choose between selling through a single ad network (which had no incentive to maximise price) or managing multiple direct integrations (which was technically complex). SSPs solved both problems by providing a single integration point that aggregated demand from multiple sources and used competition between buyers to maximise revenue.

Core SSP functions: connecting to multiple demand sources (DSPs, exchanges, ad networks, and direct advertisers); managing the publisher's ad stack (deciding whether to serve a programmatic ad or a direct-sold campaign for each impression); setting and enforcing floor prices (minimum CPMs below which impressions will not be sold); providing real-time reporting on fill rates, eCPM, and revenue; and managing brand safety preferences (blocking specific advertisers or content categories).

Header bidding is a programmatic technique where publishers offer their inventory to multiple demand sources simultaneously before making an ad call to their primary ad server. The name comes from the original implementation method: JavaScript code placed in the HTML page header fires bid requests to multiple SSPs and exchanges in parallel before the page load completes.

The header bidding sequence: when a user loads a publisher's page, the header bidding wrapper (typically Prebid.js, the open-source standard) fires simultaneous bid requests to all configured SSP demand partners. Each SSP has a defined timeout (typically 300–500ms) to respond with a bid. All bids received within the timeout window are passed to the publisher's primary ad server (usually Google Ad Manager). The ad server then competes these bids against its own direct-sold inventory and its own programmatic demand (Google AdX). The highest-paying ad — whether from header bidding or the ad server's own demand — wins and is served.

The critical advantage of header bidding over the previous waterfall model: every buyer sees every impression and can compete on equal terms, rather than buyers being offered impressions sequentially in a priority order where lower-priority buyers only see impressions that higher-priority buyers have passed on. This competition drives higher CPMs for publishers.

A floor price is the minimum CPM below which a publisher will not sell an impression. Floor prices prevent publishers from selling their inventory at CPMs so low they would be better served by house ads or leaving the slot empty. Floor price strategy is one of the most complex yield management decisions in programmatic — set floors too high and fill rates drop, leaving revenue on the table; set floors too low and inventory is sold below its market value.

Dynamic floor pricing — where the SSP automatically adjusts floor prices based on historical bid data, time of day, content category, and audience quality — has replaced static floor prices for most sophisticated publishers. SSPs including Google Ad Manager, Magnite, and PubMatic offer dynamic floor tools that aim to capture the maximum clearing price for each impression without sacrificing fill rate.

From the advertiser's perspective, floor prices affect win rates and effective CPMs. An advertiser bidding £2 CPM may win many impressions on one publisher (whose floor is £0.50) and none on another (whose floor is £3). DSPs that provide insight into floor prices by site and deal type enable more accurate bid optimisation.

As third-party cookies phase out, publisher first-party data — logged-in user data, subscription data, declared interest data — is becoming the most valuable targeting asset in programmatic. Publishers who can offer verified, consent-based audience segments to advertisers command significant CPM premiums over contextual-only inventory. SSPs are developing data clean room and private marketplace tools that allow publishers to monetise their first-party data while maintaining user privacy.

From an advertiser perspective, publisher first-party data available through private marketplace deals is often the highest-quality programmatic targeting available — because it is based on actual user behaviour and declared intent on that publisher's properties, rather than modelled third-party segments with documented accuracy limitations.

Key SSP metrics publishers monitor: effective CPM (eCPM) — the average CPM across all demand sources and fill rates; fill rate — percentage of available impressions that were sold; bid density — number of competing bids per impression (higher bid density means better price competition); and win rate by demand partner — which DSPs and exchanges are winning the most impressions. These metrics guide decisions about which SSP partners to prioritise, where to set floor prices, and which content categories are most valuable to programmatic buyers.

Advertisers interact with SSPs indirectly — through their DSP — but SSP choices affect campaign outcomes in important ways. Supply path optimisation (described in the DSP guide) involves selecting which SSP paths to prefer for each publisher. Publishers who use high-quality SSPs with low fraud rates, good viewability measurement, and transparent reporting provide better campaign outcomes than publishers using low-quality supply chains. Checking the ads.txt file of publisher sites before adding them to include lists is the basic due diligence step for supply path quality.

As third-party cookies phase out, SSPs are investing in identity solutions that help publishers monetise their inventory without cookie-based audience targeting. Publishers with large registered user bases — media companies with newsletter subscribers, streaming platforms with logged-in viewers, e-commerce sites with account holders — have first-party identity data that can be activated for audience targeting even in a cookieless environment.

SSPs are building data clean room infrastructure that allows publishers to share their first-party audience data with advertisers without exposing individual user data. A publisher can create audience segments from their logged-in user data (e.g., "users who read about personal finance three or more times per month") and make these segments available to advertisers through the SSP's private marketplace — commanding a significant CPM premium over contextual-only inventory.

Unified ID 2.0 (UID2) — the industry's primary email-based identity solution — is being integrated by most major SSPs. When a user authenticates (logs in or provides an email address) on a publisher's site, the SSP can convert that email to a UID2 token and include it in bid requests. DSPs with UID2 integration can then match that token to their own CRM data, enabling people-based targeting without third-party cookies.

The SSP's role in video and connected TV supply is more complex than in display because the technical standards, deal structures, and measurement frameworks are different. Video inventory is typically valued at 5–10× display CPMs, making the economics of SSP fee structures more significant — a 15% SSP fee on a £20 video CPM is £3/thousand impressions, compared to £0.15 on a £1 display CPM.

For CTV specifically, the SSP acts as the programmatic monetisation layer for broadcaster streaming apps and AVOD platforms. Magnite (through its SpotX acquisition) is the dominant CTV-focused SSP, with direct integrations across major streaming platforms. The CTV SSP provides: programmatic deal management; audience data activation through ACR (Automatic Content Recognition) data integration; and measurement integration with third-party brand lift and cross-device attribution vendors.

Publisher-direct CTV deals are often conducted outside the SSP entirely — large streaming platforms (Netflix, Disney+, Amazon Prime Video) sell their advertising directly or through managed service arrangements rather than through programmatic SSP infrastructure. The SSP is more relevant for mid-tier publisher streaming content and AVOD platforms where programmatic automation provides efficiency that direct sales teams cannot achieve at the same scale.

Publishers evaluating SSPs should assess: demand quality (which SSP has the strongest DSP buyer relationships and generates the most competitive bidding on their specific inventory type?); technology capability (does the SSP support header bidding, first-party data activation, server-side connections, and advanced deal types?); transparency (does the SSP provide clear reporting on win rates, CPMs, and buyer identity?); and fee structure (how does the SSP's take rate compare to alternatives, and are there minimum guarantees or revenue share options?).

Most large publishers use multiple SSPs simultaneously through header bidding — the question is not which single SSP to use but which combination of SSPs maximises competition for their inventory. The incremental revenue from adding a new SSP must justify the integration and management overhead. The diminishing returns curve: the first 3–4 SSPs in a header bidding stack provide the most incremental demand; beyond that, additional SSPs add latency without proportional revenue uplift.

Sources & Further Reading