By TGM Team · Sun Oct 18 2026 · 8 min read
..
A header bidding timeout is the maximum time the wrapper waits for bids before it sends what it has to the ad server. Set it too short, and slower demand partners never get to compete. Set it too long, and ads render late enough that readers scroll past them, or leave.
There is no universally correct header bidding timeout. The right value is the one that maximizes total revenue across a reader's session on your inventory, and it can only be found by measuring and testing. This guide explains what the timeout controls, how to read your own data, and how to tune it without guessing.
When a page loads, the wrapper sends bid requests to every configured partner at once and starts a clock. Bids that return before the clock runs out enter the auction. Bids that return afterwards are ignored, even if they were the highest.
That makes the timeout one of the most consequential settings in a programmatic stack. It decides which demand competes on every impression, and it shapes how quickly ads appear. Because it applies to every auction, a small change can have a large cumulative effect on revenue in either direction.
A bid passes through several hops before it reaches the wrapper: the exchange or SSP processes the request, routes it to buyers, the DSP evaluates it and responds, and the answer travels back across the network. Each hop adds processing and network time.
Individually the delays look small; together they can approach or exceed the window the auction allows.
When a bid arrives late, the opportunity does not wait for it. That partner's money is simply not in the auction, and the publisher never sees what it would have paid. Latency is usually discussed as an engineering metric, but in advertising it is a commercial one.
How to Choose the Right Auction Timeout for Different Advertising Environments
1. Number of Demand Partners Responding in Time
- Longer Timeout: Gives demand partners more time to return bids, allowing more partners to participate within the auction window.
- Shorter Timeout: Limits the time available for responses, meaning fewer slower demand partners may submit bids before the auction closes.
2. Bid Competition and Revenue Opportunities
- Longer Timeout: Increases the opportunity to receive additional bids, potentially improving competition and the chance of securing a stronger bid.
- Shorter Timeout: May exclude valuable bids from slower demand partners, although it can help the publisher proceed with ad selection sooner.
3. Ad Rendering and Viewability
- Longer Timeout: Can delay ad rendering while the system waits for bids, potentially affecting viewability if the delay causes users to scroll away or leave the page.
- Shorter Timeout: Can enable faster ad selection and rendering, helping reduce delays that may negatively affect the viewing experience.
4. Page Performance and User Engagement
- Longer Timeout: Can delay content or ad delivery, particularly on poorly optimised pages, potentially affecting user experience and engagement.
- Shorter Timeout: Helps limit waiting time and can support page responsiveness and user engagement. However, its effect on Core Web
Vitals depends on the overall implementation and which performance metrics are affected.
The goal is not the shortest possible auction, nor the fullest. It is the point where the extra bids a longer wait brings in are worth more than the engagement and viewability the wait costs.
Timeouts and lazy loading work together. If an auction starts only when a slot enters the viewport, any wait for bids is time the reader spends looking at an empty space, or scrolling past it. If the auction starts well before the slot is visible, bids have time to return, but requests are wasted on slots readers never reach.
The practical answer is to trigger auctions slightly ahead of the viewport, so a reasonable timeout finishes before the slot becomes visible. That combination protects viewability, which buyers increasingly use to value inventory, without sending requests for every slot on a long page.
Start with each partner's timeout rate: the share of requests where its bid arrived too late. Then read it alongside three other measures.
• Win rate when the partner does respond. A partner that times out often but wins frequently when it bids is costing real money.
• Bid value distribution. A partner that times out and bids low when it does respond may not be worth waiting for.
• Device, connection and geography. Timeout rates are usually higher on mobile networks and for readers far from a partner's servers. A single site-wide figure hides that pattern.
If timeout rates are high across all partners, the problem may not be the timeout at all. Heavy page scripts, slow ad server calls or too many partners can delay the auction before bids are even requested.
1. Record a baseline: timeout rate, bid rate and win rate per partner, plus page RPM, revenue per session and page-speed metrics.
2. Form one hypothesis, for example that a longer timeout on mobile web will capture valuable late bids.
3. Apply the change to a test share of traffic, with a comparable control group.
4. Run the test long enough to cover normal weekday and weekend patterns.
5. Compare net revenue and engagement across both groups, not CPM alone.
6. Roll out only if the gain holds across the full test period, then monitor it.
Change one setting at a time. If timeouts, floors and partner lists all change together, the result cannot be attributed to any of them.
Different environments rarely share an ideal value, so treat them separately.
1. Desktop Web
- What Shapes the Timeout: Generally faster network connections and greater browser processing capacity can provide more flexibility for client-side bidding.
- What Publishers Should Consider: Set a timeout that gives demand partners sufficient time to respond without unnecessarily delaying ad rendering or affecting the user experience.
2. Mobile Web
- What Shapes the Timeout: Variable network speeds, device capabilities, and the performance cost of delayed ad rendering.
- What Publishers Should Consider: Prioritise responsive ad delivery while allowing enough time for valuable demand partners to return bids, particularly on slower connections and lower-powered devices.
3. In-App
- What Shapes the Timeout: The timing of ad requests within the app experience, such as between game levels, alongside SDK processing and server-response latency.
- What Publishers Should Consider: Align the timeout with the available time for ad delivery within the app's flow, accounting for SDK execution, network conditions, and server-side processing.
4. Server-Side and Connected TV (CTV)
- What Shapes the Timeout: The time available before a scheduled or triggered ad break, together with latency between the ad server and participating demand partners.
- What Publishers Should Consider: Ensure the auction finishes within the available ad-break window while giving demand partners sufficient time to respond. Late responses can disrupt ad delivery or leave less time for ad selection and preparation.
In server-side and CTV auctions, the timeout is often set by how much time remains before the ad break must be filled. A late decision there does not mean a slower page; it means an empty or repeated ad slot in front of a viewer.
The timeout is often treated as the only control, but several others reduce the need for a long wait:
• Remove partners that rarely win. Fewer, stronger partners make a shorter timeout safer.
• Move slow partners server-side. This takes work off the device, though server auctions have their own limits.
• Reduce page weight. Scripts that block the main thread delay the auction before it starts.
• Fix duplicate requests. Sending the same impression through several partners that reach the same buyers adds waiting without adding competition.
The timeout is only half the equation. The other half is how quickly each partner responds, and partners can often improve their own response times once they know it matters. Useful questions to ask:
1. Where are your bidding servers located relative to our main audiences?
2. What is your typical response time for our inventory, and how does it vary by region and device?
3. Do you call downstream buyers in sequence or in parallel, and how long do you wait for them?
4. Can you share timeout and error rates for our requests from your side, so we can compare them with ours?
The answers help separate partners that are slow because of distance or architecture, which may be fixable, from partners that simply add little value. A partner willing to share data and adjust its setup usually deserves more patience than one that is not.
• Copying another publisher's timeout instead of testing your own.
• Using one value for desktop, mobile and server-side auctions.
• Raising the timeout to rescue one slow partner, at the expense of every page view.
• Judging a test on CPM alone, when late rendering may have reduced pages per session.
• Never revisiting the setting after partners, page templates or traffic mix change.
TGM works with publishers on Prebid client and server adapters and on real-time revenue optimization across web, CTV and OTT. Its agentic AI tools are designed to analyze performance continuously and surface optimization opportunities, which suits a setting like the timeout that needs regular review rather than a one-time decision.
Timeout tuning is a sensible early test for most publishers, because the data to evaluate it already exists and the change needs no layout redesign. If you are not sure your timeouts are costing you demand, TGM's publisher team can review the numbers with you.
What is a header bidding timeout?
It is the maximum time a header bidding wrapper waits for bids before passing the results to the ad server. Bids that arrive later are ignored.
Is there a standard Prebid timeout value?
No. Wrappers ship with defaults, but the right value depends on your partners, devices and audience, and should be found by testing.
Can a longer timeout lower revenue?
Yes. If late rendering reduces viewability, or readers view fewer pages per session, total revenue can fall even when more bids arrive.
Do timeouts matter in CTV?
Yes. In server-side and SSAI environments, ad decisions must fit within the time before the ad break plays, so slow demand is cut off the same way.
How do I know if a partner's timeouts are worth fixing?
Look at its win rate and bid values when it does respond. High timeouts combined with strong wins point to real lost revenue.
Should every partner get the same timeout?
In most client-side setups, the wrapper applies one timeout to every partner in the auction. That is why removing or moving slow partners is often more effective than raising the timeout for everyone.
Can TGM help optimize auction settings like timeouts?
TGM's revenue optimization and agentic AI tools are built to analyze auction performance and identify optimization opportunities across web, CTV and OTT inventory.
