Prebid docs - browse
Prebid/Settings/Prebid send all bids
SETTING

Prebid send all bids

Send-all-bids controls whether Prebid gives the ad server targeting for every bidder, or only for the bid it ranked first.

Plain English

enableSendAllBids defaults to true. Prebid's recommendation is to leave it on when the ad server can take the volume, because the ad server then makes the decision and can report every bidder's price. The keys look like hb_pb_bidder plus a copy of the winner on hb_pb with no suffix.

The cost is size. Each bidder adds a set of key-values, and GAM will only accept so much on a request. Prebid lists the ways to cut that down: turn enableSendAllBids off, cap auctionKeyMaxChars, or send only the top N bids. Turning it off without targetingControls.alwaysIncludeDeals can drop a deal that should have beaten the open auction.

Targeting has to be configured before setTargetingForGPTAsync or getAdserverTargeting runs. Changing the flag later does not rewrite keys already sent.

How it shows up in pbjs

This is Prebid's example for sending the top bid while still sending deals. It is the off switch, not the default. Leaving the flag unset means every bid is sent.

pbjs.setConfig({
  enableSendAllBids: false,
  targetingControls: {
    alwaysIncludeDeals: true
  }
});

When it helps, when it hurts

Helps

Sending every bid lets the ad server prefer a deal, a priority, or a bidder your line items treat differently from raw CPM. Historical price reporting in GAM needs those keys. If a wrapper turned the flag off to save bytes, turning it back on restores that.

Hurts

Too many key-values and the ad server drops the request or truncates targeting. Sending only the top bid hides losing prices from reporting and can suppress a deal. The right cap depends on how many bidders you actually call, not on a sample config.

What bidtune tests here

Autopilot can set enableSendAllBids to true on optimized traffic, which is Prebid's own default. The arm matters when the live setup has turned the flag off. bidtune does not force the flag off, and it does not set alwaysIncludeDeals for you.
If the page already sends every bid, this test changes nothing. Check pbjs.getAdserverTargeting() before treating it as a live difference. Debugging Prebid shows that call.
The 10% holdback keeps your current configuration. A test runs on 30% of optimized traffic, for at least 72 hours and 2,000 sessions in each arm, and no longer than 14 days. bidtune keeps the change only after the capped session read is better at two daily checks.
How bidtune tests itTHE RULES
A test takes 30% of optimized traffic
Judged against its own control. The 10% holdback never changes.
10% holdback60% control30% test
sessions per arm2,000minimum runtime72hdecides onrev / session