What is Header Bidding

Other names:

  • tagless
  • advance bidding
  • pre-bid

Problem publisher is facing

Header bidding also helps bypass the traditional waterfall approach. The three major problem of waterfall approach (ie. daisy chain) are:

  • latency
  • discrepancy
  • avg bid price rather than dynamic price.

Header bidding has been gaining popularity in the past couple of years in response to a DFP feature called Dynamic Allocation, which allows Google Ad Exchange – but no other exchanges – to bid against publishers’ direct-sold campaigns. All other exchanges must operate at the end of the waterfall after direct campaigns and Google AdX cherry-pick impressions… of course at Google’s discretion.

How it works:

They rely on a piece of javascript in the publisher's header to work. It is called before the ad server call and normally it integrates with RTB demand partners thru SSP.

header bidding

  • User requests a website
  • Header tag script redirects user to one or many SSPs (or DSPs, or Exchanges)
  • User calls one or many SSPs in parallel
  • SSPs conduct auction with DSPs and internal network demand
  • DSPs respond with bids
  • SSP determines winning bid value and returns to User
  • User passes bid value into ad request and calls Publisher Ad Server (K/V)
  • Ad server determines final line item to serve and redirects User to Marketer Ad Server (let’s assume the ad server determines a pre-bid SSP line item for this example)
  • User calls Marketer Ad Server
  • Marketer Ad Server returns final creative (via CDN)
  • User calls trackback to SSP

With this header bid acts as a floor in AdX and making AdX work harder to win the auction via DFP's dynamic allocation. In general, publishers are seeing increase eCPM by 15-25% while maintaining the same fill rate.

Headache to set it up

  • You now have far more line items need to be created in the ad server. Line item per ad size per bid value. Some publishers will group bid values into small buckets by using wildcards on the bid values. But the most granular the values, the better the yield optimization.

Reference