We love a proper tech debate at ExchangeWire. On this edition of TraderTalk we speak to Neal Richter, Chief Scientist at the Rubicon Project.

Here he discusses the infrastructure required to scale RTB globally, how Rubicon knitted together its recent acquisitions into present offering – and the problems with internal bidding within bidders in terms of how the “internal auction” practice is not giving sellers full liquidity.

Richter also goes into detail on standardisation, particularly OpenRTB, and how DealID works in the bidding process. We will be going into more on these topics in the coming weeks – so stay tuned.



  • Mike

    The whole chain around single platforms negatively impacting revenue because they aggregate bids is a pretty naive argument with a lot of assumptions. If a platform passes along the highest bid that passes quality & channel rules that idea that this has a negative impact on publisher revenue makes no sense.

    The second argument made for “full visibility is required” for an efficient market is idealist and certainly online advertising has repeatedly shown itself to be a market that operates very far from the ideal. In fact, it’s contrary to what both brand buyers and sellers want — they want “programmatic” to be obfuscated/different from premium so that they can maintain the pricing and margins on the direct guaranteed sales. Of course they also want all the deep analytics to then influence premium pricing… but neither side wants to give.

    One thing Richter should be more transparent about is that Rubicon would like to give publishers deep competitive intelligence on what buyers are bidding on and how much.

    I will agree with Richter that we need very clear transparency on pricing mechanisms and margins being made. The new games publishers & SSPs are playing with fake second-price auctions or arbitrary per buyer floor prices is total “bullocks” as the English would say. Similarly if bidding platforms are taking non-transparent margins in the middle that’s of course also flawed.