The Transparent Bidding Ad Exchange

programmaticLast week’s PROGRAMMATIC I/O delivered many interesting discussions, and one of them revolved around transparency in bidding on an exchange.

What if you had sufficient transparency on every impression’s auction, such that all buy-side participants bidding in an auction (even bidding “lurkers” or non-bidders) were revealed? Who would care about this data and why?

Hold on tight!

Hypothetical use case

A user (more specifically, a web browser) visits a popular shoe site and views a pair of shoes. They decide not to buy and move on to a local news site’s home page, where a 728×90 display ad across the top of the page is hooked up to an auction in a display advertising exchange such as Right Media, AppNexus or DoubleClick AdExchange.

A retargeting firm hired by the shoe company bids on that placement, and so do a number of other retargeting companies, since the user has visited shoe sites for the past couple of weeks. The bids arrive varied for many reasons:

  • Different price points of shoes – higher-priced shoes mean higher Cost Per Actions (CPAs).
  • Recency of a user’s visit – this week’s visitors could be more valuable than last week’s.
  • Some marketers’ first-party data suggest that the user has a history of buying shoes, so the marketers’ demand-side platform may be better informed to bid higher or lower.
  • The user visited car sites too – hey, there are car company retargeters bidding!
  • Some bidders lack sophistication and do not bid dynamically, depending on the user cookie.
  • Some bids never make the auction due to price floors set by sellers.
  • And on and on…

So if I’m a marketer and I see all these bids from millions of auctions, this data could be very valuable not only to inform my bidding but to provide a unique business intelligence (BI) opportunity on my competitive set.

On the other hand, in a fully transparent auction where everyone knows who is behind every bid (not just the DSP but the actual advertiser), presumably if you’re the marketer you’re giving yourself up too. There’s a cost.  Why would you do it?

This type of fully transparent auction would need to be incentivized somehow, assuming that the business intelligence outweighs the strategy you give up through bidding data. Is that possible?

Publishers get control

And that’s where the publisher could come in. Maybe it’s not about BI for the marketer at all. Could a publisher use this new buy-side auction data too, such that it receives better CPMs in an auction where the buy side is fully transparent? This could incent the publisher to deliver better inventory (dependencies include placement, recency in the user session, daypart, etc.).

Just like in the direct buys world, the ad exchange auction reveals potential relationships, leads that could turn the publisher ad trader into a lead generator for the direct sales team.

The continuing challenge in the ad exchange auction is to lure more so-called “premium” inventory. Providing transparency may be one step, but the intelligence layer on top, bringing leads to the publisher, could seal it. Perhaps this is what creates the programmatic “direct” or forward market sales of the future, too. The publisher says, “Let me see who’s bidding and I’ll give you the good stuff, including my guaranteed inventory.” The marketer might like that.

Enjoying this content?

Sign up to be an AdExchanger Member today and get unlimited access to articles like this, plus proprietary data and research, conference discounts, on-demand access to event content, and more!

Join Today!


  1. Buyside Guys

    Umm…Since the advent of Buyer ID, hasn’t this already happening extensively for almost all major ATD/RTG buyers in partnership with any “PMP” publisher who work with the major SSPs, PubMatic and Rubicon?

    • Andrew Casale

      BuyerID gives the seller insight into who is bidding on their supply. While this is a very useful parameter, John’s post, and the debate on the panel, centered around going a step further to make transactional bidding data more available. As in, rather than the NYTimes knowing Chrysler’s bid was from Cadreon which is current state today, also being able to enlighten Ford to the fact that they were outbid and Chrysler edged them out, to incentivize a more competitive market dynamic.

      The winning buyer pays but a penny more regardless of their WBP, so you may find your most valuable audience segments being bought by a competitor because your WBP averages are lower than the market. To give you an example.

      Marketer A and B are both vying for the same in market segments, be it travel, auto, or anything. Marketer A employs a hyper conservative bidding strategy, opting to place bids at the lowest bid price they can that still wins them impressions, assuming this is the best way to bid. Marketer B opts to bid with a healthier average bid price to maximize the supply of the segments they want. Marketer B will effectively always win the impression/audience when it overlaps with Marketer A in the same auction, yet will only ever have to pay the low price that Marketer A has effectively “set” – therefore Marketer A is giving their competitor a huge discount as well as first right of refusal over any audience they find valuable, but has no idea this is occurring.

      Price discovery is an incredibly important market quality that does not exist in today’s ad exchange dynamics.

      • Varoujan Bedirian

        Andrew, in your example, outside of fear of exposing its pricing, there are multiple and valid reasons why Marketer A would bid low. E.g. has enough supply of impressions of its target audience from other Sellers, isn’t able to make back-end eCPC/eCPA work and thus needs to bid low for economics to work, etc.

        But one reason we take for granted is Marketer A’s trust of Seller’s 2nd price auction closing algorithm.
        What if Marketer A is suspicious that Seller wouldn’t close on penny-above-2nd-price, and believes that if it bids high, it will pay proportionately higher – akin to a percentage of the Marketer’s own 1st price?

        This is especially true if Seller is clever enough to mine historical bids and trends from each Marketer for each end user (potentially factoring end user timezone-based day-part, day-of-week, and seasonality), to estimate an expected value for this impression to this Marketer, and to use this estimate as a soft floor alongside the hard floors put by the actual Publisher (could be different from Seller itself).
        This kind of bid landscape manipulation by Seller, would push an algorithm-savvy Marketer to try to bid in a manner to obfuscate as much as possible its true intentions, adding noise to the signal of bids, and muddying the clarity of bid transparency being discussed.

        Thoughts on this?

        And, as an aside, is ‘Buyer Id’ in RTB being understood as a) the end marketer, or b) the agency trading desk intermediary? Pretty sure it’s not c) the DSP tech platform that ATD is using. Curious.

      • Andrew Casale


        Great points.

        Transparency works in a multitude of ways, all of which make the marketplace work better. To the scenario you described, transparency corrects the underlying issue – if the sell side’s transactional data is transparent the buy side will gain a new found trust in how prices are established. The practices that buyers try to dodge by modifying bids would no longer be as relevant if the sell side was unable to mask price practices. You would know how each price is set, whether it be via a true 2nd bid, or a floor. And if it was set by a floor, you’d know pretty quick whether the floors are set statically, or dynamically. The trouble in today’s opaque market is not every floor is set by using your bid prices against you, yet all bidders bid with that as the assumptive model, which harms those selling in the market that are trying to play by the rules.

        BuyerID is commonly understood as the agency trading desk. In the absence of a desk involved in the buy, it would be the end marketer, or media agency as we have started to see conducting the buy directly on the client’s behalf. In other cases it can also mean the entity bringing the buy to auction – a network or publisher doing audience extension for instance.



  2. Publisher can also take advantage of this transparency to change the rule of how a winning bid is determined – not just by the bid amount alone, but also adding in “quality of the bid price”! This will have far reaching consequence than bid price. If you find this topic interesting, you can read my blogpost on “bid quality score” for details.

  3. essentially we are saying if more price transparency existed, it would drive not only higher rates for in demand impressions, but the rate would reflect what the market is willing to bear. Today the market pays well below what it is willing to pay, this devalues the whole ecosystem, not just publisher inventory. Increase the overall economy and we all win.
    Personally, I want higher quality inventory (no, I will not define here), and I am wiling to pay 10x what I am paying today. I don’t see higher quality inventory and 2nd price auction pricing’s lack of transparency is part of the problem.