Casale Media Index Report: RTB ‘Self Competition’ Is a Myth

Casale MediaThere is a common view that a big chunk of real time bidding activity consists of advertisers competing with themselves to win impression auctions. Not so, according to a new RTB report based on data gathered through Casale Media‘s sell-side Index Platform.

“There are a lot of shared brands, and also a lot of retargeting,” Casale Media VP Strategy Andrew Casale tells AdExchanger. “If you combine those scenarios you get the myth that most of the bid density is someone bidding against themselves. But bid platforms are very unique in the way they bid. It’s not just 12 auto brands bidding, or 12 travel brands. When bid activity is dense, it’s also dense from a category diversity perspective.”

The finding is just one facet of Casale Media’s new Index Quarterly Report (download), which examines Q1 and Q2 bid activity across a number of exchanges in North America. It’s a compelling read for anyone closely following the RTB space.

Here are a few other key findings:

Brand Behavior

Total RTB impressions on the Index Platform increased 28% between the first and second quarters of 2012. More striking, the number of brands and campaigns grew significantly between Q1 and Q2, owing to a combination of factors including cyclical sluggishness in the first quarter and a progressive surge in marketer investment in RTB. (One possible distorting factor is growth in Casale Media’s nascent RTB platform, but Andrew Casale dismisses that aspect: “The demand we see isn’t young. We work with every desk and every DSP. We are seeing the entire fire hose of demand.”)

Among all bidders, national brands owned 57% of RTB impressions in Q2, followed by Internet brands (29%), major regional brands (11%) and local business (3%). These ratios were fairly consistent from Q1 to Q2.

Viewed vertical by vertical, telecom was the leader in RTB market share throughout the year’s first half, with 11.3% in Q1 growing to 14.3% in Q2. Auto, computers, travel, and retail also ranked highly with more than 5% market share in both quarters. When you break it down in terms of willingness to spend, alcohol (#1 in both Q1 and Q2), travel, and auto were among the sectors willing to pay the most to win impressions.

DSP Trends

Marketwide, brands are not consistent in how they divide their RTB activity among DSPs. Some have a clear “preferred” DSP, while others divide spend between many bidding partners. The below chart shows five different (and not necessarily representative) brands.


There are two obviously dominant DSPs, with approximately 25% market share each. The remaining 50% of victorious auctions are divided among eight-plus DSPs. The below pie chart shows the spread of “won impressions” among all demand side platforms bidding across exchanges in Casale Media’s Index Platform.


Transparency Wins

Casale Media finds websites that are not transparent about selling inventory in exchange environments generated only 56% of the bid density of websites that identify the publisher source and context. That finding indicates publishers that cloak their brand over wariness about cannibalizing sales are not able to fully leverage the benefits of RTB. It’s a Catch-22 for premium media sellers.

Meanwhile pricing trends continue to be fairly erratic. Casale recorded a significant drop in highest price paid, but average CPM increased 7.8% from Q1 to Q2.  Lowest price paid also increased slightly, as shown in the chart below.


For publishers, one unfortunate conclusion of the report is that price consistency is probably not on the near-term horizon.

Casale said, “To have the same impression valued by 12 platforms at a [significant] spread, these markets have a ways to go before price discovery can really balance supply and demand, before there’s more consistency. It also potentially creates the case for opaque exchanges to become transparent.”

With more transparent marketplaces, he said, “You’d probably have a scenario where supply and demand get a little closer in alignment and value rises as it should.”



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. Who are the two dominant DSPs? I can guess one. But who is the other? I’m assuming this is US only though.

    • Exchange Guy

      Would have to be two of these three: Invite Media, Turn, Appnexus

  2. Buy Side Guy

    Self-competition to me was never really about DSP versus DSP as most buyers who use multiple DSPs still tend to use a single DSPs for each unique brand. I think the concern of self-competition stems more from how a single brand might not be able to de-duplicate adnetwork supply inventory from DSP/Exchange inventory; the theory that using both adnetworks and DSPs will somehow lift the overall price. Someone correct me if I’m wrong though.

    • I think some clarification would helpful RE: “self-compitition.” This can occur either by overlap between a DSP and adnet or via two DSPs. The re-targeting segment is of course the biggest offender as multiple bids on the same audience/cookies will ultimately drive up the cost of media. Considering Casale Media is an adnet, DSP, and exchange…I have to assume the article wanted to diffuse this “myth.” I am still not convinced.

      • Andrew Casale

        For the purposes of clarification, while yes Casale Media still operates its ad network, and operates Index Platform which powers internal and external ad exchanges, we are not a DSP, we do not bid in auctions, and the ad network component of our business is separate from the exchange.
        That being said, the underlying data we see is that same-brand biding activity, where the same brandID (something we unify across disparate DSPs) shows up in multiple bid responses to the same bid request is small. Some of the data we provided in the report also shows that the diversity of demand for the same impressions (across different industries, not the same industry) is generally high as well.

  3. Premium Publisher

    Interesting stats. It’s concerning to me as a premium publisher that 5 DSPs account for 78% of all won impressions. When each of those DSPs generally submits one bid on behalf of all of their clients, that makes liquidity pretty darn difficult to achieve. Any thoughts on that? Do ATD folks have an issue with internal DSP auctions determining whether their client’s bid wins against a competiting client and ATD on a piece of important inventory? Since most ATDs are using the same DSPs, I’m pretty sure that is what is happening. Right?

    • Andrew Casale

      This is an accurate depiction of what is happening today, and it is a problem.

      Fortunately, multiple bids per impression is actually becoming a reality, albeit slower than we all would have wanted. Our first wave of integrations along these lines are happening within this quarter.

  4. I think that it’s very important to understand the difference between bid density and the perceived dangers of self competition. From a publisher’s perspective this report is good news. Bid density from disparate sources means a robust market place. I worry, however, that advertisers will misinterpret the statement about true unique bid density to mean that it’s perfectly ok to employ multiple bidders for a single campaign.

    As an advertiser, the worry about self competition in RTB is that it creates an artificial price floor, one that is above that which the market would dictate. The headline to this article implies that there is zero self competition, an unlikely scenario as all that is necessary for self competition is for a single advertiser to have two bidders in play. A more accurate (though less newsy) headline would be “Self Competition is Not Pervasive.” If an advertiser is going to attempt a highly targeted campaign, such as remarketing, using multiple bidders makes it far more likely that these two bidders will collide on the same impression’s auction due to the narrow audience and high perceived value, and in turn be far more likely to create an artificial price floor.

    • Guy Who Gets Remarketed To A Lot

      What about this scenario? One DSP is bidding on behalf of an advertiser’s remarketing campaign and three impressions for a user in the pool simultaneously come up for bid. The DSP participates in all three auctions since the user is highly valued and it isn’t sure which impressions it will win. The DSP wins all three impressions. The problem is that the impressions are all on the same page. The advertiser just overpaid by a factor of three to hit that one user at that time and place.

      I see this happen all the time and no one seems to think it is a problem. Why?

      • Andrew Casale

        To mirror the last comment’s sentiment – what you’re describing would only be happening if the buy was performing. And to take a page out of the direct sales book, this effectively acts as a roadblock, which in direct sales is sold at a premium, not a discount.

        If this scenario were common and waste was excessive, it wouldn’t take much to tweak the bidding algorithm to include a win probability rate such that if the same cookie generates multiple impressions in the same second it only bids on x%.

  5. Why does it matter that the highest price paid came down when the average went up, as well as the lowest price paid? This means there will be MORE price consistency in the future. Extend the graph in Fig. 1 rightward and I would expect to see a convergence at a still higher average CPM–that’s better for everyone as pubs generate more revenue per impression (so what if the most expensive ones cost less? this is about volume) and advertisers inch closer to “premium” impressions in RTB that they know more about, and can more confidently bid on.