The Guardian, Rubicon Project and Buy Side Fees

You likely read this week that The Guardian sued Rubicon Project in a noteworthy ad tech case. The premise of the suit is that Rubicon overcharged The Guardian. All told, only a few million dollars are at play, which is financially immaterial for either company, and it is actually likely that The Guardian will lose, but this is a very important case given the importance of market perception here. We will discuss all of this below. 

download (9).png

The issue in the case is potentially undisclosed "buy-side" fees. For an SSP like Rubicon, the normal way it would make its money is a revue share with the publisher. The largest SSPs generally have fixed, contractual revshares - meaning they specify a certain percentage that they will take. For the sake of argument, let's say it's 20% for the SSP. In a standard second-price auction, if the first bid is $10.00 and the second bid is $0.99, then the auction clears at $1.00. The publisher would make $1.00 * 0.8 and the SSP would make $1.00 * 0.2. So it stands that the publisher may not make less than $0.80 in this auction - but can the SSP make more than $0.20? Well if it charges the buyer $2.00, because it's indicated it is willing to pay up to $10, then the SSP now makes $1.20 and the publisher makes the same $0.80. The extra $1.00 is a "buy-side fee" - meaning it doesn't change what the publisher was entitled to and theoretically is only incurred on the buy-side.  

Buy-side fees are, in fact, disclosed in the Rubicon publisher contract, though they are not specified. The Guardian conducted test campaigns recently where it purchased its own inventory and evaluated what percent of the media cost it received. If you assume a fixed and transparent DSP revenue share (this is questionable), and The Guardian purchased its own inventory in the above scenario, it would have seen a decidedly lower percent revshare than it had in its contract and likely been sufficiently irritated to file this lawsuit. But because the buy-side fees are in the contract, it's unclear what the merits of this case are. 

Two high-profile players in the ad tech space litigating fees means that these practices will be brought to the forefront. Major SSPs, with only a couple exceptions (I can only think of Index and maybe OpenX), take buy-side fees - meaning it's not a one-off situation. It is likely that The Guardian was aware of the contractual terms allowing Rubicon to take a buy-side fee before this suit was filed (lawyers should actually read the contracts before filing a lawsuit), and either wanted to take a "pound of flesh" by shaming Rubicon publicly - as it has done - or tried to settle in private with the threat of public shaming, which is now being effected. 

In the example given above, the theory is that the publisher isn't hurt because they're given all the money that they're "entitled to" (the second price) and that the rest of the fee is just borne by the buy-side. This is incorrect for a few reasons. First, there is only a finite total RTB budget for display ads. This is between 1) fixed-revenue campaigns, where a certain dollar budget is allocated, and so every extra dollar taken by an SSP comes out of dollars that could have gone to publisher's pocket in subsequent impressions, and 2) performance campaigns, where the DSP will buy as long as performance metrics are hit - so higher costs mean lower performance, which means less spend. In both cases, the money is indirectly coming out of the publisher's pocket, but it's sufficiently obfuscated and not explicitly causal. In many cases, the auctions are for header bidding or other competitive impressions. This means that a higher buy-side fee might also result in a lower bid submitted to the header (v the SSP using that higher amount as the bid), with a lower win rate and a lower overall yield.

It is likely that at least some publishers that sign or re-sign with Rubicon will have serious discussions about buy-side fees going forward. As an aside, Wall Street is particularly sour on Rubicon. Their valuation is $300M at the time of this writing and they have $200M of cash. This means that the value of the going concern is only $100M despite revenues of $227M on a global platform spend of $1B. This basically means that investors believe Rubicon's current value proposition will fade away. Just a year ago, Rubicon was valued at around $1B, but a series of missteps have materially undermined the company. This lawsuit - more than its direct impact, means that its margins may face more pressure from its customers - will not help. Compounding this issue, companies like Rubicon will face margin pressure through header bidding (Rubicon states that it is the most installed header bidding buyer - not wrapper - though being the header buyer is arguably where the actual value lies) and a march towards first price auctions. This means that Rubicon et al will likely have to restructure themselves to be able to be successful in a lower margin environment. 

Every company in the ad tech space, TripleLift included, is on a relentless march towards ever-improving efficiency and differentiation. Ultimately, companies need a differentiated product and a way to drive value for all their constituents - or they will lose their customers. This case shows the cracks developing when these requirements are not being met - publishers are beginning to demand improved margins for commoditized products. TripleLift differentiates through unique and optimized placements, whereas Rubicon's value prop will likely increasingly become the part of its business that is less easily commoditized - PMPs and the sales force that it equips to help its publishers actually improve their revenue, and whatever else it comes up with. We should expect to see more ad tech companies develop unique value props and move away from being pure intermediaries.