GroupM cited two major factors motivating its (alleged) imminent departure from the Open RTB world: 1) fraud, and 2) controlling price via scale. Fraud has been discussed plenty by others, so I’m here to talk about scale. The ultimate conclusion is this: There are clear benefits to being a large advertiser in Open RTB, but it is not because of anything GroupM -- or any other buying-power differentiated media agency or organization -- can influence.
Programmatic and (especially) Open RTB was supposed to make size and scale irrelevant in display advertising by making every impression available in individual auctions, thus removing any human bias. But it hasn’t. If anything, it has reinforced the size chasm among advertisers. Open RTB presents the double whammy of increased demand via a crowded competitive market and reduced supply via a direct correlation between relevant impression volume and advertiser size.
Paid search allowed SMBs to compete with large advertisers simply because its mechanics reduce competition and provide (effectively) infinite scale on search query variations. The crucial part is the reduction of competition: a user declares intent for one single thing at a time. Each search query makes that searcher irrelevant to the vast majority of search advertisers, since in that moment the searcher has no need of any other thing outside of what shespecifically queried. Google and Yahoo/Bing even protect this effect by rewarding relevancy. Additionally, the more specific the query, the more relevant the query is for specific businesses, thus improving the chances of winning for those exact businesses (think “taupe upholstery loveseat repair in Bucktown”).
Large advertisers satisfy their need for scale with broader terms and achieve acceptable efficiency via fancy tools, but there are still plenty of niches with the right combination of competition and relevance for SMBs to find success right alongside.
Open RTB is the exact opposite. At any given moment, a user has cookies for dozens of sites she’s visited, plus cookies for predictive and behavior-based segmentation. These cookies are used to signal the same thing as a search query: intent. But now instead of having intent against one specific thing at the moment of the ad call, that person has intent for dozens of things. So not only do SMBs have to compete against direct competitors large and small for broader intent signals (e.g. “furniture repair”), but they also compete against dozens of companies they don’t compete with anywhere else in their business.
While the demand side of Open RTB makes things tough, the problem is exacerbated by supply as well: the SMB has many fewer people to target than the large advertiser, and thus fewer impressions it wants to buy. The SMB MUST get its win rate as close to 100% as they can (e.g. spike their bids) to deliver enough scale to be successful at creating demand. The large global advertiser has plenty of people and impressions to target. Higher supply means lower cost, and thus large advertisers can be much less aggressive with their bids. Lower supply means higher cost -- another disadvantage as the advertiser size shrinks.
Back to GroupM. Scale influencing price for clients is really GroupM’s major point of differentiation, but nothing of the above is within GroupM’s control, and company principals know this. Their clients come to them for leverage they can’t get elsewhere, as GroupM makes media markets with its aggregated spend. While it will be interesting to see if GroupM merely buys up inventory transparently via private marketplaces for clients, creates a GroupM premium network, or does something else entirely (pizza party!), what’s relevant is, GroupM’s abilities could not bend the game to its advantage anymore, so its principals are taking their ball and going home.