In a recent post I talked about a post on the YPN Blog about questions people should ask ad exchanges before working with them.
Jordan Mitchell left a good comment with some questions that I thought deserved responses in a post. He starts off with:
But outside the *theory*, are the concrete benefits really there yet, or is the exchange model dealing with some basic issues?
I think ad exchanges do still have basic issues, but what business model doesn’t? Paid search still has some basic issues, but that doesn’t mean it’s not a very good and efficient business. Additionally, the ad exchange world is already way beyond “theory”. The Right Media Exchange does a very significant amount of volume, with a lot of money transacting every day. And it’s still VERY EARLY in the evolution of ad exchanges. Plus, would Yahoo! pay hundreds of millions for theory?
— it’s all remnant inventory that can’t be sold direct, so the market has already valued it to some extent and it’s not much! Then you add in the 10% intermediary fee (not paid by you) and are you as a publisher better off? In other words, are the market inefficiencies that poor where the exchange model demonstrates continued lift? I’d like to see some case studies.
Initially ad exchanges have primarily operated in the remnant inventory space. It was the area of inventory that was the easiest to get customers to try things with, and it also is an easier technology then trying to set up an exchange for guaranteed inventory. However, we’ve definitely seen people start to blur the lines with what they do with exchange technology, and you can bet that ad exchanges will move into the guaranteed inventory world. I’m not saying that exchanges will ever replace a sales team and the fact that inventory is sold on a guaranteed basis, however, exchanges may become a large part of the underlying technology platform that sales teams and ad operations staff work with to reserve, book, and deliver guaranteed deals. The lines between guaranteed and non-guaranteed will eventually blur.
Yes, the ad networks and forms of inventory allocation used by publishers out there are so inefficient that paying an exchange a small fee to have it monetize that inventory better is worthwhile. We’re working with many of the largest publishers on the web, and doing a ton for the largest publisher (Yahoo!). Also, non-enterprise publishers can work with the Right Media Exchange without paying any fees by using our Direct Media Exchange product.
You can find case studies for some non-enterprise publishers on the Direct Media Exchange site, as well about 20 featured publisher articles on the Right Media Blog. A couple of larger case studies exist for Tickle and Looksmart. There will be many more coming in the near future.
is the demand greater than the supply? I saw your other post where you mentioned 6,000 buyers and 13,000 selllers on the exchange. If there’s just a lot more supply than demand (which I believe to be the case in display ads), then I’m not sure an exchange will do much for you. If the demand is greater than supply, then you’re better off selling direct yes?
Supply and Demand isn’t really that black and white. Both can be broken down into really granular parts, which is why we actually break it down to each individual ad impression. As an example, let’s say for the entire exchange overall there is more supply than demand. That doesn’t mean all supply is worthless. There could be a lot of demand for female Canadian impressions and not much supply. By using an exchange, the publisher has a much better chance of monetizing female Canadian impressions than if that impression was randomly allocated to a USA-based ad network. The exchange allows all connected buyers to see the characteristics of that impression and bid appropriately for what it’s worth to them.
One might even argue that if there is more supply than demand overall, an exchange is EVEN MORE valuable because it increases the amount of demand you can get access to all through one technology platform. Might as well have a lot of ad networks bidding for that impression instead of just the one that you sent that impression to through a rules based ad server.
In the case where you have more demand than supply, if it’s possible to sell directly for the best prices that’s great. But might you make even more if you allowed all those demand sources to bid for each impression individually?
don’t most ad servers now allow you to optimize their queue using eCPM rules (instead of schedule)?
Yes, but that’s pretty much “dumb optimization”. Essentially what occurs for non-guaranteed inventory is publishers set up a cascading daisy chain of ad networks based on historical pricing. If you think about what’s going on you aren’t getting the most value for each impression. The ad network at the top of your chain may choose to take an impression that a network further down your chain might have paid more for. An exchange concept blows that out of the water. The historical problem is that exchanges need to have enough demand to compete with the largest ad networks that are the primary placeholders in publisher daisy chains. And it’s happening.
Great questions from Jordan, and this kind of dialog is fantastic for helping explain the concepts of ad exchanges and what’s happening with them in reality.