Why The World Needs an Ad Exchange

Preamble

The exchange is a pretty beautiful idea. There are a few companies out there pursuing it as of this moment. I'll offer my admittedly biased opinions about them in a future post. They are AdECN, RightMedia, AdBrite, and some others. Some of you may recognize these names from a previous session at ad:tech NY that I and several hundred others attended. Great session and I hope you attended and enjoyed as well.

Let me address the question posed by the title of this post first, however. I believe there are forces at play that are driving the world towards an exchange. As a symptom, we see ad network after ad network popping up – sort of an intermediary step towards the exchange. Here are some of my favorite forces:

• A frightening and dominant competitor -- perhaps you've heard of Google (ironic that I'm using their blogging platform). To agencies, advertisers and publishers they are scarily reminiscent of a monopoly. And they are most definitely not friendly (some might go as far as to say evil) in the eyes of these constituents. As much as consumers love Google, advertisers and publishers fear and therefore pretty much hate Google.
• The sneaking suspicion on the part of publishers (in particular) that there is more value to be extracted from their content. Similarly, advertisers, drunk on the returns from search, imagine that all media could be this good.
• Stock Exchange Envy. Everyone in the online ad industry is greedily looking at the stock market and the analysts running it that make millions of dollars and thinking "we should be making coin like that." We’re well paid in this industry, but we’re faced with every other person that we know striking it really really big. The exchange promises the kinds of money that Wall Streeters make.

But really the biggie is Google, isn't it? They make so much money, and they touch almost everyone lately. I’ll focus on Google and what they tell us about an exchange.

John Battelle's Searchblog does a great job of discussing Google and Battelle’s book The Search offers an excellent history of the evolution of the paid search auction in chapters 5 and 7. He points out that we in the industry (Google, Yahoo!, Microsoft, and Ask) are pretty much standing on the shoulders of Bill Gross, who effectively came up with the idea of auctioning leads to advertisers. It's that auction that makes us think about the stock market and the exchange.

The Auction Enables an Exchange

An auction needs participants. It needs buyers and sellers. The good in an ad auction has become the impression (the search world sells clicks but we're going to forget about that for a few posts -- please bear with me on that). The buyer is the advertiser, the seller is the content publisher.

Every auction needs participants. It needs buyers and sellers. The good in an ad auction will be the impression (the search world pretends to sell clicks but they have basically tricked advertisers into buying impressions. We're going to forget about that for a few posts -- please bear with me). The buyer is the advertiser; the seller is the content publisher.

An efficient auction values its goods effectively. From the seller’s position, that means selling the good for the most money possible. Because different individuals have different utility curves, the same good will be valued differently by various individuals, or, in this case, advertisers. Theoretically, if every buyer were to participate in the auction for a good, then the auction becomes maximally efficient. If one buyer participates, then the item either does not sell or only sells for some fraction of the buyer’s value for it. Somewhere in between those two extremes lives the reality of auctions for online advertising today.

There is of course mirror image issue going on with buyers. From their perspective, they want all the supply (ad inventory) in the auction so that their bid will find its best use and value.

The role of the exchange is to increase efficiency in the auction.

What Google's Getting Away With

Google is reputed to have some ungodly number of advertisers and publishers. It’s impossible to keep track, but on the advertiser side people talk numbers like 400,000. On the publisher side, last time I heard it was slightly fewer. Because Google has so many advertisers (buyers), it values ad impressions more effectively for publishers than pretty much any other auction (by the way, I’m ignoring nuances of their technology which add to this phenomenon – they’re ahead there too). In fact, they’re so effective that their standard deal with hundreds of thousands of self service publishers is to simply say “We’ll pay you something that we’ll figure out later on the fly. We think you’ll like it. If you don’t like it, you can try other networks and see if they’re better.” This drives publishers crazy but in a way it’s really not so unreasonable. If it were, AdSense wouldn’t be the most popular monetizer of remnant inventory out there. What’s frustrating to publishers is that unless they’re premium, they don’t know what share of the revenue that their site generates they are being paid. Google’s network is blind in that sense (as is YPN and others, I should point out). By the way, according to the analyses I’ve seen it runs on average around 70%.

Take, by contrast, the fee structure of the NASDAQ. There’s a one time application fee of $2,000, an annual fee of $1,200, and transactional fees between a quarter and three tenths of a cent per share. That’s not including rebates that you get access to depending on the liquidity you bring to the exchange. You as a day trader don’t even get charged these fees because they are paid by the brokerage that connects to the exchange. Those brokerages tend to charge you what amounts to transactional fees as well. What I’m saying here is that Google is extracting (more or less) usurious fees from sellers by having (more or less) cornered the market for buyers. In the long run I don’t believe they should be able to charge a revenue share, much less not tell you what that share is. The platform must charge transactional fees so that others may flourish.

I have a friend who says “Google IS the exchange,” a statement not too far off from the truth as of this moment. But they’re not the right exchange. With them in the driver’s seat it’s analogous to Goldman Sachs running the NYSE and the NASDAQ. Do we want to live with that?

What is the Alternative?

In the post exchange world we imagine a rich environment wherein a variety of players get to add value in a variety of different ways. But the benefit of the exchange will start with publishers. By aggregating the demand from advertisers, the exchange will create efficiency of inventory pricing. But the difference will be that an advertiser will have more than one broker to chose from to access all supply; similarly, publishers have multiple brokers to chose from in order to access all demand. The concept of the brokerage is new, and different brokers may have different characteristics in what they excel at.

When this happens the value of owning every advertiser, as Google does, is diminished. I believe that the value of the content that is being produced on the web will become paramount, and there will be a shift in power towards the generators and aggregators of that content. This is another reason why Yahoo!, Microsoft, IAC, FIM, and others will want the exchange to happen, though I’m not sure if they quite understand that yet.

The power of the broker will increase as well. What is a broker in the online ad world? Well, they’re Yahoo! for example. A broker such as Yahoo has both supply and demand (their own as well as YPN’s). They will try to settle trades for impressions within their own brokerage first and turn to the exchange if that transaction fails (this is analogous to the way stock brokers work). But the exchange will inform the pricing of those over-the-counter transactions – so much so in fact that there will be very little difference between the above arrangement and the act of submitting every bid to the exchange directly and using your own brokerages highest bid as a reserve price in the exchange. But that’s kind of technical – I haven’t really even gotten into how the exchange will actually trade.

An interesting and algorithmic question is how the broker will make money in the future, particularly on exogenous supply (exogenous demand is more clear cut – they make money because the inventory clears at a more favorable price than endogenous demand will provide). I don’t believe it will exactly parallel the example set by the stock market – in fact I think it will be more rational. I’ll reserve that topic for a future post. There’s much more I could say about the topic of this post and I imagine I’ll return to this topic, but right now it’s late and I imagine I’m taxing the patience of my audience for reading in a single sitting.

Thanks for reading.

DM