Open Source the Ad Exchange

My last post, long and long-winded and written when you should all have been out celebrating, sent more traffic my way than any other post in this sleepy blog’s history. I think the prospect of a new year and (in a sense) a new decade, makes us all a bit introspective and open to thinking the why about our path through the world. I’m glad people found it worth a read.

But it’s January now, and Monday. Back to work.


My work thesis since I left Omnicom in 2001 was that the advertising ecosystem was stuck in the 1970s without even the consolation of the three-martini lunch. Ad agencies–and thus advertisers–seemed to be following a rather indirect path towards connecting the right person with the right product. Efficacy was low, efficiency was poor and ROI tracking was primarily hand waving. “One to one marketing”, an idea that the internet could make real, was still used as an exciting cliche in pitch decks and book titles, but never implemented in actual marketing campaigns.

So I spent my time investing in and starting companies aimed at changing that. From early (too early, it seems) location-based marketing in Platial and exchanges in Root Markets, to tracking in Pinch Media, to data targeting in 33Across and Domdex and to demand side platforms in CPM Advisors.

But when Google launched their Ad Exchange four months ago, I decided the thesis was played out. An open marketplace can be the apotheosis of efficiency and efficacy. And if Google is putting its muscle behind that, then the rest is just filling in the blanks. While filling in the blanks is probably a better investment philosophy in the short term, it’s not the right place for investors like me who are putting smaller amounts of money to work for longer periods of time.

Then Google acquired Teracent. Then it was said that Google was looking to acquire a demand side platform. I also heard that Google was restricting sellers’ agents from working with AdX.

The key to making the exchange-as-platform useful is making it a platform. If the New York Stock Exchange owned a brokerage firm, that would be a problem. If they owned a research firm, or an underwriter, that would be absurd. But Google is, in essence, trying to do all these things. The strategy seems to be to make the exchange easy to use for advertisers and consumers because those are Google’s real customers.

This is a bad long-term strategy. Small producers of say, soybeans, are not equipped to show up on the floor of the CBOT and sell their September crop in open-outcry. And you and I are not prepared to go up against Goldman Sachs in selling our shares of GOOG on the Nasdaq. In both cases, we’d get taken. That’s why there are intermediaries. If Google doesn’t allow outside firms to be intermediaries then small producers and consumers will get arbitraged, and the information generated by the market will be poor quality. The hoped-for efficiencies will not appear.

Google knows this, of course. So why are they doing it this way?

In 2008, the NYSE Euronext Group had revenue of $4.7 billion and operating income (backing out one-time impairment charges and merger expenses) of about $1.1 billion. In 2008, the CME Group had revenue of $2.6 billion and income before taxes of $1.2 billion. The stock and commodities markets are each a couple of orders of magnitude bigger than the advertising market. Even if the ad exchanges facilitated the introduction of derivatives, like ad futures and the like, the future of ad exchanges are companies with a few hundreds of millions of dollars in revenue.

Being an exchange is not where the real money is made in the stock and commodities markets. The big money is made at places like Goldman Sachs. This is not to downplay the central importance of the exchanges: Goldman would be a much smaller and less profitable company if there weren’t exchanges. The ad markets will end up similarly, with exchanges that are crucial to the ecosystem but not that big financially, and much more profitable firms that use the exchange platforms.

Historically, exchanges have been cooperatively owned entities, in existence to facilitate the business of their members. Google, with AdX, is in the wrong side of the business. The best move, for them and for the markets, would be to spin AdX out into a Mozilla-like non-profit foundation and then work on the hard problem: figuring out how to use the newly generated information and efficiencies to make advertising more useful.

I jumped the gun. In this sector, there is a lot more to do.


  1. Jerry, again you nail it. This speaks to many of the conversations we’d had together as well, as, how to position a company in the marketplace to become the $GS.

    There needs to be an independent exchange that:
    1) doesn’t own a DSP
    2) isn’t a pub side optimizer
    3) isn’t selling the inventory it’s hosting
    4) has interests aligned with all of it’s strategic partners

    The issue that many startups & companies will face is that it’s much more profitable and sexier to be the $GS. Risk vs. reward. Why “settle” on building out the exchange?

    I certainly think they can exist – and should – as we do not need those 3 martini lunches :)

  2. Hi Jerry,

    Great post. We are working on providing pretty much exactly the independent, fair and open service you outline via the OpenX Market which we launched in April 2009. A brief sample of coverage: (techcrunch) (media post)

    Interestingly, we match all the criteria than Darren outlined in his comment.

    We’re making a ton of progress as there seems to be a real market need for this.

    Tim Cadogan, CEO OpenX

  3. Tim–

    I’ve been following what you’re doing. I agree with your approach. Google, though, with its existing market power has taken a good chunk of the market and the mindshare.

    Exchanges have both economies of scale and network effects, so there’s often one winner per category. I think it’s important to have people think about where we’re going with this while it’s still early.

  4. Jerry, you had quite a few insights from making an analogy between ad exchange and financial exchange markets. It is great. However, we need to be cautious of the limitation of any analogy.

    In my 2c, Ad Exchange marketplace is an order of magnitude more complex than NYSE! The current ad buying/selling ecosystem already underscores this intrinsic complexity – and the game just begins.

  5. Google has indicated they don’t expect to make much money from their exchange. So why do it? I think a couple big reasons are (a) data, which feeds every part of their business, and (b) a necessary feature to have on top of DFP to keep that product going for a while. Both data and their DFP customers are necessary components to their strategy to “AdSensify” (or “Searchify”, you get the idea) the display ad biz.

  6. Jordan–

    I agree that this is their strategy. But is it a good one? If the buyers and/or sellers are not too numerous, then they can pick up and leave a market that doesn’t meet their needs, is too expensive, or just plain bothers them. I think Google’s seeming plans for AdX may cross that line.

    Display is not like search: buying is more concentrated and, because display is much more complicated than search, it will remain (or become more) concentrated. Google, IMHO, is opening themselves up to a disruptive competitor.

Comments are closed.