Tuesday, June 15, 2010

Fiddling while Rome burns

I read Curt Hecht's AdExchanger interview this morning with puzzlement and, eventually, horror. Vivaki is drawing the precisely wrong conclusions from their evaluation of the situation.

[re Invite Media] Google realizes it's for the DFA stack, away from media, and they appreciate that it works just like search bid management or serving ads. It's a good thing for the industry that they're taking the interoperable view... I assume they'll eventually put them on the Google Stack, but for the time being we just want to keep progress going the way that it has been... I think [re the Invite Media acquisition] it's great that what you're seeing is some consistency where Omnicom and InterPublic Group... they both have come out supportive and positive.
If I'm reading this right, Vivaki thinks that Google wants a position in display like they have in search. Also, that Invite is not about media (I infer that it must, therefore, be about data.) And that even though Google is talking interoperability now, they will eventually integrate Invite into the rest of Google (making interoperability problematic, to say the least.) Oh, and they seem to think this is all peachy, and say that everyone else in the industry thinks so too.

This is why we can't have nice things.

For the sake of argument I'll grant that the agencies and their holding companies might not have been able to anticipate Google's dominance of search ads in the '00s. But let me nip future arguments in the bud: Google is trying to lock up display like it did search. Now you know. You're going into this particular battle with 20/20 foresight. If you do something stupid here, you've got no one to blame but yourself.

Google is a publisher and ad network. They make almost all of their considerable profit from people buying their ad inventory or that of their content partners. When any other piece of the value chain starts to look vaguely powerful, they commoditize it by buying and subsidizing someone who provides that piece. Urchin, Feedburner, Android, Teracent and (not yet subsidized, but mark my words) Invite Media.

In the short-term Vivaki will have lower costs. In the long-term the complement that will be commoditized is Vivaki. I know talk of disintermediating the agencies is as old as the DARPAnet, and I am usually one of the scoffers*. But this time it's different, for one important reason: data.

Google says that they are going to keep Invite as a separate entity. This is bull, as even Vivaki admits. The 2010 strategy du jour is to say one thing then do the opposite, and Google is a master of it**. Invite Media will be integrated with Google, and when it happens, a self-reinforcing cycle starts.

Brian Lesser said last week "we believe in the importance of proprietary technology to ensure the integrity of our client’s data... Every buy that an agency sends through a DSP makes that DSP smarter." If Google gives Invite access to the effectiveness and cost data from AdSense and GCN, Invite will have more data than anyone else in the business, by a long shot. By having more data, they can become more effective targeters, which will give them more market share, which will give them more data. This feedback cycle of proprietary learning will make it impossible for anyone else to compete in the market for targeting services. And Vivaki and the rest of the industry*** will end up as non-strategic customer service reps for Google's media planning and buying solution.

In fact, treating Invite as if Google's promise to leave it stand-alone were true is a mistake for everyone in the industry. It's easy to see, despite Neil Mohan's sweet-talking, that every DSP has to assume that each of their orders on the Google ad exchange will get seen by Invite. It's difficult to invest in novel strategies when you know your competition will see them in real-time. Almost everyone else in the industry has similar problems, or just the awful problem of potentially dealing with a single supplier/customer.

But the conundrum the other exchanges have is the most interesting. Vivaki implies that Microsoft is psyched about the Invite acquisition. That's ridiculous. The integration of Hotmail inventory and AdECN into Invite is not a result of Microsoft wanting to work with Google. Integrations take time, so this one certainly started (and was probably complete) before Microsoft even knew about the Invite acquisition. Moreover, while Microsoft has proven itself to be a difficult political environment for ad companies to thrive in, the people there are not stupid, not by a long shot. And being excited about giving your nemesis access to your trade secrets would be very, very stupid. The same is true of Yahoo! and AOL.

Because when Invite is integrated into Google, it seems reasonable to assume that Google will:
  1. Start cherry-picking the other exchanges' best publishers; and
  2. Start front-running the other exchanges, keeping the demand for themselves****.
By giving Invite access to their marketplaces, Microsoft, Yahoo! and AOL give Google access to data about position and price of every ad that runs through them. They would be giving Google the very data it needs to outcompete them. If the other exchanges allow this, they won't for long. Because if they do, they won't be in business for long.

Darren Herman said to me "it's like we're paying Google to take our business." It would be one thing if companies lose to Google because Google just flat-out does things better: there's no crying in baseball. But the game's barely started. Keeping progress going the way it has been is the wrong strategy. Recognize the threat and respond.

* When entrepreneurs tell me "if the agencies don't adopt our technology the agencies will become irrelevant," I say "if the agencies don't adopt your technology, then you will become irrelevant." The agency owns their customer, the marketer, and they are good at and jealous of that ownership.
** Google's switch on mobile phones and Apple's bait-and-switch to app developers are two recent examples.
*** I was a little puzzled by Hecht taking everyone else's lip service on the Invite acquisition at face value. While everyone I know is still pondering what it means, no one is really fully on board. Of course they say they are, but when was the last time you heard a holding company executive (Martin Sorrell the exception proving the rule, as always) say anything revealing? For what agency execs really think, read Darren Herman's post on Dart and Atlas.
**** It's widely rumored in the industry that Google has a double standard in exchange pricing between people buying through Google's user interfaces and people buying through the exchange API. If Invite is an insider, it shouldn't surprise anyone if they get preferred access.


bhalliburton said...


As Jonathan Mendez says, while everyone else says they want to win in Display, Google is the one acquiring every leader in every market. DoubleClick Exchange and Right Media exchange interfaces are impenetrable. Google just acquired the friendliest UI for buying off of exchanges at a price that will quickly be rendered meaningless by any increase in volume on the DoubleClick Exchange or by leveraging the benefits of front-running inventory.

Also, one cannot discount the fact that Invite actually built all of their stuff in Python. The integration will be faster than people expect.

Jonathan Mendez said...

Spot on Jerry.

Thanks for the shout-out Brent and your last point re python is especially important.

Anonymous said...

You only need to read Google's job postings to understand their strategy. Nary a mention of the word "agency" in this one:

Vertical Head of Display, Google Display Media Team

As a Display Sales Manager, you'll work to manage, motivate, and develop teams that will provide online display advertising solutions for leading Fortune 1000 companies. The primary responsibility of the Display Sales Manager is to drive new business revenue for YouTube and the Google Content Network with Fortune 1000 advertisers in your vertical...


Craig Swerdloff said...

Ad networks are seeing pressure on at least two fronts; cost of inventory and buying efficiency. As Google gets better at monetizing publisher ad inventory, which they quickly are with AdX, the cost of inventory for networks is rising. With Google's acquisition of Invite, advertisers/agencies can efficiently access inventory and data themselves. Ad networks exist to buy inventory cheaply, package it with data appropriately, and sell it to advertisers and their agencies at a premium. That model is going to get much harder to do.

Scott Hoffman said...

Jerry, great post - all I can say is that some agencies digital media arm see the train (Google) a coming, but they can't get off the tracks - when the collision happens...

Ken Rona said...


Very nicely articulated post.

There is one advantage that you did not mention. Invite is going to be in the Google Cookie space, meaning they have the greatest reach of any DSP. And reach is critical for almost anything having to do with running an online campain. Targeting, optimizing, remarketing, etc. The indies will have a hard time duplicating 100% reach without working with major partners. I would expect a push and pull to operate in terms of taking out the Indies. RM might realize that they need to support the dark pools and the indies realize that they need big partner reach.

Jerry Neumann said...


Absolutely agree. Good point.

Love your blog (http://dafacto.blogspot.com/)