The Official Blog of Cogmap, the Org Chart Wiki


Archive for the ‘Online Advertising’ Category


The Lure of Mediation

Thursday, March 17th, 2011

UPDATE: I love blogging because from time to time someone corrects my ignorance and I become smarter. I just received a lovely email indicating that AdMeld is now selling directly to agencies. While my source KNOWS STUFF, I am emailing Ben Barokas for comment. Shit, when did I turn into a reporter?

I am amazed at how many companies – ad networks, DSPs, and others, are considering getting into the mediation business these days.

The lure of mediation is obvious: If you see 100% of a sites inventory, you are seeing a much higher frequency of impressions, giving you more opportunities to monetize, you are potentially seeing some users that didn’t have the frequency to reach you in the daisy chain (super profitable), and you have more information about the absolute frequency of users (how many ads they have seen on the site).

This leads to a number of great monetization improvements:

  • You can front-run inventory
  • You have insight into the aggregate value of users
  • You can front-run inventory
  • You can front-run inventory
  • You can front-run inventory
  • You can front-run inventory

The terrible thing about mediation is that with great power comes great responsibility.

I feel like a proper mediation platform offers a few things to publishers:

  • A responsibility to maximize yield for the publisher – i.e. not front-running inventory
  • A responsibilty to offer transparent reporting

Many of the best known mediation platforms – AdMeld, Pubmatic, etc. – explicitly do not sell to non-ad networks. This allows them to report transparently on ad network performance with no risk of conflict and desire to front-run inventory. This also means that there is no difference between maximizing yield for them vs. maximizing yield for the publisher.

If I were an ad network, I would love to offer a mediation platform, because I want to front-run inventory, but the offering is so disingenous to customers, it is hard to imagine that it is a sustainable business.


Are DSPs Building Justifiable Valuations?

Wednesday, March 16th, 2011

I am reading more and more stories about DSPs raising large piles of money:

Lotame: $34m

MediaMath: $24m

DataXu: $19m

TradeDesk: $2.5m

TradeDesk appears to have a nice, sane strategy. If Invite Media was the industry leader and they were worth ~$80m, then all the rest of these companies are a long, long way away from justifying their valuations.

I am hearing more and more stories about how much of the revenue flowing through these platforms is essentially ad network business where the DSPs are taking on risk to run performance campaigns for agencies. While that is really interesting, is that revenue valuable? While it seems like the IPO market is opening, acquisition should still seem like a viable way out. If you build a $200m top-line ad network, I am unsure if anyone will acquire you today.

I wish I had more to say here, but why don’t people comment and we can go from there.




Is Twitter Evil?

Sunday, March 13th, 2011

It is super tragic watching Twitter struggle to find a business model. As has been widely documented, they have made, at some level, a fundamental shift that implies that people should no longer be able to build businesses using the Twitter API.

Frankly, this change in their Terms of Service is nothing less than shocking:

Courtesy of TechCrunch

Now, I am sure that Twitter represents this as “focusing development activity”, but when you look at developer-focused companies (“developers, developers, developers”), developers have always migrated to the most profitable areas and been slowly pushed out of business niches by the integration of functionality. If companies make Twitter clients and fill them with ads, will those end up being more popular than a Twitter client without ads? If Twitter can’t make a client people want to use, what does that mean?

The flip side of this is that, obviously, Google doesn’t let third parties scrape their organic search results and not show the ads.

Twitter is struggling toward some sort of balance, but it is very tough to turn a ship.

What kills an ad network?

Tuesday, February 15th, 2011

Many people commented last week on the AdWeek article about’s decline – an article that I foreshadowed with my magnum opus dissection of ad networks: Whither Ad Networks.

While I agree with the gist of the article (obviously, based on my previous writings) – Ad Networks are screwed by the advent of exchanges – I think there is a bit more to the story.

  • sales organization basically disappeared for quite some length of time under Tim/Jeff
  • technology execution was not great/distracted during this key period

One of the things that shocked me – absolutely shocked me – was Tim’s approach to when he took over Aol. Rather than treating as the jewel in Aol’s crown, he focused on the content side of the business.

Maybe, coming from Google, he was spoiled by the best ad inventory in the world (pages with search results) and wants to recreate that awesome inventory, knowing that advertisers will follow. But what he had at was probably the third biggest aggregation of advertisers buying inventory in a marketplace after Google and Yahoo!. One would have thought that he would value that highly and focus on how to get more advertisers participating in that marketplace. His approach to achieving this was to minimize the sales force’s role in bringing advertisers into the marketplace and bet heavily on self-service. Unfortunately, self-service grew, but not in a Googly way. Google’s self-service was the way that people could access the most valuable inventory on the Internet (search results). The result was an incredibly strong gravitational pull. Further, the self-service platform was not production ready when he started to downsize the sales focus.

I suspect (and I have no knowledge of this from an insider perspective) that some of the consultants and advisers he brought in from Google and/or internal Aol people counseled him that Google + DoubleClick would run over and crush that entire business. The result is he became less invested in the business, took the sales force and diverted them to selling premium content.

Did execute poorly? Did the market change? Sure. But if you take 75% of the sales people and tell them to go sell Aol, that is bad, too. technology didn’t execute crisply either. They are in the midst of an ad server consolidation project now that will hopefully leave them with a more nimble, flexible architecture, but as a result of this and previous, similar projects, following the market with mediation technology, APIs for programmatic buyings, DMPs, and other technologies didn’t really happen. ended up bidding into exchanges to expand their reach (just do simply low frequency bids on run of network-ish inventory, low-tech), but not becoming an exchange because that required more technology execution.

So while it is easy to say that a new technology era happened, Aol failed to respond to that change and that is why they are no longer successful, that overstates the effect outside factors had on the state of today. I agree with the insider comments in the AdWeek article that internal miasma played a role as well. The decision to aggressively shrink the sales force and underinvest in the technology stack no doubt had huge impacts on the business.

Now, maybe Tim made the right decision – maybe executing on the tech would have not been successful regardless, so Google would have won, so this was the right thing to do. There is no way to know. My point is divorcing the external pressures from the internal changes is impossible, so no single factor pointed the way to the decline of the organization.

When you look around online advertising, DNA is imprinted on many of the most successful companies out there. Sadly, when you look at Aol, there is precious little DNA left and the DNA there is no longer given the tools to build that business.

Mail Ad Inventory, History, and Frequency

Tuesday, February 1st, 2011

AdExchanger said something the other day that really resonated with me:

Display advertising around email is no longer the bottom feeder of inventory I once envisioned.  “Sure, it’s cheap, but it’s email inventory” used to be a common refrain.

Times change.

In fact, over the course of the past few months, numerous industry people have suggested to me that the display ad inventory wrapped around email is a top performer relative to price as ad networks, demand-side platforms, trading desks and “all of the above” buy email inventory.

Funny he would say that: As I recollect, Yahoo Mail was always some of the best performing ad inventory around. Bottom line: It had huge reach and it reached the kind of people that tended to convert on offers. True story. In fact, Yahoo divided mail inventory into a few frequency tiers and priced the low frequency stuff quite exorbitantly as I recollect. And we paid. We paid! We bought all we could afford. It was consistently good inventory.

When Yahoo bought Right Media, a big part of the financial justification was that they could push all the inventory into RMX and see financial upside. (10% RPM lift on a gajillion impressions is how Biz Dev guys do spreadsheets to justify big deal acquisitions like that one.) Despite this, when push came to shove, it actually took a fair bit of time to migrate the inventory over simply because people like us were writing such big checks and it was hard for the sales organization to come to grips with the fact that they were not going to simply accept a giant order from us at the beginning of every month.

Of course, the high frequency stuff, like all high frequency stuff, was/is basically worthless. So there you go.

Is search inventory awesome because by it’s nature it is low frequency?

Mobile user agents is kind of a big problem

Tuesday, January 4th, 2011

Right when you thought browsers were crazy, we enter the mobile era – and I am here to tell you that mobile user agents suck!

Hint: If you are trafficking campaigns in DFP and you target iOS, it only targets iPhones, it will not target iPods. There is no way to target iPods!

Mobile user agents are like rabbits – there are a bunch of new ones every week. If you think about it, we had huge problems with browser user agents and there were only a few and they only came out with a new one once or twice a year. New phones come out every day. And they have things like “new screen resolution” that break everything.

Collectively, that is a loud sucking sound.

The Difference Between Ads You Love And Ads You Hate

Friday, December 17th, 2010

Scott Kurnit’s new AdKeeper is offering the following challenge, outlined by AdExchanger and Forbes:

In an interview with Ad Age’s Edmund Lee, AdKeeper CEO Scott Kurnit discusses his company’s ad-banner-keeping technology and boldly challenges Ad Age readers: “I will make an offer of $100 to Ad Age readers for the first person or the first 10 people who find me an ad — they’ll have to capture that with a camera or something — that wouldn’t be more useful to a reasonable segment of the audience if it was kept for later.” Not only is Kurnit giving away $100 (potentially), his company is also providing “keeping” services for free to big brand marketers for 6 months.

This is a hilarious offer because only an ad that explodes is not worth “keeping” for later. If it was a University of Pheonix ad, one that I would never click on, it would be worth keeping for later because I might change my mind. Irrelevant = worth keeping. If it was an ad I wanted to click on, “Check out new TVs at Best Buy”, then wouldn’t I consider surfing some more and keeping it for later? Maybe. Relevant = worth keeping.

Only an ad that said, “click in the next 60 seconds or lose me forever”, an offer essentially impossible to manage given Internet latency today, loses value in the future.

What about the 5th ad on a Google search result. It is simply wrong. But it might not be later? Seems easy to either take Scott’s money or have Scott deny that I am correct. The email where I read of this challenge had a Google sidebar filled with ads for Django consultants. I am not even remotely involved in a Django project right now. Does Scott owe me $100? Might I want Django consultants later?

Frankly, my original thesis when I started this post was that only the most viral and interesting ads are worth keeping to most consumers. Once again, proof that lucking into mind-blowing creative is probably the greatest marketing there is.

I am not a believer that consumers will want to keep enough ads to make it interesting and the ones they do want to keep are already viral Youtube sensations, so the earned media effect of keeping an ad is probably de-valued. My wife can’t figure out how to get interstitials offer her browser fast enough. The ability to keep an ad and replay it later does not make the ad more engaging on an a priori basis. The ad must already engage to generate a keep. I think that is the bigger problem in media today. What ads here engage me? Elf Myself? Subservient Chicken? Having said that, my impression (I don’t know any of them except Sim Simeonov) is that Scott and his team are super smart. It will be interesting to see where they pivot from here.

Reaction to Yahoo’s Acquisition of Dapper

Thursday, October 7th, 2010

AdExchanger didn’t ask me for my opinion in their blurb of opinions on Dappers acquisition, so I thought I would blog my own.  Here you go folks:

“Yahoo’s decision to acquire the premier event marketing company in the online advertising space should come as no surprise.  Whenever I go to a Fixing Advertising event, I always felt like it was over-run with Yahoo people.”

That is what Dapper does, right?  Throw those great parties all over the world?

AppNexus sponsors Cogblog, raises $50m Series C

Tuesday, October 5th, 2010

Coincidence?  I think not.

Ad System Ecosystem of 2012

Tuesday, October 5th, 2010

Everyone loves a good ecosystem slide.  In fact, in the ad space, virtually everyone has seen the GCA Savvian slide re-used in 100 different contexts:

When Terence Kawaja left GCA Savvian to form Luma Partners, he promptly put out another one.  In fact, he gave a pretty laminated version to AdExchanger while I was standing right next to John and when I asked for one, I got blown off!  True story.  Anyway, this is the ecosystem in 2010 according to Luma:

I have good news, and a revelation for people in the online advertising space.  I have traveled into the future and brought back Luma Partners slide from 2012.  And lo, the slide was presented:

Or was this from a secret internal Google presentation last week?  Only TechCrunch knows for sure!