The Official Blog of Cogmap, the Org Chart Wiki


Archive for February, 2009


iPhone Paid vs. Free Debate

Wednesday, February 25th, 2009

Lots of people in the blogosphere talking about Greg Yardley’s analysis that it doesn’t make sense to give your app away for free on the iPhone because the economics of paid are better.  Couple of points he didn’t make in the deck that I wanted to make sure people thought about as they read it:

  1. It appears that the new metaphor that Apple has propogated that people pay for iPhone apps has worked.  On the Internet, people expect everything to be free, but on the iPhone, people expect to pay for some portion of their apps.  We can suspect that this is true because Greg indicates that the difference between free vs paid is 6.6 to 1.  That is a lot different that the Penny Gap that most web developers build around, assuming that it is more like 100 to 1.  If it were 100 to 1, the economics change pretty quick.
  2. There are probably data normalization issues here.  How do you really compare apples to apples?
  3. Finally, Greg assumes that everyone is using ad networks to monetize all of their inventory.  If you look at the big guys on the Internet, they would tell you that no one but arbitrageurs can survive at those kind of CPMs.  Just like on the Internet, apps should try to figure out a way to charge premium CPMs.  If it was a $10 CPM, all of the sudden it is a more interesting question.

I love how Greg put together a nice data driven post.

4th Generation Ad Networks: Were they actually 1.1?

Wednesday, February 25th, 2009

Jeremy Liew has a nice post on 4th generation ad networks.  Go read it, I commented, but I wanted to expound on my comment a bit.

In my mind, as he described it, 1.0 was inventory aggregation.  DoubleClick ad serving was, as he described, one approach here.  However, before his 2.0 – “Data” – there was “optimization”.  Optimization theoretically gets more interesting as you add data.  Initially the only data point was “They are looking at this page”.  Now we can layer on more data.

In a world flush with data, he who can most effectively utilize all this data wins, right?

Trying to post more!

News Loves A Good Sound Bite

Wednesday, February 18th, 2009

The local news crew has locked on to the newest Facebook sound bite:

“We are the 6th biggest country in the world!”

To wit:

So I was thinking, what soundbites could I produce for the media that feed on this particularly meme?  Here are a few:

  • Google is twice as big as the United States – you would think that would be good for the stock price.
  • AOL is the 4th largest country!
  • Twitter is one of the 100 largest countries in the world!
  • Cogmap would be one of the 200 largest countries in the world!
  • The recent “Conficker” virus attack would make Conficker the 90th largest country in the world!  Clearly a plot!
  • Bebo is bigger than Iraq and twice as hard to withdraw from!
  • Apple would be the 7th largest country!
  • Wikipedia is almost, but not quite, bigger than the United States.

Last but not least:

  • Microsoft Windows is the biggest country in the world!

To compare services to countries is preposterous.  Lots of people have diverse constituencies. zips toward the top

Tuesday, February 17th, 2009

Everyone in the office is abuzz with talk of  Fmylife has skyrocketed to the top of traffic statistics in a very short period of time.


Fmylife is a remarkably simple application.  There are only three things you can do on the site:

  1. Read stories
  2. Vote that the person deserved or did not deserve the terrible thing that happened to them
  3. Submit a one paragraph story that starts with “Today, ” and ends with “FML”

The design has two images: a “new” icon next to new FMLs and a “share this” icon.

So this is not popular because it has a sleek web 2.0 design or some amazing functionality.  Almost any Rails developer could probably code the entire site in a couple of days.  This was simply good execution of an idea and effective development of the concept as a marketing meme.  Amazingly, when you look at technorati, there haven’t been a lot of major blog posts about fmylife to date and not even a substantial amount of long-tail blog posts.

What made for great execution?  The barrier to participation is remarkably low.  Voting is a simple click with no login.  Submitting a story involves giving yourself a nickname, picking a category, and writing a sentence.  With barriers this low, it is easy for even casual readers to get involved – and many do!

We can also look at the ranks relative to other web sites:


Wow, FML is as big a hit as MSNBC. I bet it costs a lot less to build and run. It grew faster than Hulu and Hulu had the best content licenses in the history of the Internet.

A great example of the power of a great idea.

Great Example of the Pressure of Quarter to Quarter Earnings

Tuesday, February 17th, 2009

A great critical analysis of the challenges that MapQuest faces is on Silicon Alley Insider, written by the former CTO of AOL, namely someone who theoretically had the power to make change and apparently did not?  I hate to see crying out of school like this, although I think he does a nice job of breaking it down.  Let me be clear: I don’t know him, have never met him, and look forward to being his friend as he is way more of a power broker than me.  I hope that accusing his post of crying out of school does not taint our relationship.

This is also a great example of something people love to see in blogs: The comments are more constructive than the initial post.  Very engaging material is an SEO and traffic gold mine.

Follow on comments discuss how opportunities to invest in MapQuest were passed over because contributions back to the company were needed to make quarterly numbers.  This long-term vs short-term tension is found in every public company as their revenue numbers become increasingly difficult to hit.  When McKinley discusses how straightforward the Google UI is vs the ad-ridden Mapquest UI, the next question is obviously one of “How are you supposed to build a sustainable business”, because Google’s is clearly heavily subsidized today.

Of course, now you are seeing these same problems at all kinds of businesses as the economy goes south.  Google is starving the losers.  If Google Maps had not taken the share it had taken in the last year, it could easily find itself on the Google scrap heap as they arbitrage their business to make increasingly challenging quarterly numbers.

Speaking frankly, some of these same challenges exist for me, working at an AOL subsidiary, but the problems are far more nuanced than the comments on this blog imply.  I think everyone feels like we are still in the second or third inning of the ball game.  There are literally hundreds of tremendous growth opportunities facing the company and we are forced to make decisions about which to pursue.  Even if AOL required no contribution, while we could choose more, we could not choose them all.  It is easy to point the finger a few years later and say, “We should have done that one.”, but every decision made involves dozens of negative options.

Furthermore, AOL has macro resource decisions that need to be made.  They are making decisions across the organizations: Should investments be put into existing profitable properties or invested in new initiatives that have not yet gotten traction.

Other comments discuss the choices made in API delivery: Google offers a free API with no support, Mapquest requires payment and extensive support.  Mapquest targets the corporate behemoths, Google targets the long tail.  Probably a valuable lesson.

Finally, a great comment by a former AOLer discusses the SEO impact of Mapquest decisions.  Mapquest did a poor job of SEOing their site.  When they dropped from #1 to #2 in map search results, traffic dropped tremendously.  Now they are #4 and Google is #1.  Frankly, that could explain virtually the entire tradeoff in traffic.  A 20% traffic decline for Mapquest as they drop from #1 to #4?  Sounds right.

GroupM Makes Their Advertising Data Confidential

Monday, February 2nd, 2009

GroupM has added a new clause to their terms and conditions indicating that they own all of the data associated with ads they buy and that data should be treated as confidential.

Traditionally, the standard has been that the advertiser, the media buyer, and the publisher all share a right to the data.  Now GroupM has claimed that they and their advertiser (contrary to PaidContent’s coverage) own the data.  There are some caveats, however, that are elucidated by GroupM’s COO:

“The real issue around that is the old terms and conditions didn’t really accommodate the data question properly, and our clients as you might imagine, have so much data generated from advertising these days, that they wanted to make sure their insights about their advertising is kept confidential. And that the data is used by the publisher only in aggregate, not at a granular level,” Montgomery explained to Online Media Daily.

He says it is not the intention of the new GroupM T&Cs to preclude publishers from utilizing data derived from GroupM’s online media buys in “aggregate,” and that the effectiveness of many of GroupM’s online campaigns depends on publishers doing just that.

“We acknowledge that they have to use the data in aggregate, and in fact, in order for them to optimize and benefit our clients,” he says, conceding, however, that data ownership will likely be the most contentious area of the new T&Cs.

(courtesy of MediaPost)

It is a shame that his comments are not part of the legal agreement because it is really kind of unclear what he means.  I know he doesn’t want me to tell competitors, “This is how this campaign performed.”  When he says, “in the aggregate”, can I tell competitors, “This is how these kinds of campaigns perform”?

Furthermore, how does this relate to blind behavioral data.  If a company was gathering behavioral data and they wanted to say, “This user likes car ads”, is that OK?  Is that aggregate data in that we lump them into an “auto” behavior?  I assume it is.  I assume if I create a “Likes Audi cars” behavior from an Audi campaign and then sell it to Ford for conquesting, that is considered too specific.  This leaves a lot of room, but it would be nice if the acceptability of aggregation is defined better.