Cogblog

Today I had this sudden sensation that many of my posts tend to be about how things are a little harder than people expect them to be.

Today’s “actually, this problem is pretty hard” is based on an article in Advertising Age about Portfolio.com.  Portfolio.com is planning to extend the reach of their web site by buying exchange inventory, allowing them to sell more of their audience in the face of sold out inventory.

Advertising.com has been selling this off and on for years under the name “Audience Extend”.  Theoretically, this is an extremely powerful product for an oversold publisher because it gives them access to the exact same audience on relatively inexpensive inventory.  This can be valuable to an advertiser that is trying to reach this audience and valuable to a publisher eager to continue to take revenue.

I have always found that these deals are extremely challenging to make work because in a situation like this, the salesforce of the publisher (in this case Portfolio.com) is used to representing a premium inventory placement (it is sold out!).  Once you start working to educate the sales force about how to sell in a network model (sometimes blind placements, uneven, usually non-guaranteed distribution of advertising across placements, etc.), you are taking your sales force out of your sweet spot and asking them to do something different.  Furthermore, usually you need to find someone on the network side who is experience at the network model sales approach and involve them in either training or directly supporting the sales cycle.  Obviously, this is not a job you can ask your worst salesperson to do.

Given the margin compression that results from the network sharing revenue with the selling publisher, plus the cost of inventory, it can be hard to justify on the network side investing appropriately in the partnership.  In the same vein, when the network offers to put forth substantial resources, the publisher sometimes becomes concerned that the sales force may take their eye off the ball with the constant barrage of network rhetoric.

Managing partnerships is a delicate challenge that necessitates both parties investing significantly upfront to maximize the likelihood of success.  It will be interesting to see if Adsdaq and Portfolio.com make this work in any context outside of a one-off deal opportunity.

So everybody knows that Technorati has been flailing around looking for a business model.

Unfortunately, the business model du jour is Yet Another Ad Network.  So of course, Technorati announces they are starting an ad network.  The only thing that could have been more depressing in terms of expressing strategic genius would have been announcing a bidded text link marketplace.  Frankly, this announcement has all the same problems.

What makes business models like Glam have a prayer of success relative to ad networks such as Advertising.com and Yahoo! are that they have a very focused group of inventory and advertisers that they are matching up.  This allows them to have a smaller sales force and publisher services team and still have good coverage in their target market.

Technorati launches an all things to all people strategy and says they will be able to overcome Google/Yahoo/Ad.com due to their awesome contextual search technology?  To be able to pay publisher more, I think it is a function of two things: Optimization technology and a strong marketplace of advertisers to bid up inventory opportunities.  Contextual targeting is a tiny part of overall optimization and without the marketplace of advertisers, great optimization doesn’t create huge value because if you are decisioning off of one advertiser, there is only one possible outcome, regardless of how much you optimize.

I can’t see how Technorati overcomes the marketplace building barrier.  For that matter, virtually any company executing a YAAN strategy today puts themselves behind the eight-ball.  To get access to inventory, you have to pay higher than Google and Ad.com.  They are setting the marketplace floor because they have a strong enough market to buy every impression and offer something.  To pay more than Google and Ad.com, that implies that you have to have some level of optimization and the marketplace.  Glam did this by taking their tens of millions in venture capital and using it as a loss leader to overpay for inventory relative to the value of that inventory to Glam.  Technorati doesn’t have that capital runway, despite the $7.5m in new capital they just raised.

Odds that Technorati can focus on non-premium inventory and build a marketplace fast enough to reach escape velocity: FAIL.

I say all of this in a nice way.  I love Technorati.  Use it all the time.

Just saw Facebook talk at Graphing Social Patterns.  It was basically a 45-minute ad for Facebook ads!  Drill down on their hyper targeting mechanisms, performance, etc..  As the speaker said, Facebook is an advertising driven business, so they walked us through their focus.

I think Facebook completely missed the boat here.  First, this is a geek conference.  They should have given this presentation at Advertising 2.0.

Second, I think this is not actually the optimal strategy.  What are they doing to help people using the platform make money!  Developers, developers, developers.  To justify a $15b valuation, what they need is an ecosystem.  They need multiple billion dollar companies using their infrastructure and then a tithing system to take a slice of the revenue.  To really scale, rather than focusing on selling their ads, they should focus on enabling people to sell their ads.

I love Google’s announcement that people can sell their own Youtube inventory.  Obviously, they are headed in the completely opposite direction by engaging the masses to make Google money.  Conceptually, they have the right strategy.  Tactically, they could still blow this - the minimum deal size of $10k seems to fly in the face of the validation Google has gotten around building a long tail of advertisers and publishers.  The smartest thing about this deal is that the concern with selling Youtube inventory has always been dicey content.  If people sell it themselves, then they represent their own content and ensure quality.  That essentially gets Google out of the quality discussion by having the inventory manage its own quality.  A slick solution to the problem.

Glam Media tries to do something new, which is nice, but not necessarily a brilliant idea.  They launched GlamTV, which licenses video content from third parties and allows their publisher network to then show the video content with Glam-embedded ads.

I assume this was driven by a lack of web sites that had significant video volume in Glam’s sweet spot, so they are trying to build out the content themselves.  My gut is that the best web sites in Glam’s network won’t want to use this syndicated content, so the upside to building their own video inventory and then building a video network on top of it will be minimal.

As I have long said when it comes to start-up ad networks, it just takes a couple of deals to look like you have a nice little business - they could close eight or nine $50,000 tests the first month and think (and sell to investors or acquirers) that this is the base of the hockey stick when it is really just the same kind of bump a new web site gets when they get TechCrunched.  Huge traffic for the first week and then a tiny fraction of that traffic actually sticks around.  Getting $50 cpms, as they have discussed at length, is great, but there are a bunch of other important questions: Sell-through rate and eRPM.  If they are selling all of the inventory at $50, that is a hit it out of the park business, but if publishers aren’t getting a couple of bucks (both the syndicator and the syndicatee, I would suspect), then it will be hard to build a viable business.  Can they build to a decent sell-through that nets out to an eRPM that high?  Seems challenging.

Some days I feel like 90% of my blog posts are posts disagreeing with other blog posts.  Am I the alternate voice crying out in the wilderness?

Anyway, today’s post that I take issue with comes from Silicon Alley Insider.  I actually have to dig in with a little of the line-by-line:

The burn rates and valuations of many Web 2.0 companies are predicated on this article of faith: The Web will benefit from a massive share shift as marketers choose the measurable (online advertising) over the not-so-measurable (newspapers, TV, etc.).

Wait a second, it is not just about measurable vs not, it is about where people spend their time.  I read recently that almost 20% of the media that people consume is online.  The percentage of US ad spend that is online is in the single digits.  So you have to assume that online ad spend in the US will double or triple in the next few years.  That means 30%+ sustained growth rates.  Even if ad budgets shrink, growth should still be significant.  Anything less creates a substantial arbitrage opportunity for smart marketers.

But what if at least one of the core assumptions — that online ads motivate consumer behavior — is wrong, or maybe just half-right?

The beauty of a measurable medium, including mediums where transactions are measured, is that prices become appropriate.  The reason we have banner blindness is that people just consume more pages, seeing more ads.  Many pages have literally a dozen ads on the page.  Those ads are, in many instances, virtually worthless and will be priced accordingly.  6.5 billion pages on MySpace every day: Mostly not valuable.  But that is ok, things will be priced accordingly.

Then SAI tries to question the conclusion:

Does it hold water? Certainly banner advertising had been hit by the perception that it is roundly ignored, but a raft of research concludes that video advertising — even annoying pre-roll ads — has some of the highest recall in the business.

What does this have to do?  Nielsen says TV works because it is entertaining and you are focused on it…. just like pre-roll.  Pre-roll is different than banner ads.  Pre-roll is annoying for the same reason commercials are annoying:  I don’t want to watch them and I can’t ignore them.

Post on Silicon Alley Insider that made me actually click through for the comments: Ad Sales Reps, supposedly scarce, are a dime a dozen.

My immediate reaction to this was a) Great, you can hire the worst sales people from Yahoo and AOL, no problem.  That’s encouraging.  b) Is the guy that was having success (or struggling, assuming he is not the top guy) selling Yahoo’s home page going to be the guy that monetizing neverheardofit.com?  The brand difference is pretty big.

The first sales guy I ever hired worked at Netscape and went from selling Netscape to selling our neverheardofit product.  He failed miserably.  He was used to people taking his calls because he was from Netscape.  He was used to having good lead flow because he was from Netscape.  People called him.  It was a different kind of sale.

You have to have sales people that are appropriate to the life-cycle of your business.

Article by the New York Times saying that companies gather a lot of data about you.

Big shocker, albeit this is fairly mildly exaggerated.

What it appears to me that comScore measured was the frequency they see users on their properties.  Does Yahoo! really get interesting data from people refreshing their mail again?  MySpace “has the opportunity to gather” a lot of data, but does refreshing your page every 60 seconds for 3 hours really mean a lot?

They bash Conde Nast for not gathering a lot of data, but the chances they have to gather data are probably far more contextually interesting than basically anything that anyone does on MySpace.

In fact, I would contend that a big driver of the Beacon program that Facebook introduced was an attempt by Facebook to acquire information that was actually valuable from a targeting perspective.

Yahoo can tell that I am extremely interested (high frequency) in emails my wife sends me (granularity).  Good luck selling that data.

AOL acquires Bebo.  I know nothing about the deal and very little about Bebo, but I like the sound of it and I want to go on the record with that opinion.  Driving integration with AIM is a great idea and one they have already tried unsuccessfully with AimPages, but I think you have to keep pushing on this until you get it right.

Also, owning a big social network is good.  MySpace and Facebook are not owned by Google, MSN, or Yahoo.  So AOL has something they don’t have.  That is good.  And page views are growing (albeit low value views).

Realistically, I suspect that no one (maybe Google?) knows more about monetizing social inventory than Advertising.com.

We live in interesting times!

Compete.com launched a new product called Behavior Match that I read about in an advertisement cloaked as data driven article in Seeking Alpha.  (Reprinted (or originally) on their blog)
This comes dangerously close to my day job, so I want to be clear that I am just reacting to what I read in that article.  The product might be great.  Compete is certainly great.  Etc…..

Anyway, I have to say that I feel like they kind of miss the boat.  I have always felt like looking at the size of the target market and looking at the composition index can be misleading (less so if the composition index is off the charts).  So in the first example, Compete.com indicates that “Among this list, aol.com is the best advertising opportunity for Johnson and Johnson to reach young and expecting mothers”.

Is it really?  Probably, relative to a site like Yahoo, this is true.  The CPMs are likely comparable and the fact that AOL indexes higher implies that fewer impressions are wasted on poor fit.  However, if sites on the list like Myspace and Youtube have CPMs that are much lower (likely), the cost per relevant impression may be lower.

Let’s move on to the second chart.

Here, one of the stated conclusions is: “A campaign focused across many torso domains has the same reach opportunity as a larger internet property”.  (A torso domain being neither head nor tail of the Internet.)

Once again, that is a broad statement.  But what if all the people visiting hbwm.com also visit kidprintables.com.  Then you are not expanding reach, you are just uncontrollably increasing frequency.  You may run a campaign on all these torso sites and only reach 50,000 people.  Furthermore, you probably paid a premium price.

What is interesting about this is that much of this new product is seemingly available in comScore today.  In fact, comScore offers reach penetration reports that will show how many of the hbwm.com people also visited kidprintables.com.  Where comScore falls down, and a service that would have huge value to advertisers, is that their tool is overly simplistic.  I want to create bundles of sites and compare their reach in certain demographics and the aggregate composition index.

A tool like this would allow advertisers to make smarter buys and enhance our understanding of how a target market uses the network.

So everyone in the world is reporting that Facebook had an all-hands where they talked about 07 results. 2007 revenue: $150m.

I went on comScore and tried to figure out total page views. According to comScore (frequently cited as under-reporting): ~167b page views.

So the effective RPM of Facebook pages is theoretically: ~$0.89?

In the big scheme of things, that is neither here nor there. What I find really interesting is this other TechCrunch post: Lookery guarantees 12.5 cent CPMs for Facebook apps

That implies that this ad network is doing a much poorer job than Facebook of monetizing relatively similar inventory. Yow! Obviously, the next step is to drill down to the performance of the inventory, but I am bound by a variety of contractual pieces of paper to not discuss my findings here!

Update: More celebrity sightings!  Scott Rafer critiques my critique in the comments and then I agree with everything he says, which don’t deny my earlier points.  Definitely click through your RSS feed and read all about it.  Alas, he didn’t create a Lookery chart on Cogmap, so I had to.

It is always cool when people, whose blog you read, suddenly read yours for a moment.