Cogblog

True or False: Current usage of view-through conversions is not a success metric customers benefit from, but rather a pricing mechanism by which agencies are able to imply a correlation with performance while networks are able to use their reach to spend large budgets, pleasing all parties except the advertiser.

Discuss.

“Today, in response to feedback from many of you who run branding campaigns, we’re announcing a new filter that allows you to show your ads only on AdSense sites among the 1000 largest on the web, as defined by DoubleClick Ad Planner. This new feature will ensure that your ads reach a large number of users, but only on well-known sites best suited for branding goals.”

- Google Inside Adwords blog post this week

Yeah, Ad Networks sell the same way all the time.  comScore top 100, top 500, top 1000.  It is kind of amazing that Google has not released this feature previously, but I am sure that what held them up was that the Product Manager responsible kept de-prioritizing this because it is, if not “evil”, a kind of silly feature that generally serves to mislead people that are not paying particularly close attention.

Perpetually, this is positioned by the sales force as “branding inventory”.  It is implied that the largest web sites in the world are generating inventory ideally suited for large brands to advertise on.  That is just plain silly.

The amount of page views a web site generates has absolutely nothing to do with the relative brand safety or value of brand association that a web site has.  Let’s spend 30 seconds looking at this list of top 1000 web sites:

• Facebook
• Blogspot
• MySpace
• Photobucket
• Orkut
• WordPress
• Blogger
• Partypoker
• Megauploader
• Imageshack

Those are just some of the non-Chinese sites in the top 60 (so maybe 1/2 of the english language content available) and let me tell you:

• These sites are not just predominantly UGC content, they feature tons of NSFW content.  (Unlike, say, LinkedIn, which, while on the list and primarily UGC, is mostly brand-safe.)
• Further, the inventory that Google is getting from some of these guys (e.g. MySpace) will definitely tend toward the distinctly less brand-safe and NSFW.  MySpace is probably not sending their best impressions to Google.

Popularity of the web site is not even loosely correlated with brand safety.  The strategy for this stuff is simple.  For most ad networks (and this is probably still true for Google, although maybe slightly less so), most of their ad inventory came from these sites anyways.  After all, these sites account for the majority of impressions on the Internet.

This “Top X” assurance comforts media planners in some way while only cutting out 10% – 30% of the inventory the network had for delivery.  The result is plenty of room to optimize for the network and the deal is closed, so a victory all the way around.  Buying “Top X” inventory is a way to pay a premium for the inventory you were probably going to get anyway without significant brand protections for advertisers.

People used to say that there were 400 ad networks out there in their hey-day.  I think there will be an order of magnitude more DSPs.  Here is why:

Building a DSP is easy.

Ad networks were vertically integrated examples of our marketplace.  To build an ad network, you had to build an exchange, sign up publishers, and build a DSP.  The market has evolved to a point now where getting dramatic reach and scale at a level unimaginable 5 years ago just takes 10 lines of PHP (so that is like one line of Python?) with Right Media’s PHP library:

 
 
 
 
 
 
 
 
// get params from command line
list(, $SOAP_BASE, $ADV_USERNAME, $ADV_PASSWORD,
       $ADV_ID, $LI_ID) = $_SERVER['argv'];

// get handles on Contact and Publisher services
$contact_client = new SoapClient($SOAP_BASE . 'contact.php?wsdl');
$campaign_client = new SoapClient($SOAP_BASE . 'campaign.php?wsdl');
$target_client = new SoapClient($SOAP_BASE . 'target_profile.php?wsdl');

// login and get auth token to be used later for other API calls
$token = $contact_client->login($ADV_USERNAME, $ADV_PASSWORD);

// create campaign data holder
$campaign = new stdClass();

// set advertiser entitity id and description fields
$campaign->advertiser_entity_id = $ADV_ID;
$campaign->description = 'example campaign';

// create a new campaign based on campaign object defined above
$campaign_id = $campaign_client->add($token, $campaign);

// link the campaing to the line item
$campaign_client->addLineItem($token, $campaign_id, $LI_ID);

Fearsome.

I used to tell people all the time that starting an ad network is the easiest thing someone can do: Get some inventory, call 20 agencies and tell them you have a new algorithm to drive performance, and they each give you a $20k test budget!  Voila, you did $400k in your first quarter.  Agencies felt pressure to find test budgets for everybody because, if a client were to ask them “What do you think of X”, you can’t say, “Well, we never tried X”.  Agencies felt an almost fiduciary responsibility to try new stuff.

Now, if you didn’t perform, the next chunk of dollars was tougher, but you had runway instantly.

We are seeing the exact same dynamic in DSPs today.  If you mix the data and inventory a little differently (and it would be hard not to), voila, you are worthy of a test.  Agencies are playing the field today, the great rollup of DSPs that everyone is so looking forward to has not yet happened, and agencies expect to and are prepared to try lots of different things.  All you have to do is perform after that and you have a business.  If you eat your margins early on, offer layered in retargeting, etc., the odds that you can artificially inflate performance in a way that makes your business look interesting is high and this gives you more runway to work.  Convert 25% of your test budgets to $200k renewals and you have a Q2 business doing $1m in revenue.

Building a DSP is cake.  Locking in data or inventory or building an algorithm that creates great performance for advertisers over time by arbitraging data and inventory is what will separate the winners from the losers, but it will be non-obvious in the first 6 months of working together who those guys are for agencies.  Remember when Glam spent millions of dollars of VC money buying inventory at a loss to lock in exclusive access to inventory, then when they had the advertiser base, they crushed payouts to pubs? There are a lot of the same kinds of problems that get slathered over early on in this market. Building a DSP that can look at 10 billion impressions a day vs. 1 billion cost-efficiently is interesting, but no one will need that scale for a year, so no one will know who can do it better/faster/cheaper.

I worked at Ad.com and I am not gonna lie, I took away from that place a sense that algorithms are hard, you need tons of data to figure out how to improve them, and there are few shortcuts other than “been there, done that”. Unlikely that any small company has a better algorithm today. You need that kind of algorithmic skillz. (And of course, Google > Ad.com > other networks) And even Ad.com would say that their algorithm has tons of room for improvement. Ask me, I know ten. But even algorithmic improvement has trade-offs: You can factor in more data to improve results, but that requires more data points to test which requires larger test budgets. Bummer. To limit testing, you need to limit the factors you evaluate. The result is plain vanilla.

The market for DSPs is white-hot, expect 4,000 of them, but most of them will be frauds.  Real differentiation and competitive advantage in a space where virtually everyone has access to the same inputs is hard.  Remember the Netflix competition.  Given a data set (100 million data points), improving predictive results is incredibly hard.  With a million dollars on the line, it took 3 years and dozens of people to generate a 10% improvement.

You heard it here first.  Breaking news as it happens.

http://www.flickr.com/photos/nickbilton/2709378784/

The Advertising Network business is dramatically changing due to massive market fragmentation.  It used to be that ad networks were “the way” that people acquired significant reach and met performance targets in campaigns.  Publishers chopped up all of their spot market inventory to dole out to a variety of ad networks and then the networks packaged and resold that inventory in a variety of reach/brand or performance bundles.

I have spent a lot of time thinking about how things will work out for Advertising.com.  After having worked there for many years and recently leaving, I wondered, “what will become of Ad.com?”  I have reached some conclusions that signal to me a dramatic market shift that will rock every ad network and result in the dissolution of most of the advertising networks we see today.  I expect the number of advertising networks to fall precipitously (less than 40?) in the next few years and the networks that survive will be far smaller.  The spot market is going to work a lot differently in the future: inventory will be acquired differently and the inventory will be sold differently.

First, the inventory that ad networks acquire today will be bought via exchanges tomorrow.  Publishers use yield optimizers to manage daisy chains today, but simple daisy chain management can never realize the true value of inventory to individual bidders without a real-time auction of the impression across the entire set of interested parties.  The result is that a few yield managers are slowly turning themselves into RTB exchanges (Admeld, etc.), while other people buying spot inventory directly from publishers will probably see that inventory get sourced to exchanges instead.  Why would a publisher not maximize the availability of that impression to advertisers?  Limiting an impressions availability to a single network or group of networks artificially limits the yield opportunity to publishers.  Exchanges, a technology platform with a minimal transaction fee, will be the primary means by which spot market impressions are liquidated.

Why did no advertising networks seize the opportunity to transform into an exchange?  ContextWeb did.  Alas, exchanges are kind of a winner-take-all market – a natural monopoly.  Why not Ad.com?  Probably margin compression.  Exchanges are a technology platform model.  They are supposed to run on razor thin margins, keep the team relatively lean, and make it up on volume.  The largest ad networks had the benefit of scale, but the margins were much more significant due to the value they added in the process: Sales and publisher support far above the call of duty for an exchange marketplace.  Exchanges are wild places: The ignorant don’t have their hands held, they are killed and fed to the crazier animals.  Anyone remember all the awards that Ad.com got for customer service?  While Ad.com had the biggest network and incredible optimization technology, they were, fundamentally, a service model.  High touch.

So advertising networks are going to have to get their inventory from the same place everyone else gets their inventory from: The exchanges.  They will rapidly (extremely rapidly) become disintermediated from publishers.  Who do they work for?

Not big agencies.  Big agencies are licensing DSP technology to manage their own spot market buying interfaces. The value that networks previously provided of tremendous reach at low CPMs is no longer their solitary purview.  DSPs are a technology platform solution that enables buying into exchanges in the same fashion with a lower margin architecture.  The overhead of sales and support services provided by ad networks is superfluous for Madison Avenue.  For years now, Madison Avenue has been confident that they could vertically integrate into networks and claw back margin.  Now they are being offered products that enable that.

Why aren’t ad networks becoming DSPs?  Two reasons, I think.  First, the same margin problem.  This is all about startups disintermediating the ad network market by taking the 30% or 40% margins that networks made and cutting it into several products that run at 10% margins and serve broader markets.  If you can capture all of Havas’ future online media spend, you have a nice business.  Second, this is about building a real technology product.  Many ad networks had minimal technology infrastructure and were primarily sales and service businesses.  The networks that had technology did not have products.  Ask anyone that has commercialized a product.  The last 10% of the product that you have to build to make the product usable outside your organization is just as much work as the initial 90%.  It is tough.  And it is a core competency not necessarily found in these organizations.  It is something different.

So who uses ad networks?  Small agencies that don’t license DSPs may use ad networks as a buying interface, but even this will be slowly picked off as some of the DSPs that fail to get traction on Madison Avenue unveil their long tail solution or Google meets their needs.

The real customer is direct response advertisers.  Direct response advertisers will still want to buy on a CPA and will be looking for an intermediary willing to take on the risk of marketing their products.  This is an area where skilled technicians can add a lot of value.  Sales people that recognize what CPA deals might work and what is effective CPA pricing to generate the margin needed to drive substantial volume for a DR advertiser will be critical.  More critical, however, is this: The algorithm.  He who has the mightiest algorithm wins.  Now there will be lots of different ways to arbitrage exchanges in the best interest of DR advertisers, so there is plenty of room for a number of networks and the market will certainly evolve substantially, but it all rests on the algorithm of the ad networks.  Each network will go get data and feed it into their system.  Their ability to use that data will demonstrate their utility or lack thereof to potential advertising partners.

What value does a Tribal Fusion add in a world like this?  How much longer can “retargeting networks” justify generating huge margins for cherry-picking exchanges?

This is a great market – and a growing market – but I suspect that we will end up with far fewer and much smaller ad networks.  Margins will be slimmer.  But a few algorithms will rise up, using cheap sources of data that facilitate this slim margin business, that yield such benefits to direct response advertisers that the networks that have designed these algorithms will be sustainable businesses.

In a decade, we will look back on the 2000-2009 period and say, “It is amazing that an entire industry came and went.”

I have noticed recently that tons of people that have built interesting tools to help people create ads: ClearSpring, AdReady, PaperG, and Jivox among others, for some reason spun that into creating ad platforms of some sort.  That made me wonder:

Why?

Those are actually pretty different core competencies and there is a lot to do when it comes to building great creative building tools.  Are there not opportunities to license this technology?  Is anyone so extraordinarily focused on solving this problem that they don’t think about anything else?  When I look at what AdReady built several years ago, it has not really evolved.  Is “good enough” acceptable?  Why can’t anyone focus on being best-of-breed in a few key areas any more?  It seems like that would be a good idea.

Maybe the market is not good enough.  People like Sprout, KickApps, and Gigya that started out in this space have since moved on to other things.  But Pointroll got rich and sold out doing this.  Medialets started out as a network and has actually decided to focus just on making great creatives.

The approach of people that want to do a lot of things and have this as a “differentiator”, but not the absolute focus of their business sound like people that are unlikely to win.  Focus, focus, focus.  EOM.

http://lifeinthenhs.files.wordpress.com/2009/02/facebook.jpg

The blogosphere is abuzz, errrr, atwitter, err, atalking (“TALKING – A NEW TOOL FROM MICROSOFT – TALKING IS THE BEST WAY TO TALK TO YOUR FRIENDS!  START TALKING TODAY!  IT IS SO EASY TO START TALKING, A TWO YEAR OLD CAN DO IT!”) about how Facebook wasn’t making any money from Ad Networks and their proprietary self-service ad platform is making all the money.  While I am a believer in self-service advertising………..

I have to say that checking out the trees in this forest is fairly instructive.

First, this data is probably wrong.  Inside Facebook revises the run rate from $150m to $50m because it turns out that the $150m includes Microsoft’s cut.  Is Microsoft really paying out a 33% rev share?  Unlikely.  It is far more likely that Facebook keeps at least double that.

Second, Facebook is courting hard-core DR guys on their self-service platform.  The ads they run are ads that Microsoft would never approve.  Those ads perform better because they are designed to.  Facebook is doing things they would not allow Microsoft to do: run brand-degrading ads that are probably illegal all over their site.  This renders higher CPMs.  True story.

Third and more important, Facebook wants them to fail.  Yeah, you heard me right.  Facebook is the reason why banner ads fail on Facebook.  Why would their self-service ads succeed when their banners fail?  Is there some magical structural difference?  Nope.  The difference is this: When you buy a banner ad impression on Facebook from Microsoft, you get exactly that: 1 impression on Facebook.  When you use the Self-Service platform, you get something a little different: A CPC campaign targeted at a man in Wisconsin who likes cheese and the green bay packers.  Exactly what you wanted.  Facebook is not passing data to advertisers on their banner network so their impressions look like the worst ads on the Internet: non-contextual, brand-unsafe, high-frequency UGC inventory.  The ad that shows there is a speculative guess with respect to performance.  The advertising version of throwing stuff at the wall and seeing if something sticks.  Advertisers have no framework for segmenting the inventory so optimizing it is impossible.  It performs like crap because it is crap.  Microsoft can’t sell that on a CPC.  Without data layered in, it performs too poorly to make money on a CPC.

Facebook is selling advertisers on the self-service platform highly contextualized, highly targeted inventory on a performance basis.  Advertisers like that.  Advertisers want to buy that.  Not high-frequency, nameless, faceless impressions.  End of story.

I was reading Zach Coelius’s article defining DSPs the other day and it made me wonder, why does it seem like new companies, generally, are winning the DSP wars?  For that matter, how did no one jump into the space now dominated by Pubmatic, Rubicon, and AdMeld?

• Zach says that DSPs should have access to billions of impressions per day.  Ad.com had that years ago.  Valueclick was big.  Right Media, theoretically, could have tried to spin around and be a DSP.  One wonders how they let so many new companies enter this gap.
• Zach emphasizes reporting, impression attribution and other features as key aspects of a DSP.  Traditional third-party ad servers are all perceived as excelling in this area.

Certainly, there are features that they did not have, but it seems like extending the product would be easier than building a new product.  I can point to several things that probably made transitioning to this new exchange/DSP world difficult:

• The quarter-to-quarter pressure of publicly traded companies.  The ability to invest to meet new market technologies is frequently limited for large companies.  If you look at exchanges, Right Media was a well-funded private company.  DoubleClick was going nowhere until it went private.  Then it was able to escape some of the quarter to quarter pressure and before you know it, they used Project Wolf to sell themselves at a huge mark-up to Google.
• Legacy architecture.  Sometimes it is simply easier to build from scratch than to build with a legacy code base.  Particularly with the growth of new abstractions for rapidly developing technology such as Ruby on Rails.
• Margin and other business structure limitations.  This is the real killer.  There was a certain expectation for how the relationship worked and breaking new ground was hard.  When people have revenue and profits, sacrificing those if you want to change the structure of the industry is hard.  This is classic Innovator’s Dilemma kind of stuff.  Making an advertiser or publisher think of you differently if they are already a substantial source of revenue is a big risk.

How did the big ad companies of the first half of this century sleep on these new opportunities in the market?  Give me your thoughts.

I love Junior Hines blog.  He rarely lets on what he is thinking, but the drift you get from reading the things he finds interesting tell you what a star he is.  But I absolutely love this statement:

Putting banner ads on a social networking site was more or less bolting on a scalable advertising system onto a site with massive traffic, regardless of the fit with the product experience.

While I think, at some level, this is certainly true, I also think that Facebook never really tried to make it work.  The reason that their ad platform is better is because they are using their data, but they would never share that data with third parties.  If you wanted relevant ads that performed well on your site, why never share?  For the ad networks they worked with, every impression was the 500th ad some guy has seen on the site and he hasn’t clicked on anything yet and we aren’t optimistic about the future.  Of course the ads they have to show are going to perform poorly.

Just saying…

http://www.flickr.com/photos/casalemedia/3471980910/

Brian Tomasette continues to have some nice posts on his blog.  His most recent discusses how it is possible to wire together DFP to work like an ad network technology stack.  In fact, he points out, Collective Media uses it this way. (or used it this way.)

Couple of thoughts:

1. Love that Brian is willing to put it all out there like this.  Maybe he doesn’t view it this way, but I view this as a pretty good knock on Collective Media (more on that in a second).  When I was an AOL employee, I tried to be pretty restrained in bashing competitors (although I am sure some would disagree).  It is very tough for Brian to be a sales guy at AOL and make fun of Collective.  But telling the truth is, at some level, never wrong…. it is just not PC.  And really interesting!
2. Lot’s of ad networks are doing this.  I mean a lot.  If you and I started an ad network tomorrow, we would probably do this.  If Brian started a network tomorrow, he would do this.  Great optimization only comes if you have the liquidity to have things to optimize off of.  So the focus of every small network is signing up publishers and advertisers.  It is rarely the algorithm.  The result is that people get just enough technology to get by.  Essentially, you fix your technology costs as a percentage of revenue when you use this.  It isn’t necessarily great tech, but it would take a few people a bit of time to do better.  Most people say, “why bother?”  One of the main reasons that DMX was shut down by Right Media, in my opinion, is that people were simply using it to run tiny ad networks and, given that most of the inventory was just more frequency from the same publishers and more CPA ads from the same advertisers, they were not adding a lot of value.  This is the “not premium” that Yahoo was complaining about.
3. The advent of exchanges has made this even more popular.  Now we don’t need to front the inventory to run a tiny ad network and we barely need to optimize to go out and sell retargeting.  All you need is an account on DoubleClick or Right Media’s exchange and you can start selling network services.  But all this inventory is the same!  Zero value has been added, it is just more sales people selling the same inventory.  And CPA advertisers don’t particularly care, although as they allow varying frequencies of inventory to be inundated with tiny advertisers, they lower the eCPA for the big guys, potentially hamstringing campaigns.  This is why the really big guys are a little careful about this stuff.

When a new network calls you, do some diligence.  Take their tags.  Look at ‘em.  Are they Right Media tags with a cname?  Do a traceroute.  Look at cookies they set.  This isn’t hard and doesn’t take a lot of time.

There are tons of people starting ad networks.  Generally, those people aren’t engineers, they are ad network sales people.  Are they really building something unique or are they cherry picking inventory off of exchanges?

UPDATE: Let us be clear, I mis-characterize Brian as making fun of Collective in this post.  All he actually does is state the fact that at one time they used a third party’s tech stack for optimization technology without rendering any opinion regarding the decision to do that.  As Brian remarks in the comments, he respects Collective Media.

Lot’s of people, and AOL is a well-documented party here, are trying to figure out how to cut the number of ad views on a page and increase the CPM they charge advertisers while not drastically reducing their page CPM.  Similarly, decreasing low quality “picture galleries” or breaking articles into numerous pages, can decrease “available pages” while increasing the quality of the product being sold.

Unfortunately, that is incredibly effective and easy to execute in straightforward direct sales, but has relatively low perceived value in the world of exchanges.  In exchanges, where people are typically buying audiences, one impression feels as good as another to most advertisers.  When Fetchback wades into an exchange bidding $2.00 CPMs and charging their advertisers $6.00 CPMs, it doesn’t matter if there are 8 ads on the page or two ads on the page.  The biggest challenge in a retargeting campaign is acquiring volume, so they are probably happy to get it any way they can.  Generally, this ends up being true for lots of advertisers seeking audiences.

Further, advertisers have very little data about aggregate ad frequency per user (is this the hundredth ad they have seen on the site, on the exchange, or on the network today?)  That data would help in the valuation process and is notoriously absent.  Of course, as we discussed, only the most sophisticated advertisers would be able to value it.

Finally, were one to be one of those sophisticated advertisers, it would probably be all about the performance.  Determining lift in performance for pages as the ads on them change is critical to effectively managing an exercise to decrease aggregate ad views across a web site.  Split testing is important and I can tell you from experience that it needs to be a real split test.  Simply changing and comparing a past period to the present exposes the test to too much variability of advertising supply.

Most publishers will struggle to justify the resources to conduct an exercise like this.  Particularly when, unfortunately, given the simplistic perspective of many advertisers, there are reasonable odds that this will not move the needle at all in increasing the true net value of their sites inventory.

While it is in everyone’s best interest (theoretically) to have fewer, better performing ads on a page, in an exchange where little is known about any given placement, a bad actor can exploit good actors in the system to unfairly maximize his yield at the expense of other players.  This results in a prisoners dilemma situation.  The result is that, in many tests, publishers may find themselves in a death spiral of adding more ads to inventory to increase the effective yield of a page.

It will be interesting to see how we evolve the Internet to be a good place for advertisers and readers.