Cogblog

I find it creepy.  It is a good thing I don’t follow New York Times reporters on Twitter or you might find that splashed across a major paper today.  These days I feel like I see completely incongruous ads showing on inventory and I know they tracked me there.  (I have not read the Times article or any other related articles, for what its worth.)

(This is “LeadBack” or retargeting, as explained on the Ad.com Japan site today.  Notice how the cookie looks like a golden coin!)

As the story was told to me, John Ferber was actually the person that invented retargeting.  He had a bolt of lightning in his brain one day that made him see how the technology that was being used for analytics and conversion pixels could be applied to deliver ads to people after they had visited a website and he named this “Advertiser LeadBack”.  (Advertiser LeadBack is a designated trademark of Advertising.com, a wholly owned subsidiary of AOL)  Advertiser LeadBack was a phenomenally successful product.  So successful, in fact, that advertisers would ask other companies for “LeadBack”.  Ad.com felt like they were the Coca-Cola of the trade, the brand identity of their retargeting solution was so popular and well-known.

One battle we continually fought at Advertising.com was the battle to cut down the number of retargeting campaigns in a buy.  (I thought I had posted the white paper I had written on this subject to the blog, but I cannot seem to find it.)  Essentially, our thesis was this:

• Advertisers wanted to spend more on retargeting because it performed so well, but they still frequency capped campaigns because everyone agreed that high frequency was not improving performance.
• Ad.com had the biggest network from a reach perspective.
• Most of the other networks were buying the same inventory to get reach, just different frequencies.
• Ipso facto, buying a campaign on another ad network was typically simply increasing the frequency to a user rather than reaching new users.
• Buying retargeting campaigns from other companies was wasteful.

As it turns out, every time we won that battle, I now think there was another effect: We decreased creepiness online.

I find it creepy.  No lie, it is totally incongruous to see some of the ads that I see on some of the sites that I see them on.  We talked about creepiness all the time at Advertising.com.  Ho Shin’s (Ad.com General Counsel) favorite example was “If I visit a Korean site and then visit the Washington Post, we can’t show an ad in Korean.  It would freak people out!”  While the ads I see are all in english, I get the same effect regularly today.

I attribute four things to the rise in the creepiness factor:

• Small advertisers having access to retargeting technology.  At Advertising.com, we had huge minimums: $25/k month.  Only the biggest advertisers in the world advertised with us.  If you advertised with us, you were spending an absolute bucket of money advertising online.  Only a tiny sliver of your budget was being devoted to retargeting.  The result was that people did see these ads all over the place and most of them weren’t retargeted.  Further, people had become accustomed to the idea that big advertisers ads would appear in random places as part of a network buy.  So if you went to a big web site and then saw an ad for that website somewhere, you could write it off to big marketing campaigns.  Not too creepy.  With the influx of small advertisers, now people see ads all the time that they know had to be targeted to them.  These small advertisers are not buying ad space on all these sites!  They are only here because I am here.  Creepy!
• Customized creatives.  It is one thing to see a Best Buy ad.  Best Buy is a big brand, they spend a lot on marketing.  It is another to see the TV I was just looking at being shown to me again.  Creepy.
• The dramatic increase in retargetable inventory.  When most people bought most of their retargeting via Advertising.com, while we had huge reach, we actually worked very hard to limit the amount of frequency that we bought.  This was because typically we didn’t have that many ads to show any given person.  (As I said, we had high minimums which limited the number of advertisers we worked with at any given point in time.)  The result is that you would only see a couple of ads.  To extend the Best Buy metaphor, you would go to twenty or thirty web sites and you would see a specific ad that had been retargeted to you on one or two.  The campaigns were frequency capped and we simply didn’t have that much inventory.  A few billion impressions per day.  Now, with the advent of exchanges, people can peak at 10 billion plus impressions every day and people are much looser with frequency caps because the ROI continues to be strong.  The result is that you might see these retargeted ads on 10 or 15 of these web sites.  Even the biggest advertisers probably aren’t buying all the inventory directly!  They must be following you.
• My final point is, in some ways, simply an extension of the above: Quality of inventory is declining.  I don’t mean quality in an absolute sense, but in that abstract sense that advertisers tend to think of quality: Big Brands.  As long tail inventory gets exposed to exchanges, it just feels less like an advertiser bought the space and more like the advertiser bought you.  Best Buy inventory appearing in Yahoo! Mail?  But of course!  Best Buy ads appearing on random blog about punk rock?  Bizarre.  Creepy.

I recognize there is no going back, but to deny the creepiness is a lie.  We always discussed how it would be “best” if people felt like these ads were not following them at all.  The point of these ads, from a squishy brand perspective, is that when a consumer sees these ads, they bolster the credibility of the advertiser – “Wow, that web site is big enough to be advertising on some of my favorite sites, I trust that web site more.”

It would be interesting to see someone A/B test some of these hypotheses to determine their impact on the value of retargeting campaigns.  Get to work!

I have been gone from AOL for almost a year now so I know less than nothing about what is going on in the world, but I wanted to throw out a random opinion or two on these rumors.  I have already given my opinion on the future of Ad Networks, so, no surprise, my opinion ties in closely with what I see going on out in the market.  Here are my theses:

1. Ad Networks will acquire most of their inventory by bidding into exchanges
2. Algorithms and access to exchanges will rule the day

Given the lack of announcements by AOL about the status of RTB bidding activity and anecdotal evidence regarding the state of the engineering team and the current focus of the engineering team, one could reasonably postulate that AOL has not been able to invest like they would have liked in scaling their bid engines to look at billions of transactions per day.

This implies that acquiring AppNexus, which has the bidding engine scale and is one of the premier intermediaries between exchanges and advertisers, might be a great tech fit.

AppNexus is a fairly unique company that is laser focused on building better technology than everyone else.  I am a big believer that whoever doesn’t acquire them will kick themselves for missing that opportunity.

So are they interested?  Fo’ sho.

Could it happen?  Seems unlikely.  For a company that has raised $15 million and is in the fairly unique position that they are in, they probably won’t want to sell for less than $100 million and it will be difficult, given the debt convenants AOL has, to do a larger deal.

Of course, this blog post could probably have been compressed to a tweet for all I know: AOL probably wants to buy AppNexus, as do many, but it seems unlikely.

So I went to SXSW for the first time in 2009 and I had a great time.  (I didn’t go in 2010 because I was in San Francisco the entire previous week and my startup travel budget was blown.)  I tell everyone the same thing: At SXSW, because it is out in the middle of Texas, everyone is a visitor, no one is running home to their family, and the result is that everyone is totally open to meeting people.   You can walk up to anyone and say, “What is your story?” and they will totally engage with you.  It is a great way to meet new people, network, and build interesting relationships.

Sounds great, right?  But here was the problem: I could not find my people.  I ended up saying, “What’s your story?” and learning tons about crazy obscure movie directors and probably 124025820628 graphic designers.  I wanted to meet my people.  My online advertising people.  But I was unsure how.  There were exactly 0 panels or events for online advertising.  I ended up hanging out with Greg Yardley most of the time and while I thought that was great, he can testify to how annoying spending an extended amount of time with me is.

I have become convinced that what the world needs is an advertising panel at SXSW to provide a forum for online advertising geeks to find each other and commune at this esteemed event.  Thanks to the power of panelpicker, that can now happen.  Selecting speakers for SXSW is partly driven by user voting, so now is your chance to vote for next year:

Click here to go vote for my panel (Registration Required):

http://bit.ly/aq1llV

I know registering is a drag (or as we like to say, conversion rates drop precipitously), but think about it, every other panel has to get people to register, so if you just put yourself out for 2 minutes, we can have a higher conversion rate than everybody else.  That translates to winning!  This is going to become THE EVENT for online advertisers at SXSW.  We are going to be uber-connecting people.  Help us make it happen.

They are the people that price CPA deals.

A lot of them have fancy titles like “Director of Yield Management”, some have lowly titles like “Manager of Ad Operations”, but at Advertising.com, they were known simply as Delivery Managers.  Every ad network has them and they form the lifeblood of most yield optimizing organizations.

I suspect that, simply given the age of the organization, Advertising.com had the first delivery managers.  This is the story I was always told about how it was created:

When Gar Richlin joined the organization, many CPA deals were essentially being negotiated by the sales organization – a group that did not have the in-depth knowledge of the network required to understand what reasonable price points and volume correlations might look like.  This is critical in an ad network because a low CPA, while appealing initially to an advertiser, might turn out to drive limited volume because a low eCPM might prevent the campaign from getting widespread network distribution.  Working hard up front to negotiate a fair CPA would maximize the volume of conversions an advertiser receives, making happy advertisers and happy networks.  Gar created a ninja team of the best and brightest half-dozen or so people in the organization.  They were given ultimate power in accepting or turning down CPA campaigns from advertisers.  They tended to be senior people that had the respect of the sales organization, but they were not sales people.

This was a key inflection point in the organization.  Suddenly customers were given reasons to pay a higher CPA.  Delivery managers would come to sales people and say, “If customer X would increase their CPA by Y, we think we could generate Z more conversions for them per day.”  That moved the needle for customers.  As I said, many, many times at Ad.com, “If a sales guy runs into his bosses office and says, ‘If we pay $11.00 instead of $12.00 per sale, we will get 200 more sales today’, I think the head of sales does that deal every time.  And the next day, if he runs in again and says, ‘Paying $13.00 instead of $12.00 will generate 200 more sales today.’, the head of sales does that too.  Right out to the absolute margin of profitability.”

Delivery managers are the people that take on the risk in the organization.  Their willingness to take responsibility for making a CPA campaign work makes or breaks the success of an organization.  Advertising.com used to have a guy, Rich Morrissey, that was the absolute guru of pricing educational campaigns.  He had done so many of these deals and managed so many campaigns that he knew where we would get volume at different price points, what effect different creatives and landing pages would have on conversion rates, and what fair prices were.  He priced all these deals and we used to joke that if he were hit by a bus, it would cost us hundreds of millions of dollars.

The in-depth knowledge of how the network works is what allows these people to do their jobs.  What is a good frequency cap for campaign X?  What is a reasonable starting CPM to bid if we get paid on a CPA and how much can we spend before we have to stop or change it?

A delivery manager that knows the answers to these questions are worth their weight in gold.  On the one hand, they are developing such an esoteric, specialized skill set that they are virtually unqualified for other jobs.  They have a resume with one programming language on it: Ad Networkian.  On the other hand, the demand for these people is exploding: Inside agency trading desks, DSPs, Exchanges, Ad Networks, Publishers (up to and including Facebook), and many more all need these skill sets.  In a story that rocked Baltimore, Tribal Fusion opened an office in Baltimore just to hire Advertising.com delivery managers.

I am predicting now: If the people you have doing your delivery management are people that you treat like “Ad Traffickers”, they will find better homes that value them more appropriately in the near future.

What do you call your delivery managers?  How did that position come to exist in your organization?  Share your story now!

Was talking with one of the smartest people around the other day and he said to me, “Yeah, but what happens when the government stops subsidizing all of these online universities?”  And I said something to the effect of, “What?”

So then I did some research.  Lo and behold, he is right.  (No surprise there.)  The federal government is loaning 25% of for-profit students $30k each to attend school and then 40% default on their loans.  So it seems reasonable to imagine that the government will cut these guys off at some point.  And these guys are big spenders in online advertising.  A conversion is worth a ton to them.  They pay some of the highest CPAs in the industry.

Other is the online universities bucket in this table.  As you can see, they comprise approximately 10% of online advertising spend today and are expected to be (according to Forrester), one of the fastest growing segments of the market.

Is this sustainable?  As we have already demonstrated, there are people that know way more about this stuff than me.  Maybe Jay or someone can weigh in?

Second price auctions seem like they are all the rage.  One of the challenges in second price auctions is that a bid could theoretically be very high, yet the payout actually be very low if the gap between a bid and the second bid is significant.

One situation where this could be problematic is if the publisher in an ad auction requires a minimum payout higher than the second price but lower than the highest bid.  Theoretically, the impression could be lost by paying the second price in a situation where the bidder has clearly indicated his willingness to pay more for the impression.

This is increasingly critical as audience targeting and new retargeting techniques become more common because the spread between the sparse high performing data points and the “other impressions” is growing rapidly.  Media buying efficiencies based on techniques such as RTB are resulting a world of have and have-not impressions.  So how do publishers and advertisers find equilibrium?

Google has overcome this by allowing publishers to introduce a floor that acts as an artificial bid.  If a floor was $1.00, but the second bid was less than $1.00, the price paid by the winning bidder becomes $1.00, meeting the publishers minimum expectation while delivering the impression that the winner wanted to buy.  This sounds great!

There is a problem though.  Because the bids and asks (unlike in many electronic financial exchanges such as Archipelago) are closed, publishers increasingly feel pressured to use floors to daisy chain impressions, manipulating bidding to maximize revenue.  Floors are used something like this:

• Suspecting a situation with significant disequilibrium between first and second bids, a publisher might set a floor of $5.00.  When the floor is higher than the first bid, the impression is passed back to the publisher.
• The publisher than sends the impression back to the exchange, with a new $4.00 floor.
• Repeat until auction is executed.

In a world of RTB, this is a situation hard to police and frankly, one that the exchange is not super-incented to fix as it extracts higher eCPMs.

What makes this really nuts is that, as publishers create and manipulate tiers to attempt to discover bid prices, advertisers are encouraged to adjust bids to attempt to discover artificial floor prices introduced by publishers.  The result is a constant moving target by both parties, adjusting floors up and down by pennies and dimes while advertisers adjust bids and frankly, experience wholesale changes in liquidity (campaigns come and go) that throw an unpredictable wrench in all these algorithms.

I know lots of startups are thinking hard about these issues.  It is great that we live in a market with so much white space.

What is going on with international advertising?  Can anyone tell me how effective international monetization is?  I ask because I generally thought of those impressions as not particularly valuable, yet InMobi and Hi5, predominantly international advertising plays, just raised a ton of money.

Quick story: Guy at competing ad network announced several years ago that they would pay a certain amount for a certain kind of impression that I coveted.  But what was funny was that it was a flat CPM with no frequency cap and no geographic restriction.  We promptly offered more than double to publishers, but frequency capped it and only took US impressions.  However, the restrictions were in the fine print.  So I see one of the senior guys at the competitor a few weeks later and he says, “I don’t get it, how do you monetize all the international stuff!”  So we were cherry-picking all of the good impressions and then publishers would ship them all the leftover international and high-frequency impressions.

Anyway, what is state-of-the-art for monetizing international impressions?

Companies are gearing up to make an absolute fortune in behavioral targeting by scraping Google search terms off of publisher referral data.  (Some already are!)

The theory is simple: Search has credibility with advertisers as a very effective behavior with high conversion rates.  The only people with active search behavior data today is Google.  When someone searches for something and then clicks, their browsers HTTP_Referrer is set to the last page they visited, possibly something like: http://www.google.com/q=Searching+For+Cars (Where “Searching For Cars” was their query on Google).  So if a behavioral data gatherer like Magnetic is sitting on a publisher with a lot of search driven traffic (most web sites), they can gather a lot of referrer data.  Logging this allows for (As Magnetic’s web site describes it):

Magnetic™ is search re-targeting. The Magnetic data marketplace empowers advertisers and publishers to use search data as the key indicator of intent and re-target campaigns to the most relevant audience online. With more than 270 million search profiles, Magnetic significantly lifts the value of media and improves campaign performance.

A few years ago I had a similar, albeit less brilliant idea: It would be easy, in the same vein, to write a few lines of javascript that compared lists of sites to a browsers user history.  In that way, you could easily sell conquesting: Dell could show ads to everyone that had visited apple.com AT ANY TIME IN THE PAST.  All, without requiring Apple’s consent.  And I was able to get that data because I had a deal with Website X that allowed me to run tests against user history data in the browser without the user knowing.

Now, maybe a post like this will cause some start-ups to start coding away on this (no need, really.  It is only about 10 lines of javascript, then you have to ajax-ly forward it to your server.  Here is how.).  That is not my intention, obviously.  The real question is: Is that OK?  And if it isn’t, how is it different than the search retargeting we are seeing blossom in the market today.

What about other uses of that technology?  Could Microsoft offer special discounts to people that had visited Apple.com recently?   Is it OK to essentially steal that data from a user?  Should you know what competitors a potential customer is talking to?

I posit that this whole area is about as far into a potentially unethical gray area as one could go.  There is basically no difference between any of these examples and they all make me want to turn off referrer data in my browser.

What do you think?  Are you OK with every website knowing every web site you visit and query you type into a search engine?  That seems like where we are going.

Just wanted to do a quick shout-out to the first sponsor of my blog, AppNexus.  AppNexus is clearly poised for transformational brand awareness the likes of which few companies in the space have ever seen.

Thanks, Brian O’Kelley!

Reading Eric Porres post about transparency in behavioral targeting on AdExchanger reminded me that I have been meaning to crank out a post about vector targets.

If we want to really simplify behavioral targeting, a behavior can be distilled to three attributes:

1. The event: Is this a visit to Kelley Blue Book or a visit to Yahoo! Autos.  A visitor to one or the other could perform differently.
2. Recency: Was it a week ago or a month ago.  Was it an hour ago?  True story: Retargeting cookies you dropped in the last hour perform a zillion times better than anything else you are doing.  Are you watching those?  Are you segmenting them out?
3. Frequency: Did they visit twice?  Three times?  That performs differently.

Painting in extremely broad strokes here, there are a couple of kinds of behavioral bucketing strategies that are commonly used.  The simplest would be to simply describe the above behavioral attributes: “Our auto intender population is made up of people that visited AOL Autos at least 3 times in the last 60 days.”  This bucketing approach is nice in that it is quite transparent to an advertiser what constitutes a population member, however the population size is fixed, which can limit deliverability for a campaign.  We only have X number of people with those attributes.

Typically, the way a vendor might work around this is tracking populations with varying recency and frequency.  So if someone wanted to spend a bigger budget than could be delivered at 3/60, we have a 2/60 or a 3/90 that we could offer to advertisers and that becomes our auto intender population for this RFP.  Another way to flex the population is by mixing behaviors.  We might acquire auto intender data from 10 different publishers.  By squishing this together, it is possible to generate larger populations – “3 visits to any of these 20 sites in the last 60 days”, but it makes regressing performance more challenging, particularly because the data points around frequency, recency, and the sites used to generate the data are typically not shared with the advertiser.  But then, advertisers aren’t asking for it.

In an effort to recognize the performance variability in different inventory, as well as recency and frequency, most behavioral targeters have transitioned to a vector approach.  To illustrate from an anonymous vendor deck that most of you have probably seen (and frankly, an awesome example of modern BT):

So as people surf the behavioral data providers data sources, their scores against a variety of behavioral targets are incremented and decremented based on how well the behavior provider thinks they will perform against an offer.  This is nice in that it allows you to easily keep track of many variables.  Rather than tracking recency and frequency and sites visited by user, you track a single number and then you have things that modify the number all the time.  Data storage is much less sophisticated.  The downside is that you are forced to trust the publisher to accurately score users – A “strong performer” in theory could have become a strong performer in a number of ways and the advertiser can never unwind it to see if there are attributes other than “vector score of X” that correlate with performance.

Vectors are great for data exchanges because you can easily vary the vector score to deliver appropriate campaign volume.  If you only have a few people with a vector score of 1000, you can dial it down to 600 and have a 10x bigger population.  Or something like that.  Even in the above example, it is impossible to tell when they stop being a great target because it is all relative.

Vectors today are used to gauge all kinds of things, even things like gender – “This person visits sports sites every day, ipso facto, they are male.”

Unfortunately, all of this ease of use is hamstrung by the fact that vectors obfuscate, even for the person generating the data, how someone became what they are.  In an effort to create a compact data structure that can return real time bids in 250 milliseconds, we have decided to limit ourselves to looking at the forest instead of trees.

Now, today, there isn’t much to really complain about regarding this vector strategy or simple bucketing.  Few, if any, advertisers are prepared to consume every individual behavioral data point and regress on it to determine optimal converting behavioral attributes.  Because advertisers aren’t prepared, publishers are using vectors and not really storing the data in a way they could give to advertisers.

When I think about the future though, I expect someone, somewhere, is going to want this stuff.  Thar’s is gold in them thar hills!