Tattertech Blog

I wasn’t going to touch behavioral as a topic again for a bit, but everyone’s talking about a recent story on Clickz (such as on Mashable, where I left several comments) I figured I would throw a things into the conversation.

While I am a huge advocate of the wonders of Behavioral Targeting, I have some issues with this concept aside from the privacy issues.  In regards to privacy, again as long as it’s opt-out I don’t see the tremendous concern.

A BT network today generally needs to establish specific behavioral contracts with publishers to make use of their users.  So if a network wants to gather behaviors from a major site (for the sake of argument) ESPN, they need to get a legal agreement with ESPN over the terms of use.  ESPN will then set up the behavioral coding from the BT network and the network can begin targeting the ESPN users across any site they touch.

The ISPs in this case will simply catch the packets on their own hardware between the end user and the publisher (say in this case ESPN) and resell that to the advertiser.  They will then pay whoever they buy the actual ad space from, but not the property that provided the relevant piece of data.  The argument I assume being that because the transaction occurred on the ISP’s hardware they have full right to mine that data.   The only disadvantage I see for the ISP is that one can assume they won’t be able to go to market with the name ESPN as a source of their behaviors, they can simply say, “Here are users who are interested in sports/sports related content.”  A ‘traditional’ BT network will most likely provide the advertiser with a list of publishers that make up the population bucket they’re selling.

If I were a major content provider with a huge brand name (say an ESPN, CNN, etc), I would be a tad upset at this development.  The ISP can capture their users with limited cost, buying remnant inventory from a ValueClick is dirt cheap, and yet the ISP can command a decent premium from an advertiser thanks to the potentially detailed targeting.  If I’m a major brand, this lowers the value of my users by creating a place to get them dirt cheap.  While a BT network could arguably depress the CPMs a publisher earned, the publisher has control to cease the deal and the receive an additional revenue stream (without using any inventory such that it’s essentially pure profit).  ISPs can easily undercut both the content provider and the BT network by avoiding the second payment.

All that said, I can’t say if publishers have any significant leverage against ISPs here.  With a typical BT deal, the BT network can’t use a publisher’s users without consent.  Theoretically, an ad network could compile behavioral data of users across the network they show ads on, however by going to market with this product (especially if it lists publishers by name) the publishers within the ad network can simply remove the tags and refuse to work with them.  As such the BT network is forced to seek the publisher’s opt-in.  The content providers may not have any means to cut off the ISP (short of damaging their own traffic and existing revenue streams) and could be locked into allowing someone else to monetize their users.

I’m not sure what the solution is to this, short of forcing the ISP to go down the route of a traditional BT network.  But forcing them to go down this route, by paying content providers for their users removes a good part of the competitive advantage the ISP has and probably makes the idea worthless.  In any case it will be interesting to see how this progresses.

I took longer getting to this one than I planned, because shortly after writing the first post the FTC came out and (shockingly) did the correct thing on the topic of privacy.  The FTC announcement can be read here, but the general point is that the industry should have the opportunity to police itself.  This is a great way for the FTC to say they’ve looked at the issue while changing very little.  Most major players in the market already have opt-out language available, and shortly before the FTC announced these principles AOL even announced more robust opt-out abilities. As the news is fairly old at this point, I won’t go any further into the FTC announcement (plenty have covered it already).

Instead I’ll just cover some major points on the issue in general:

As it stands with most behavioral tracking, no personally identifiable data is stored in the cookie.  With a few exceptions, all that is tracked is that a browser navigated to a site that demonstrates something deemed of value (perhaps looked at a sports article).  That browser cookie space then carries either the data itself or a link back to data that ties the browser to sports content.   Advertisers are then targeting simply anonymous groups of browsers with a sports value.  No personal information is passed to the advertisers and it does track you anywhere outside of that browser’s cookie space.

Now with something such as Beacon the issue gets cloudier.  This no longer occurs behind the scenes with anonymous identifies but rather clearly linking your identity with the advertisement and worse yet, displaying to a third party (your friends).  Typical behavioral advertising is simply a transaction between the advertiser and your cookie space (resulting in only you see ads related to your behaviors).

Again, as the issue has been beaten to death lately, I’ll cut the privacy discussion short but in my opinion with a reliable opt-out provided, the benefits of better matching ads to me outweigh the non-issue of a cookie (that I can delete manually anyway if I don’t trust the company)  that just has some data about certain types of sites I’ve been on.

Thanks to Google/DoubleClick, the FTC, and of course most recently, Facebook, privacy concerns have become front and center on the internet. While I’ll touch on the privacy concerns later (and thoughts on how to mitigate them), it’s important to first have a conversation on who benefits from behavioral advertising and why.

To properly scope the discussion - and to give the most basic background on what behavioral advertising means I am discussing here advertising that uses data determined from the user’s actions prior to showing the ad as behavioral (stored by cookie or login). Most likely this will concern tracking that the user visited some site or section of a site and using that to determine what ad to show to the user. This can obviously be across ad networks or within a publisher or portal.

So who benefits? Let’s start with the obvious…

The Advertiser:

It’s not difficult to see how the advertiser can benefit. In the case of a remarketing campaign, the advertiser gets to push the message to user’s who have already shown interest and push the user towards converting (or at least visiting again). Even outside of remarketing, the advertiser can tailor a message to an individual with specific interests or seek users likely in the market for the advertiser’s products/services. The advertiser can focus their efforts on the correct types of users (at least if they understand their customers) and cut out wasted impressions.

The Publisher (and/or Ad Network):

This one’s not too difficult to understand either. In this case the publisher (or ad network depending on the circumstance) can demand higher price from the advertiser because of the increased relevancy and value of each impression. And the publisher doesn’t even have to use up their most valuable impressions on the advertiser. Instead of having to put the ad on the front page or next to the applicable content (inventory that is already valuable and has a high price), the publisher can provide the same level of value outside of those premium impressions. This means greater monetization across their entire inventory. The location is no longer the primary concern for price, but rather the user being reached.

The User:

And here’s the crux of the deal - the end user benefits too. Rather than seeing useless ads that have nothing to do with actual interest, ads are delivered that are relevant to current interests. A greater debate about the good of advertising aside, a person interested in purchasing a car in the next few months is going to be more interested in ads related to cars than ads about cellphone plans. Behavioral even in its most simplest form is the recommendation engine of the internet advertising. It funnels the long tail of advertisements to the users that are potentially the most interested.

In short:

• The advertiser reaches the proper potential customers with the least waste.
• The publisher/network potentially increases the value of its least valuable inventory and drives higher margins/revenue without increasing page views or other metrics.
• The actual user sees more relevant ads instead of random clutter and hopefully derives some value from the advertiser’s directed message.

The largest cost in this ecosystem (although there are some technology and scaling concerns behind the scenes) is the user’s sense of privacy while I’ll tackle in my next post (such riveting topics as opt-in vs. opt-out, acceptable tracking, etc).