Opt-In ISP Targeting: ATT Mulls a Different Route

As the NebuAd controversy has materialized, one of the principal criticisms of their agreements with ISPs -- and about BT in general, has been the reliance on an opt-out model. As you know, opt out requires the consumer to actively engage and tell the ISP and BT providers, "No, don't track me." The alternative, opt in, hasn't gotten a lot of traction, at least in the US, principally because there is an expectation that most people would not bother to opt in even if they were fine with the idea of targeted advertising.

Based upon the feedback provided by major ISPs in their responses to the House Energy and Commerce Committee, however, it appears that one major ISP is considering opt in as the right thing to do. This according to an analysis last week in the NY Times Here is an excerpt from Saul Hansell's analysis of ATT's vision:

While the company said it hadn’t tested such a system for monitoring display advertising viewing habits or committed to a particular technology, it expressed much more interest in the approach than the other big Internet providers who also responded to the committee’s letter.

AT&T did however promise that if it does decide to start tracking its customers online, it will “do so the right way.” In particular, the advertising system will require customers to affirmatively agree to have their surfing monitored. This sort of “opt-in” approach is preferred by privacy experts to the “opt-out” method, practiced by most ad targeting companies today, which records the behavior of anyone who doesn’t explicitly ask to not to be tracked.

The passage and entire article interested me greatly because ATT fully discloses their keen interest in participating in the online ad revenue boom, but their parallel determination to ensure that consumers are satisfied with their approach and choose to participate.

This is nothing short of gutsy. I have to admit that I have never seen opt in as a realistic solution for the current approaches to BT because while the consumer does receive value in the form of better online content paid through higher CPMs, they do not get something tangible in their hands for participation. I think this benefit is too abstract for most consuemrs to grasp and value. I do NOT subscribe to the theory that targeted advertising has significant consumer value to the average consumer, at least in the rational abstract.

Is ATT right? I hope so -- this is definiotely a situation in which I would be delighted to be wrong. And perhaps they are anticipating a different consumer value proposition for the solution they ultimately select. Is it possible that ATT is seeking a solution that provides real consumer value in exchange for participation? And I mean REAL consumer value, not their ersatz value trumped by companies like Phorm with their phish protector.

It's going to be interesting to see what happens.

Thanks for reading, and don't forget to write.

Rep Markey Wants More Info About Embarq and NebuAd

In this MediaPost news story, they report that NebuAd continues to be under Congressional scrutiny despite the generally friendly reception they got last week at the Commerce Committee hearing on BT and privacy.

According to the piece, ISP Embarq CEO Tom Gerke was sent a letter that questioned whether his company had provided robust notice to consumers about the tests they ran with NebuAd.

The text of the letter, which I found on Congressman Markey's site, appears below:

July 14, 2008

Mr. Tom Gerke
Chief Executive Officer
Embarq
5454 W. 110th Street
Overland Park, KS 66211

Dear Mr. Gerke:

We are writing with respect to a recent test conducted by Embarq to tailor Internet advertising to the web-browsing patterns of individual Embarq subscribers. We are interested in the nature of this test as well as the impact that this test, and the underlying technology it employed, could have on consumer privacy and other issues.

We understand that Embarq conducted a test earlier this year in a select community in conjunction with NebuAd to create consumer profiles for the purpose of serving ads to consumers based upon their search and surfing habits. As you may know, questions have been raised regarding the applicability of privacy protections contained in the Communications Act of 1934, the Cable Act of 1984, the Electronic Communications Privacy Act, and other statutes, to such practices.

In particular, we are concerned that Embarq may not have directly notified the subscribers involved in the test that their Web use was being analyzed and profiled. We therefore request that you answer the following questions in order for us to better understand the nature of the test conducted, its impact on consumers, and the broader public policy implications of this technology.

1. In what community was the test conducted and how was that community chosen?

2. How many subscribers were involved in the test?

3. How did Embarq notify subscribers in the affected community of the test? Please provide a copy of the notification. If Embarq did not specifically or directly notify affected subscribers, please explain why this was not done.

4. Did Embarq conduct a legal analysis regarding the applicability of consumer privacy laws on the service used in the test? If so, please explain what that analysis concluded.

5. Please explain why Embarq chose to conduct the test allowing consumers who objected to "opt out" rather than first asking customers to "opt in."

6. How did Embarq notify subscribers in the affected community of their opportunity to "opt-out" of the test? If Embarq did not specifically or directly notify effected subscribers of the opportunity to "opt-out," please explain why this was not done.

7. How many subscribers in the affected community opted out of participating in the test?

8. Did Embarq conduct a legal analysis regarding the adequacy of the "opt-out" notice and mechanism employed to allow consumers to effectuate this choice? If so, please explain what that analysis concluded.

9. What is the status of the consumer data collected during this test? Has it been destroyed?

Thank you in advance for your attention to this matter. We respectfully request a response by Monday, July 21, 2008.

I don't know if Embarq notified their customers or not beyond including info about it in its privacy policy, though this passage from the MediaPost article indicates that many ISPs that worked with NebuAd did not.

But software researcher Robb Topolski, who recently tested NebuAd and concluded that the program violated users' expectations of privacy, said the vast majority of the Internet service providers who worked with NebuAd did not seem to send separate notifications to subscribers. Instead, they apparently placed information about the program in their terms of service, privacy policies or other lengthy documents subscribers generally ignore.

I am anxious to see Embarq's response. What constitutes robust notice is ill defined by the government, at least in form. The government, to my knowledge, does not have a proscribed process by which consumers are to be informed.

Is it enough to put it in the privacy policy? Is it enough to put it in a brief and well organized privacy policy? If they put it in the privacy policy, do they then need to alter the customer that the privacy policy has been altered? If so, how must they notify? Would an on site notice do it? Is email OK? Do they need to send a letter?

Presumably the answer to this relates to whether each of the tactics described above resulted in satisfactory levels of consumer awareness.

The googly, from Embarq's perspective, is that the generally accepted means of notification in BT has been in privacy policies. Google, for example, does not send out letter before you download their toolbar telling you that all the places you visit are fair game for analysis.

Will ISP targeting be held to a higher standard than the rest of BT? I think that would be dead wrong. To me, the difference between ISP targeting and traditional BT from a privacy perspective seems to relate to the amount of info collected. A notificaiton process is either right or wrong, whether the BT provider collects 20% of my web visits or 100%. And if i am not mistaken, there are currently a number of companies out there diligently pairing BT data with PII, and they are doing so with modest consumer notification. For example, portals and Facebook. NebbuAd may be collecting more information, but other companies are collecting more PERSONAL information.

I'm not sure what I think the standard should be in terms of the form of notification. But I am sure that it should be applied to all BT, not just the technologies that collect the most complete picture. Because if the latter route were taken, at what point would the amount of data collected lead to the requirement of more outbound notification practices? 99%? 98%? 73%

Off my soapbox.

The challenge of this kind of notification is one of the classic push me pull yous of marketing. Often, the government looks at measures like opt out rates to determine whether the average consumer could be reasonably assumed to be notified. There is a lot of grey area between tucking it away where few will see it and sending out letters or emails.

One of the most interesting answers will be to question 5 -- about why they chose to do opt out rather than opt in.

Markey also released a statement when he informed the world about the request for information.

"Surreptitiously tracking individual users' Internet activity cuts to the heart of consumer privacy. The information collected through NebuAd's technology can be highly personal and sensitive information. Embarq's apparent use of this technology without directly notifying affected customers that their activity was being tracked, collected, and analyzed raises serious privacy red flags."

Ouch.

Thanks for reading, and don't forget to write.

Veoh Announces BT Video

BT and video haven't really gone together very well -- in large part because you don't cookie a video stream. There are other factors as well -- such as the need to categorize the increasingly ginormous quantity of video available on which to advertise.

Enter Veoh, which announced its launch today of a BT video solution. The platform allows one to target on interest areas -- like tech -- or more esoteric segments, like the fans of a particular show.

This MediaPost piece lays it out in some detail.

I presume that this is either being done through the delivery of tandem banners bearing cookies or through tracking based upon logged in users.

It'll be very interesting to see how well the market accepts this innovation.

Thanks for reading, and don't forget to write.

Ad Network Glossary

A nice, succinct glossary of BT terms is available here at AdNetworkX.com

Thanks for reading, and don't forget to write.

ValueClick Tests Its Sales Mettle...And Comes Out Fresh and Sparkly

What a great time I had last Tuesday as a judge in ValueClick's (VC's) sales pitch competition to get their key folks ready to sell Predictive BT, their new entry in the crowded and confusing behavioral arena.

So I must admit that for me this event had two big things going for it: I love hearing a good pitch (I really do,) and I am fascinated by BT.

ValueClick had five people prepare their own presentations and selected 5 mystery judges to roleplay actual clients with actual challenges and predispositions toward VC. While each seller's counterpart posed a specific kind of challenge, it was up to the judges how they behaved, engaged (or didn't), asked questions. We could lob easy questions or try to set their feet on fire. It was all unpredictable just like an actual sale must be.

Each judge engaged with one sales person but rated the performance of all five by listening and filling out a quantitative survey.

Selling BT is always a challenge, and it is particularly difficult in light of:

How different VC's solution is from other BT platforms
The frosty privacy environment
The seemingly counter intuitive concept that there is an ideal number of data points to truly predict future behavior.

But their sales team was up to the challenge.

Let's pause for a moment for a commercial message, both because I heard about this solution five times, and because I have always liked ValueClick -- their products and their people.

VC's BT solution differs from other BT models in four key ways:

1. The model attempts to distinguish between "in markets", "drive bys", and "fanatics." By which I mean that they believe that there needs to be a system whereby people who are truly contemplating a purchase are separated from browsers and people who devour content on a subject but are not actually in market. Think cars. I hear about the new Scion XB replacement, and visit MotorTrend to read about it. And then forget about it. Or, I love cars so much that I visit MotorTrend 12 times a week, even if I am not about to buy a car. Or I am someone that is weighing options with a finite number of contacts with relevant brand information online.

2. The model does not assign individuals to one "bucket" but rather enables marketers to target people who may be prospects in a variety of categories.

3. The approach offers a real melange of data points -- from classic banner ads to search (on VC's comparison sites and via buying 3rd party data,) to affiliates via the CJ.com subsidiary. And it does this without needing PII. VC's subsidiaries in multiple lines of web business are a real strength here.

4. The model is constantly adjusting to changing circumstances. I liken this to a broad set of dynamic regression analyses, though that may not be how it works. That's just how I understand it to work.

So my role was to be a person who knew a lot about BT and ran digital planning and buying for a credit card. My plan was to be polite but disruptive, which is a common problem with people who know a lot about something. They try to run the meeting, and in the process make it very difficult to tell a cogent and coherent study. The meeting can devolve into racquetball, with the prospect hitting the ball in different directions and the salesperson forced to essentially do polite combat, returning the ball defensively. At the end the salesperson is tired and no sale is made. S/he didn't lose per se, but not losing is not the same as winning.

My plans were foiled by Nick Kounalakis, the salesperson I was assigned to role play with. Here's the play by play.

0. Before I could throw a wrench into the Nick show, he recited all the research he had done about my supposed problems, and had really deep knowledge of the special issues of my company. He also identified points of relationship -- shared life situations, etc., which made it much tougher to give him a hard time. Not that I intended to be hostile, but it was all part of his approach to directing our interaction and moving toward a close.

1. I tried to take control of the meeting by setting an agenda of three topics that relate to the product but don't allow the salesperson to tell the story. Nick got the better of me by reasserting control of the discussion by making me accept his agenda. Which I was sort of forced to do out of a sense of decorum and because his agenda was far more coherent than mine, What he did was listen to what I said and reframe it in the context of the story he wanted to tell.

2. I had a great deal of questions about privacy and issues surrounding it. And Nick answered every one, but each time reframing it into the next bullet point on his slides.

3. I tried to lead him astray in the discussion of the range of data points they use by focusing on Search and whether Google and Yahoo were part of their model. They aren't, of course, but rather than telling me that, he told me what WAS included, which included a variety of very appealing contexts that I had never really considered before.

4. I tried to steer him to a discussion of cookies and the issues that cookie deletion creates. And dammit if he didn't shut me down with a 4 second answer. And I want to emphasize that I didn't FEEL shut down. But I was.

5. I tried to get us on four different tangents, but Nick was like a chess player, always redirecting me to the point he wanted to make on the slide.

I was particularly impressed with how he created a broader sales platform -- that BT would be a component within a larger set of products VC could offer a credit card issuer. Mo' money was his very logical objective, and hey, if a salesperson can devise a solution that spends mo' money responsibly, then why shouldn't s/he get it?Which was very sharp because the challenge of BT in the credit card business
is that it really isn't going to be the lynch pin of a buy -- credit card companies will always buy a lot of cheap junk because it is DR-efficient and retargeting because it works like a dream in this category.

In short, he sold. He didn't just teach, he didn't just explain, he used education and empathy to meet my roleplayed needs.

Nick won the contest. And by focusing on him, I do not mean to diminish the excellent selling that was done by others in the competition. Each of the contestants did a great job of dealing with the at times rather difficult role players.

So congrats, Nick.

And even more to the point, kudos to ValueClick for recognizing the complexity of this sale, offering a comprehensive training environment, and giving their new messaging a dry run before they went out into the market so they can be better prepared when they actually deal with prospects instead of thespians.

Thanks for reading, and don't forget to write.

Gooey Brother is Watching With the Google Toolbar

I've talked a couple times about the Google toolbar and how it holds the potential for an enormous amount of data collection because the T&Cs that no one reads permit Google to collect all of the user's browser based searching behavior.

Now comes this post on TechCrunch which points out that the new Google Ad Planner may be using this toolbar data for ad targeting.

Google isn't saying if the toolbar data if now being used for targeting. Which makes me all but certain that it is. If it wasn't, I think they'd be all over themselves to deny it.

The thing is, if the data are anonymized I personally don't consider it a major privacy problem. WHAT IS A PROBLEM is the secretiveness of this and other companies who seem to have forgotten that nothing goes unnoticed on the medium that they helped pioneer.

Say TechCrunch:

If that is true, Google should disclose the fact, even if it is only using the toolbar data in an aggregate, anonymous form. Because most people who download the toolbar are probably not aware that the data it collects about their surfing habits can be used to target ads back at them.

To paraphrase Gilbert and Sullivan, "I am the very model of a GODDAMN TROJAN HORSE."

Thanks for reading, and don't forget to write.

Self Regulation: Can BT Do a Better Job?

So let's talk about the NAI for a bit. NAI is the Network Advertising Initiative, the group that is defining standards for BT and privacy, and also houses the BT opt out mechanism. The mechanism enables the consumer to opt out of being racked for BT.

I have some positive comments and a peeve or three.

First Positive: Our industry voluntarily put up an opt out site. Remember how tooth and nail the telemarketers fought opt out? OH. MY. GOD. did they lobby against it. Back when Do Not call was an issue, I had a friend in Congress who told me they had NEVER gotten so much mail about anything. Those cards and letters crushed the dinnertime cold call lobby only because Congressmen and Senators genuinely thought the wrong vote would lose them an election.

It took a law to make the Do Not Call happen, and the telemarketing industry had so overstepped the bounds of acceptable business practice that everyone with a browser opted out in about ten seconds after the site when up. Death of an industry, more or less, and THANK BLOODY HELL for that. I spit on your graves, prospector telemarketing kingpins.

So the BT industry put up their site without a law -- indeed to avoid one. But whatever the reason, I'm happy it is up. And generally works.

Second Positive: (After a bit of manual doing) I opted out as an experiment, and the opt out worked. I don't get the cookies anymore. It's as if my McAfee scan gives me an incredulous look when it does its auto scan and finds nary a tracking cookie. It's like the Maytag repairman of 2008. Bravo.

Third Positive: The NAI is putting BT standards together in a very complex space. This isn't simple stuff. My next post is about their proposed standards -- which have been out for public comment for a couple months. I applaud them for trying.

First Peeve: The site is not terribly easy to find, and it's the place people need to go to opt out if they so choose. The site has a Google Page Rank score of 8, which is high, but I type NAI into Google and the site comes up fourth. On Yahoo it is not on the first page. Call me crazy, but an org with the sharpest webbies on the planet could make itself easier to get to. Type "Opt Out of Cookies" and NAI comes up 5th. Type "Behavioral Targeting Opt Out" and the NAI is not on the first page. I don't think that the NAI is deliberately hiding (heck, not if it has a PR of 8,) but part of being a responsible self regulation org is making yourself easily available to consumers. Look, the world is give and take. I am not suggesting the NAI OEM a widget with the title "You want to opt out, don't you?" with a Big Brother picture above it, but an org of the biggest web smarties in the universe can make a site that is easier to find than this. I mean, search NAI on Yahoo and it's not on the first page???

Second Peeve: That bit that so many web publishers are saying about consumers enjoying the Internet more because of targeted ads. Well, sample size of one, but that is a load of crap in my experience. I opted out to see how different my Internet experience would be. Frankly I didn't notice one at all. OK, actually I perceived that I saw more Mortgage snake state banners, which seems logical given that mortgage DR is total bottom feeder untargeted sub a buck CPM crap. Net net, selling consumers on the idea that the internet is a magical experience because of targeted banners -- it just ain't believable. Or true, come to that. And please spare me comments about the surveys that say people prefer targeted banners. Yadda yadda. It's crap and you know it. In fairness, there is nary a mention of this on the actual NAI site, but it is a common message coming from members. It is on page one of their proposed standards doc.

Third Peeve: When I went to opt out, I had to fill in which ad networks I wanted to opt out of by ticking 16 boxes saying I want to opt of 16 networks. I find this silly. What is the consumer REALISTICALLY going to do. Surmised internal dialogue of Mabel in Decatur: "Well, I want to opt out of Advertising.com, but those nice people at ValueClick, I think I will keep their cookies." Again, I don't think this is deliberate barrier creation, but it strikes me that some consumers might THINK they were being put through hoops. Which is not a good thought to provoke if you are trying to head off regulation.

Fourth Peeve: Even when I tick the 16 boxes, I got an error message that for some reason I am not opted out of a couple of the networks and need first try again (didn't work) and then go to their sites to finish the job. Hello! A faulty app is NOT my problem. That is not a good consumer experience. Opt out either works or it doesn't. If for some reason it is actually difficult to make a perfect opt out mechanism, I say work it out, because it's either make it work or have the feds order you to.

So, in sum, I applaud the efforts of NAI to self regulate. I just think they need to step up the game just a bit. While as a citizen I am not opposed to government regulation of any industry, I fear the sort of regulation that may come from our government because so far at least I am totally unimpressed by their understanding of the Internet. Ignorance does not begat good law.

If you've read two posts on this blog you'll know that I think BT (at least the traditional ad network sort) is just fine, and that it is essential to pay for the web. But advertisers are always suspect in the eyes of consumers. We need to prove we are worthy of their trust. NAI can help, and is helping. I just think it could help more.

aCerno: The Funnel Frontier

Those familiar with the team at Catalyst:SF know that our folks have worked to develop more effective marketing for a variety of ad networks and rep firms over the years -- I counted 15 last time I polled the group.

So it was somewhat surprising to me personally when I found a large ad network I had never even heard of, much less understood.

That network is aCerno, the digital side of I-Behavior, which has years and years and years and years and years of very successful experience in direct mail.

DIFFERENT PAGES, DIFFERENT BUYING PROBABILITIES

aCerno is just plain fascinating, so let me tell you a little bit about them. But first we need to draw distinctions between four kinds of web pages: HTTP, BUYPATH (HTTPS and HTTP portions of the buying funnel,) SOCIAL, and SEARCH.

HTTP pages are...well, the bulk of web pages online.

BUY PATH: Pages that are actually part of those last moments of a consumer's decision to buy something online.

SOCIAL: Social media pages. These are also HTTP pages (almost always), but the reason I draw a distinction between HTTP-SOCIAL and HTTP-OTHER will be more transparent in a moment.


SEARCH: The pages that collect our queries and report back their findings for our queries.

Now, if you stop and think for a moment as a direct response marketer might, each of these different classes of pages is, on average, going to have a different probability of reaching someone ready to buy something. For instance, social media discussions CAN be about buying stuff, o'course, but tons of them are just talk. For a social net to provide valuable BT info, they will need to be able to draw such distinctions. No doubt that's part of their plans.

Next in line would be HTTP. It is such a broad category that naturally different pages would have different probabilities of reaching a prospective buyer. For example, if one is browsing the page for the book What Happened? on Amazon, versus looking at the front page of Politico, there are naturally different levels of probability that one is going to buy the book.

Most BT companies focus on the HTTP pages at the top and the middle of the buying funnel -- they are where ads run. Unfortunately, though, most of the best HTTP pages for predicting prospective purchases are NOT available for advertising. Amazon won't sell a banner to BN.com on the What Happened? page. But the NYTimes review page for the book is available, and that'd be a goodie for a BT ad. What the traditional BT companies have, though, is VOLUME -- billions of ad placements on millions of pages. And when you aggregate that info, you really can identify people who are more likely to buy. That's what the whole industry is based on.

Next in line would be SEARCH. Search is a task oriented process by definition, and if you could know who had typed "Buy What Happened" into Yahoo Search, well then you've got pretty good odds of reaching good prospect eyeballs. Which is, of course, why SEARCH is such an enormous portion of total online marketing spend. BUT in many cases, there is still a little emotional difference between searching and buying. Search is a great way of increasing your odds, but there is still some waste.

Which brings us to BUYPATH pages. Naturally there are different sorts of such pages. If I get halfway through the shopping cart/buying process and then abandon, mine'r probably good eyeballs for retargeting or for intercepting a customer from some other retailer. Similarly, if a BT company could see that I made a certain purchase, they naturally could determine other related items I might buy in the near future, or when I might be willing to buy again.

SO WHO SEES WHAT?

"TRADITIONAL" BT COMPANIES: Most BT companies see all of the HTTP pages that have "their" ads on them. Note that they don't see ALL HTTP pages -- only the ones where their ads run. They don't see HTTPS, most SEARCH, or most SOCIAL. Generally, HTTPS is considered "too hot to handle" by most companies. That's mighty personal stuff, no matter how anonymized.

RETARGETING COMPANIES (a service that some of the basic BT companies also provide) generally know what you do and where you go on a marketer's site within the HTTP pages.

SEARCH SITES/PORTALS: They can see anything you do on their pages. And you might think it stops there. But if you have a toolbar, like the Google toolbar, they have the ability (and the "right" -- page 137, paragraph four, subsection 11, desk reference 22833495 of that 6 point type agreement you agreed to without reading it said so) to see anything you do that goes through the browser. The extent to which these companies actually DO look at all that info is unknown to me. Most people believe that Google uses more of this info than the other Toolbar-ed portals -- after all, their entire model is based upon knowing you and what you want. Yahoo and MSN can use it, but as I read their privacy policies and T&Cs, I don't think they ARE using it at this time.

SOCIAL NETWORKS know what you are doing on their pages. They know what apps you have, what you chat and blog about, and the like. But their reach beyond their own walls is limited. And it appears some consumers like it that way. Facebook's Beacon debacle really stems from their learning what you did outside their walls.

Gentle reader, if you're flaggin' in terms of attention span, I ask you to bear with me for one more paragraph, because it really is about to get VERY interesting.

THE FUNNEL FRONTIER



So, if you were keeping tabs, you probably noticed that BUYPATH pages are generally not tracked except by the marketers who own a site. And these marketers don't have a view into the sites of other retailers, so their view is fairly limited.

aCerno saw this buying funnel opportunity and ran in. If you think about it, buying funnel tracking is pretty analogous to what DR companies do with Direct Mail (DM.) Everything that can be tracked IS tracked with DM, and there's no aversion to collecting and using PII neither. Heck, the whole business is based upon appending demographic, psychographic, and purchase info to PII.

But let me be clear, aCerno is NOT collecting PII. Rather, they are using random numbered cookies to track users anonymously. I'm not sure if they track HTTPS or just HTTP, but they're right down in there in the bottom of the funnel in that customer stream. In a recent interview with Internet Retailer, archived here, aCerno's CEO explained the tracking process thus:

“If you’re interested in buying a big-screen TV, and you’re on the web going to e-commerce sites and looking at different brands and price points, what we see is a cookie number unique to a browser that is associated with all that viewing of products and information about big-screen TVs,” says Tom Sperry, CEO and chief privacy officer of aCerno. “We know cookie 456789 is probably very interested in purchasing a large-screen TV. Our advertisers can deliver a message to that user and influence that purchase decision on which brand he’s going to buy and where he’s going to purchase.”

The model aCerno uses is to combine data from over 375 major retailers and develop behavior profiles of consumers based upon this cross seller experience. Many have compared their model to the convention among catalogers to pool mailing lists and interest areas to collectively improve ROI. The list of retailer sites is a secret. And for damned good reason. Talk about opening the kimono! These companies are entrusting aCerno with amazingly sensitive info from the perspective of their businesses.

WEB NOT DR ENOUGH?

What aCerno says, interestingly, is that before they came along, online worked under a broadcast rather than a DR model, and that their network changes all that. Because they collect these buyer profiles and then buy ads from publishers to target audiences off the retailer sites. Let me be clear, this may sound like retargeting but it is actually much larger than that. This is full fledged cross category predictive BT based upon BUYPATH info.

I've never used them, but the model intrigues me, and I think it is worth a look for DR marketers. Like any other decision to use a network, one must compare sales and pricing to other networks to find the most efficient opps. But their model is very different, so if I had lots of units to move, I'd RFP 'em. Never hurts to see...

Thanks for reading, and don't forget to write.

Linkstorm Redux

Those who are regular readers of this blog know I am a big fan of Linkstorm, and the recent interview of David Sidman, their CEO, had an interesting new take on the company -- at least it was new to me. In the interview, conducted by Phil Leggiere and still available on MediaPost, Sidman posits that their product is a different take on BT. Their DHTML banner product lets consumers find what they want rather than us trying to surmise their wants via past/predictive behavior. (The interview is from April, but in case you didn't see it, give that link a click.)

Wait, that is in incorrect simplification of their view. Not only does he communicate that their menu banners are a form of BT, but also he explains that they can be a complement to BT. Rather than me paraphrasng the content, here's an interview excerpt:

So, to use a football analogy, behavioral targeting will get you into the red zone — but no farther. We don't do our own ad placement. However, if a customer already using BT has the data to know you are a certain profile, we can take that data and further customize the way menus are structured. If you know from their browsing that a customer, let's say, is a bargain hunter, then you can organize your different offers by price range.

Nice POV.

Thanks for reading, and don't forget to write.

iMedia: ValueClick and Predictive BT

I loved the breakout session on predictive BT. Hosted by of ValueClick, and featuring a great case study about Cricket Wireless, the session really laid out the latest in BT in clear and concise ways. And Matt Boyd is the consummate gentleman, acknowledging his company’s perspective while offering a much broader and less self serving view of the BT industry.

For his gentlemanly conduct alone, Matt and his team deserve to be on your next RFP. Not to say that those are the only reasons why they should be on your list. But we should also reward the good guys who capture the spirit of the event and recognize that providing info is the best way to win.

Also, because I knew Matt when he was just starting in the biz and knee high to a grasshopper, and it’s nice to see his remarkable talents rewarded!

Thanks for reading, and don’t forget to write.

iMedia: United Online

I just finished my first iMedia event, a dinner hosted by United Online. The dinner itself was very low sell (a wise move for a Saturday night,) but the back of the invitation really got me to thinking. United Online, in case you live under a rock, is a collection of web businesses including:

Classmates.com: The site that can put you in touch with the people of your past. Built on a subscription model and an extensive email sponsorship program.

NetZero and Juno: Value-priced ISPs that essentially trade consumers lower dialup service for liberal privacy policies that enable marketers to get insights and precise BT.

MySite: A web hosting and site builder offering not unlike parts of register or GoDaddy, but less expensive.

Names Database: Classmates broadened into a searchable database to find people you know or want to know.

MyPoints: The loyalty cum marketing platform.

FTD: They of flowers fame.

At first glance, it seems an odd collection of businesses. But look closer, and you begin to see the essence of the new web – the database and behavioral marketing. United Online has, through its various businesses, a massive collection of people’s profiles that could prove extremely useful for marketers. Whether as a research platform or as BT-enabled ad and marketing platforms.

Naturally the amount and types of info that they have from these disparate businesses varies, but in sum you can see how their core is data. With MyPoints, they probably have access to people’s bloodtypes ( ;-) ) while with Classmates considerably less. But United Online is widely considered one of the richest database companies around that is marketer friendly. Definitely an option to consider next time you want research or BT.

One interesting aspect of their businesses is to consider the health and stage in the product life cycle. A look at their annual report demonstrates steady declines in dial-up ISPs, but very strong growth in selling premium subs to Classmates. And the recent acquisition of FTD adds another interesting layer to their operations. And I’m guessing that being a data delivery device for florists is about as profitable as would be printing money. Attract eyeballs, match orders to florists, get a big cut. Good eatin’.

United Online makes solid money too, so the model seems to have some merit.

Sort of a new definition of synergies in what otherwise looks like a mapcap collection of stuff. But perhaps Untied Online is like McDonalds, which is really a real estate company that sells burgers.

United Online is a data company that provides a variety of disparate ancillary services. They serve a nice bowl of chili, too, and for that and the thoughts their company has driven for me tonight, I am thankful.

Thanks for reading, and don't forget to write.

A09275 Third-Party Internet Advertising Consumers' Bill of Rights Act of 2008

Well, that is a seemingly dry title for a post on a blog that features lip syncs on Fridays. But this bill, currently before the Assembly of the State of New York, has some wide reaching implications for our industry. Though I have to say, upon reading the bill, it's a little less scary (to my nonlawyer eyes) than industry mavens would seem to believe.

Currently with 57 sponsors, this bill would seem to make law what the NAI is trying to do on a voluntary basis. The bill was inspired by the acquisition of DCLK by Google, and the justification submitted with the bill reads as follows:

Recent acquisitions of third party advertising companies by large Inter-net companies have raised serious and important privacy issues with respect to who is monitoring an individual`s online behavior for marketing and other purposes. While many companies have implemented rules and polices to protect consumers, other have not. It is the responsibility of the State to adequately protect and inform its citizenry of who is tracking their legal online activities and subsequently selling that information.

Assemblyman Richard Brodsky, who appears to be the force behind the bill, is, it should be noted, not without understanding of digital. He has written numerous internet related bills, and his concerns in this case appear to be related to the fact that while many of the ad networks out there are voluntary participants in NAI, some are not.

It won't surprise you that I hope the bill is not enacted. Though, it should be noted that state law regarding the Internet generally is toothless because the Internet is by its very nature Interstate, which makes it the purview of the Federal Government. But the bill is significant, and the number of sponsors should alert all of us to the fast that consumers are not comfy with the current state of affairs (or perceived state of affairs) as regards BT.

I think part of the challenge is that US communications law was almost entirely written before the Internet was even a gleam in Al Gore's eye, so rules and best practices regarding the space go forward in fits and starts. The FTC made a concerted effort to address the internet in the period 1998-2001, and held a major set of discussion on BT specifically not all that long ago, deciding at that time that opt-out was an acceptable way for us to go forward in the US. Their recommended privacy guidelines live here.

This post on Ars Technica outlines a lot of research on how consumers feel about BT, or what they believe BT to be. And based upon my reading, there really is a great deal of consumer discomfiture on the topic. Which brings us to what I think would be 5 ways we can begin to address consumer concerns going forward:

1. Be transparent. Every time someone tries to do something BT related without full disclosure or CLEAR AND CONSPICUOUS disclosure, the conspiracy theories start flying.

2. Be real about benefits to consumers. I accept that there are some benefits to consumers re targeted advertising. But being real, in my opinion, requires that we state the whole truth -- that that the real reason our industry needs the higher CPMs made possible via BT is to pay for the web. I read 35 newspapers a day (or at least the page one headlines) and I don't pay a nickel for any of them via subscription. I pay $25 for DSL a month. About the price for a subscription to ONE of those 35 papers. Something connected to me has to pay for that usage.

3. Give people a reasonable way to opt out. I think that if people know there is a way to opt out, that will satisfy them -- that they won't all go rushing to opt out. This isn't like getting telemarketing calls -- this won't be like the DNC registry. Making the opt out system voluntary as well as easy to find online wil head off a lot of legislative efforts.

4. Stop with the 35 page T&Cs. Check out the eBaum's World privacy policy. This is how it should be done.

5. Not be Chicken Littles every time anyone in government brings up the topic, but rather engage them and ensure we're on the same page. Hey, I am opposed to legislation, but the way to avoid it is for us to act responsibly, not to villify people with concerns.

Thanks for reading, and don't forget to write.

Digital Thought Leadership Outside of Area Codes That Begin With 9 or 1

I had the good fortune to speak at the Las Vegas Interactive Marketing Association's "Behavioral Bustle" event yesterday, and I am so glad that I did. It was great to meet such a smart and energetic bunch of people doing so many innovative things with digital.

And it got me to thinking about how coastally biased our industry is. There's seemingly a major digital event every week in NYC, LA and SF. But what about the Second Cities? We're not talking about small markets. Over 110 people showed up to hear me drone on about privacy issues with BT, and these are not digital dabblers. The industry in LV is apparently very strong, particularly in the area of BT.

LVIMA's head, Shawn Rorick, put on a great event featuring Specific Media and DrivePM, both of which put together great presos on the state of the industry and where it is heaed. Liberty Callas of DrivePM did one of the best presos on the topic of BT I have ever seen, explaining nuances and really getting to the heart of one of the most critical issues: scalability.

So why is our industry so coast centric? It made sense perhaps 5 years ago during the digital nadir, I mean, you have to fish where the fish are.

But the digital ocean is now truly global, and markets like LV -- and DC -- and Portland -- and Seattle -- and Atlanta -- and Minneapolis -- and Chicago -- and St Louis -- and Dallas -- and and and are growing dramatically, and are clearly underserved. Wander on over after you read this to Lee Odden's Online Marketing Blog for a more complete list.

Fortunately, the associations for these markets do seem to be getting some sponsors, but the rest of the world needs to wake up to the dollars and minds being showcased in cities other than those served by SFO, LAX, and JFK.

And thanks again, LVIMA!

Thanks for reading and don't forget to write.

BT WEEK POST 3: THE MECHANICS OF BT

The Mechanics of BT


The following is a simplified description of how a BT ad is served. When a user requests a page managed by a BT enabled ad server, the user’s PC also shares the cookie info. Because the behavioral targeting platform stores the user’s behavior with that cookie text on its servers, it can instantly know what sorts of ads would be most appropriate for you.

Imagine that your profile tells it that ads for real estate, financial services, travel, and auto are relevant to you. Based upon this info, the server could compare the revenue it could earn from serving you various ads. Perhaps the travel ads are worth a $12 CPM, while real estate is $15, financial services are worth $19, and auto are worth at least $20. The server selects auto and identifies the type of auto ad relevant to you. The ad is then served on the page as it loads for you. All this activity takes places in the milliseconds between the time you request a page and the time it appears on your screen. See below:



There’s a lot more to it than that. For example, an ad server might decide to serve the travel ad if the time limit for serving the travel impressions is about to run out. Or more than one BT provider might have the rights to sell a particular ad space. But what’s above? That’s the gist.

BT DEPENDENCIES

The system naturally depends on several things:

1. That the consumer has a cookie. Increasingly, consumers are eliminating cookies from their PCs periodically. Comscore research from 2007 reported that about 30% of consumers deleted their cookies at least monthly as a way to reducing their privacy concerns.

2. That the consumer has not opted out of BT. Partly as a way of avoiding federal regulation of BT, the BT industry has set up a voluntary opt out program that lets users choose not to participate in BT targeting. The jury is still out on whether this voluntary program will be deemed sufficient to avoid federal regulations or statutes. The Network Advertising Initiative website (http://www.networkadvertising.org) allows consumers to opt out of ad networks individually or collectively. See below:



Opting out doesn’t reduce advertising delivered, only the extent to which it is behaviorally targeted. When I opted out of all programs as an experiment, I appeared to get a lot more mortgage and other low CPM advertising. Which makes sense, of course.

3. That a BT enabled ad decision server is monitoring the ad placements on the requested page. Not every ad unit is served by a BT enabled server. But an increasingly large proportion are. After all, there is A LOT of money to be made. And advertisers LOVE this technology because it leads to tremendous results.


BT INFO WILL CONTINUE NEXT WEEK. BECAUSE FRIDAY, AS ALWAYS, WILL BE LIP SYNC DAY AT OLDMTA!

BT WEEK POST 2 - PRIVACY

COOKIES, BT, PII, AND PRIVACY

Before we go on, it’s important to take a little detour into the topic of privacy online. The Internet was never designed to be an extremely anonymous or information secure environment. It is a medium of information sharing.

The Federal Trade Commission of the US Government has made it crystal clear that the owner of a PC or other connected device is the controller of what information is shared by that device with marketers and other organizations. Web sites essentially need your “permission” to place a cookie on the machine or extract cookie data or other information from you.

It is relatively easy to refuse to accept any cookies on your machine. It’s a simple setting change.
But here’s where the plot thickens. Most websites require that the information sharing between you and them be a two-way street. In order to view all of the content or use all of the functionality of a site, your PC must be set to accept cookies. The sites require it. If your PC refuses to accept cookies, then you will be refused some or all experiences and information on that site.

You must be willing to share SOME, BUT NOT ALL, information in order to use the web. Reputable comapnies that use cookies generally agree not to collect or store “personally identifiable information” (PII) about you in their data collection and usage efforts unless they get your permission. Your name, address, social security info, and other highly personal info are not fair game without your explicit permission. Anything that uniquely identifies you as a person is off limits. To a BT platform, you are simply an anonymous individual that clicked on a Lincoln Navigator ad. Among other things.

According to Wikipedia, the following types of info are PII, and cannot be collected or shared without permission:

• Full name (if not common)
• National identification number
• Telephone number
• Street address
• E-mail address
• IP address (in some cases)
• Vehicle registration plate number
• Driver's license number
• Face, fingerprints, or handwriting
• Credit card numbers
• Digital identity

Personal info that is generally NOT considered PII include:

• First or last name (if common)
• Country, state, or city of residence
• Age, especially if non-specific
• Gender or race
• Name of the school they attend or workplace
• Grades, salary, or job position
• Criminal record

Reputable sites don’t take your PII. Disreputable parties might. This is one way that identity theft occurs. A cookie sent by a reputable company (for example, a major ad network) is generally acceptable to PC protection software from McAfee and others. A program designed to take other information is blocked. Or usually is. ;-(

But of course the water here is a bit murky because if you collect enough non-PII, you can probably identify a single individual if you want to. For example, I am a single, Scion XB-driving, homeowning Oakland CA resident that buys 4 books per week online, generally buys green, travels 12 times a year for business and 6 for personal (tickets purchased online.) I use Colgate, have a Sonicare I bought online, prefer foreign language films and buy my tickets online, and have a penchant for purchasing DVDs of sit-coms from the 70s and 80s. I send flowers online about 8 times a year. I have a Facebook profile, a MySpace profile, and a Linked In connected to hundreds of people. That, I am guessing, describes exactly one person in the US.

But truth be told, if someone really wanted my identity, there are far easier ways to get it than exhaustive analysis of online behavior.

And companies active in BT ARE NOT collecting PII. In fact, the Network Advertising Initiative (NAI) has just proposed a set of voluntary guidelines that govern what can and cannot be collected. They are currently seeking public comment on these standards.

There are a few other issues to consider:

1. How long do companies keep my data? Does the consumer have the right to expect that their data expire after a reasonable space of time?
2. How well protected is my data? Can it be easily stolen? Again, not such a major problem if there is no PII
3. Is the privacy policy of sites that collect (or allow others to collect on their pages) BT data written in a manner that is compliant with FTC standards?
4. BT can be argued to be a customer service when it increases the relevance of advertising on a page. But what if the BT is used to vary prices, so that, for example, a person making $65,000 a year is charged more for a book than a person making $45,000 a year. Few would consider that a beneficial service.
5. What dangers are there in collecting certain kinds of BT, like health info? Can this info later be used to deny the consumer insurance? Or eliminate him from consideration for a job? Of course, given that PII is not being collected, these worries are probably unwarranted. But some people still have them, and that is an issue for the BT industry.
6. Are the BT data stored in your cookies correct? If I search for a cancer site on your computer, it’s not MY cookie that is getting credited with this search. Again, in a world where PII are not collected it won't make any difference to the consumer -- but it does point out a limitation of BT.

At the risk of passing the buck, it should be noted that these BT privacy questions are really part of a larger issue of privacy in America, and the extent to which Americans have a legal right to privacy. Do we? Jurists disagree strongly on the extent to which we do. The Fourth Amendment states:

The right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated, and no Warrants shall issue, but upon probable cause, supported by Oath or affirmation, and particularly describing the place to be searched, and the persons or things to be seized.

But privacy rights are going to be THE defining issue for the next several decades in US Courts given the rise of digital media and the ability of the government and other organizations to easily collect, analyze and use information about us.

In the EU, a single difference in the way BT is regulated creates a far different degree of privacy concern. In the EU, users must OPT-IN for BT, whereas in the US they must OPT-OUT of many forms of it. The new NAI standards actually change that a little -- some kinds of BT data are considered sensitive enough to warrant requiring consumers to opt-in. The rationale for the US decision is that the OPT-OUT method results in far higher potential ad revenues for US content providers, meaning that consumers get access to more and better content supported by this higher level of revenue. US sources argue that EU residents are disserved by the OPT-IN policy there because less content is available to them. Naturally many other US sources and EU sources disagree. But so far the FTC has stated that OPT-OUT is the right policy for the US.

I agree with that decision. Consumers have "free" access to trillions of pages of info. Somehow that needs to get paid for, and so long as there is a strong wall built around PII, I think any technique that improves online revenue is a good thing.

Let the flame mail begin!

Thanks for reading, and don't forget to write.

NEXT: THE PROPOSED NAI STANDARDS

BT WEEK BEGINS!!!

This week I will be discussing some of the nuts and bolts as well as noteworthy issues related to behavioral targeting (BT.) BT is hot, both as a marketing technique and as a topic for discussion, particularly among those who debate privacy issues. So let's get the ball rolling here at OLDMTA!

WHAT IS BT (IN ENGLISH)

Behavioral targeting is based upon the truism that what we do can help predict what we plan to do and will do in the future. The idea is that there are certain kinds of online (and other) actions that can improve the likelihood that an individual will respond to digital marketing activity.

An example: most people who are going to buy a car in the near future will not do so on impulse but rather will:

• Research cars online in car sites
• Research cars in car content on portals and other sites
• Search for cars and cars information
• Click on car ads
• Talk about cars on their personal pages, blogs and email accounts
Etc.

By monitoring the behaviors of car searcher individuals, companies can better understand what will interest them, making them more likely to respond to digital auto marketing activity.

So that's BT in a nutshell. BT collects that info and uses it to target marketing efforts to more likely responders.

WHY IS BT “HOT”?

The power of such information can be enormous. If you consider that about 150 Million Americans over 18 are online, but that only about 4.2 million Americans are “in market” for a car in any three month period, it’s natural that automakers would be interested in weeding out those individuals so as to concentrate their marketing efforts on these people.

Contextual targeting (meaning, placing your ads and other marketing activity in related online content) is a start. By placing a car ad on a page about cars, you have SIGNIFICANTLY increased your odds of finding the in-market shopper. The more precise the contextual targeting, the better. For example, placing a hybrid ad on a page about hybrid cars helps out a lot. It stands to reason.

Demographic targeting can also help. If you know that affluent Northern Californians are more likely to buy a hybrid, targeting ads to those individuals can further pinpoint a company’s ability to reach likely buyers.

Psychographic targeting can also help. If you place a hybrid ad on a hybrid car review that is published on a “green” site, you’re getting even closer to perfect targeting.

Context and demography and psychographics have been tools for marketers for decades. But the two way nature of the web has added an additional tool that appears to trump all of these classic approaches.

TWO WAY IP-BASED MEDIA

Every Internet connection is made from two ends – your ISP and a digital device, like your PC. Your PC has a distinct “address” that it shares with no other machine. That address is called an IP Address. Since your IP address is yours alone, the flow of information through that address reflects your online behavior.

As part of the effort to personalize your Internet experience, companies use “cookies” to “remember” who you are and what you like. Cookies were originally developed for sites to maintain “shopping carts”. They are now used for a much broader range of personalization and tracking purposes, among other things. A cookie is a small text file that is placed upon your hard drive when you visit a site. If you have a My Yahoo page, for example, when you request info from the My Yahoo server (e.g., to load your personal My Yahoo page,) your PC pairs your request with the contents of the cookie text file to tell the server what your preferences are. The My Yahoo server can then provide the content you selected for you’re my Yahoo page.

Another kind of “cookie” is a tracking cookie, which is placed on your computer when you take a certain action online. For example, when you click on a banner, a tracking cookie is placed upon your computer to record what you did. That way, if you do something that a marketer wants, like make a purchase, the marketer will know what marketing activity drove your action. The marketer wants to know what drove your action so she knows what works in the marketing plan. The banner that drove your action gets a “point” based upon your subsequent actions.

You might think that the cookie goes away if you click but don’t buy immediately. Usually it doesn’t because most people take action long after that first click. You see a Lincoln navigator ad, click, look at a video, and then go away. But perhaps two weeks later you decide to get a quote from a dealer and head back to the Lincoln site on your own. If you request that quote, the banner you originally click on will usually be credited for that action if you perform it within a certain period of time – often 30 days.

But there’s more to the story, because the company that served the ad can also record your action, and use it to better understand who you are and what you are interested in. Suppose, for example, that you clicked on that Lincoln Navigator ad, then later went about your business online. The company that served the ad could use the info to categorize you as a luxury SUV shopper, which would be important info both for Lincoln and any other car company that sells luxury SUVs. Since clicking on that Lincoln banner makes it FAR more likely that you are in-market for a luxury SUV, your eyeballs are worth more to Lincoln and any other car company than can purchase ad impressions in front of you. So the ad server can sell those ad impressions for far more money.

Naturally, the more info that a company has on your behavior, the more it will know about what interests you. And the more it knows, the more likely it can determine that you are a prime audience for a multitude of companies in a multitude of categories. That means your eyes are worth more to a variety of companies in a variety of contexts.


TOMORROW: THE TECHNICAL MECHANICS OF BT (FOR DUMMIES)