Commentary
Behavioral Targeting: Putting Lipstick on a Pig
- by Jaffer Ali , August 10, 2009
All the lipstick in the world can’t make behavioral targeting any less porcine.
Before we go any further, let’s define what we mean by behavioral targeting since it means different things to different people. For the purpose of this article, the FTC definition will suffice:
"... the tracking of a consumer’s activities online – including
the searches the consumer has conducted, the web pages visited,
and the content viewed – in order to deliver advertising targeted
to the individual consumer’s interests."
For all the hype, behavioral targeting has such limited utility that one can only shake one’s head in bewildered disbelief at its growing acceptance. The army of mathematicians that once were lionized on Wall Street has invaded Madison Avenue. Computational finance wizards now seek to become computational marketing mavens.
They are finding a receptive audience among agencies, VCs, industry trade publications, ad networks and marketers across the country. Desperation has a way of bringing people together. But there is a major problem; the same underlying math that destroyed Wall Street underlies BT and will destroy online marketing (indeed, BT-inspired online performance has declined more than 95% over the past decade) . The adherents of behavioral targeting seem to be blind, deaf and overwhelmingly dumb to the metaphysical issues surrounding BT and to the existential implications confronting it.
Consider some of the specious and/or fallacious claims of BT proponents:
1) Human behavior is rational.
2) Future behavior can be predicted on the basis of past behavior.
3) Stalking, collecting and measuring minute surfing behavior is neutral and does not influence the behavior being measured.
4) Behavioral data cannot be traced back to the individual.
5) People actually want relevant ads.
6) Behavioral Targeting will remain legal.
Let’s take the issues one at a time.
1) Our economic models all assume that human beings act rationally. But today, amidst the carnage of the financial system, we see that our “animal spirits” are more highly engaged than reason. Case in point: John Nash’s game theory presupposed rational behavior and he has since recanted his position.
Alan Greenspan said that there was a “fundamental error in their model of human behavior”. That error was rationality. Emotions rule human actions much more than reason or rationality. By developing marketing methodologies that ignore emotion and elevate rationality way beyond what is reasonable, behavioral targeting has a flawed foundation.
2) There are many aphorisms that support the notion of predicting the future by looking back. But try driving your car looking only in the rear view mirror and see how that works for you! This, according to Nicholas Taleb, is nothing more than data bias, sometimes referred to as “data-snooping bias”.
BT allows for thousands of hypotheses to be tested within a very short time frame. The statistical significance becomes more suspect the more we drill down. The idiotic “one-to-one” mantra has led us off the statistical cliff in search of “hidden” relevance. This "science" has produced a performance death spiral with CTRs now hovering around an abysmal .2%.
3) More and more people are becoming aware that their surfing behavior is being monitored. Some do not like it at all. But awareness of the practice can actually CHANGE THE BEHAVIOR. The literature on this is quite overwhelming, supported in large measure by common sense. That is why the FBI practices secret surveillance. They know that once their monitoring becomes known, behavior changes. BT literally affects behavior once those being stalked realize that Big Brother is watching.
4) Many privacy statements of BT practitioners claim that no personal surfing behavior can be tracked back to the individual. With ever growing databases of personal behavior, the potential for matching - even without direct information being stored - is now possible approximately 75% of the time. The matching will get “better and better” (or, worse and worse, depending upon your POV.)
5) Even if we could inject relevancy into the equation, it's a huge leap to suggest that people want more advertising. Marketers like to say, “if we can only deliver the right ad in front of the right person at the right time…” But as audience targeting techniques have become more refined, what has happened to performance? Once again, it bears repeating: our industry-average click-through rate has plummeted to .2%, when 10 years ago it averaged 5%! It doesn't take a genius to conclude that in an on-demand world, nobody demands more advertising, yet despite overwhelming evidence to the contrary, advertisers and marketers still insist that the Emperor is fully clothed.
6) There are many legal hurdles on the BT horizon. Professor Andrea Matwyshyn of the Wharton School said:
“…questions are inherently intertwined not just with privacy
laws, but also with contract law, computer-intrusion law[and]
consumer-fraud law.”
If you are a betting person, how much would you like to wager that the practice of behavioral targeting will remain legal? We all allocate resources where we think they have the greatest chance of success. Given the hurdles confronting BT and the looming legal battles, is this really where you want to allocate your time and money? Or are you just putting lipstick on a pig?
20 comments about "Behavioral Targeting: Putting Lipstick on a Pig".
Read more comments >
Next story loading 
Bravo. I couldn't agree more.
I read this article from top to bottom - and agree with every single point made. Absolutely brilliantly written.
By the way, we call it "feeding strawberries to pigs" -same concept, just another take on it!
Me too. Well said, particularly the last point. Given the trend in "Do Not Call" lists, who would rationally invest many resources in the next unauthorized intrusion technology?
To me, the industry has failed to focus on utilizing new technologies to put the customer truly in the center of the marketing model. Behavior Targeting is using technology to sustain the "product-centric" model that the industry has used since the industrial revolution led to mass marketing: I have a product to sell, and I want to use technology to help me better target who to deliver my sales messages to. The customer's choice is still "Yes/No" . . . buy or not buy. . . redeem or not redeem.
Our company is focused on enabling agencies and marketers to shift to a customer-centric marketing approach, where the interactivity of existing technologies are used to determine how to present an offer which represents a solutions to specified needs, not modeled needs, obtained via interactions with the customer . . .to tailor the offer to the customer's responses . . . and to give the customer the choice of how they'd like to be "rewarded" or incented to respond.
How do you think response rate, customer experience and loyalty/retention/share of wallet would be impacted by such an approach? Right . . .that's what we're seeing.
Same technology can be used with message targeting . . . for returning visitors, a simple "what messages would you like to see" can increase response rates, click throughs, and ultimate sales, as well as return visits, page views and site visit duration.
It all comes down to marketing approaches driven by the value of relationships versus today's transaction, and treating existing or repeat customers differently than we treat anonymous or new customers . . . and oh yes, being willing to REALLY put the customer in charge of the relationship.
Agreed. More and more it seems as though contextually and socially relevant media placements are truly what drive consumer engagement and ultimately purchase intent.
Nice article. An interesting counter-point I don't often hear.
Just one little point on: "our industry-average click-through rate has plummeted to .2%, when 10 years ago it averaged 5%!"
Maybe perhaps that's because of the growth of the online population and the shift of online ads from something new and specialized to a more general market?
I don't think the drop in CTRs can be entirely contributed to the dearth of audience-targeting technologies.
Andrea, good observation. I think it's more likely the case that the same over-saturation of the channel by advertisers is largely driving the decline in CTR. Same experience with TV ads, as the industry increased the ad time per program and reduced ad length, and in Direct Mail, as the quantity of direct mail delivered per day skyrocketed. Anytime a channel gets over-used, response dulls.
The key point is, I think, that even with sophisticated algorithms driving "better" targeting, increased relevance of individual ads was no where near enough to offset the saturation of irrelevant ads being delivered per day (per hour, per minute?). So to Jaffer's point, is that amount of time, effort and money perhaps misplaced?
Well put, Robert.
But I think there is a parallel dynamic at work, that being the ongoing integration of the "on-demand" mindset. While media fragmentation claims its obvious toll, overt consumer efforts to block/bypass/eliminate commercials (Tivo, DVRs, etc., pop-up blockers, spam filters, etc.) is a clear indication to me that given the choice, we don't and won't choose ads.
If you're like most web users I know, you've already trained yourself to avoid eye contact with the right-hand column Adsense clutter. And forget pre-roll effectiveness. Why would anybody welcome a commercial online that they are willing to pay extra to purposely avoid offline? Makes no sense whatsoever as the dismal numbers indicate.
People talk about relevant ads vs. irrelevant ads when the real debate should be ads or no ads.
Although i usually have to read Jaffer's articles twice with a dictionary close at hand, I usually "get it" and agree with him 99.23% of the time.
Well done as always.
This is one of the most inane, bizarre articles on BT that I have ever read. I can't possibly disagree more with nearly all of Mr. Ali's points. Where to begin...
1. Limited Utility - Every targeting methodology, and every marketing channel and approach, has limited utility. That is why smart marketers have a stable of activities - the marketing mix - that they employ.
2. Desperation - I'm not sure who Mr. Ali is speaking with, but the marketers I see and hear every day are not desperate. They have a healthy desire to not waste money if they don't have to, and they are under continual pressure to perform - but that's their job, and I don't know any that I'd classify as "desperate."
3. The broad-brush "fallacious claims" that the author ascribes to apparently all BT practitioners underscores the broad brush view he apparently has of the industry. Too many generalizations here to even begin dissecting.
4. Of course humans are irrational beings. But marketing & advertising is all about placing one's best bets on what activities will accomplish awareness, elicit interest, and inspire action. There are no magic formulas - it's trial and error, no matter what the channel, no matter what the technique. Mr. Ali seems to suggest that since BT isn't perfect - as if any other method is - that marketers should therefore abandon thoughts of using it as best they can for what it can do, and use other tools to accomplish other things. One works with what one has to accomplish the mission. The goal is to use each each technique and channel in as balanced a way as possible in order to balance business needs with business restrictions. But Mr. Ali would have us believe that since aberrations may occasionally occur (the irrational economic situation he uses as a comparison last occurred 80 years ago), marketers should dismiss any attempt to design marketing programs with intelligence that will work the other 99.9% of the time.
5. Mr. Ali neglects to cite his source in saying that in 75%+ of cases individuals can now be positively identified down to the individual level. Regardless, the notion that consumers are just now catching on that they're being tracked is laughable. Of course they're understanding that more now. And they also understand that the reason their mailboxes are stuffed with catalogs at Christmas time is because the fact that they bought from a catalog was tracked. And that the mail they get offering new services is because their residential move was tracked. And so on. Mr. Ali apparently disdains the whole of the direct marketing community with his rant, as the "porcine" practices he derides have been performed in some way for probably 80+ years now.
6. I have never once - in 12 years of offline direct marketing and now 10 years in online marketing - heard a vendor/supplier suggest that if ads can get more relevant that consumers will then be overjoyed to get them. But Mr. Ali here misses the same point missed by nearly every study I've ever seen on BT, targeting, data, privacy, etc. The question always seems to be posed as if people like advertising, or if people like (or are comfortable with) their information being tracked - but the question is always asked in a vacuum, as if there's no tradeoff for them if things were to change. The correct question is more like, "Would you rather a) Have free access to the vast majority of content you enjoy online, the TV programs you enjoy, the radio you listen to, etc. and have some of your information responsibly tracked in the hope that more things you might want or need can be presented, or b) Not have any information tracked and pay full price for all these things you enjoy?" There is a tradeoff. Period.
7. The repeated citation of .02% response rates is not only greatly overgeneralized, it also is misleading because Mr. Ali places no context around it. What are the response rates of typical banners ads? Of video ads that Mr. Ali's company hawks? Of email? Of magazines? Of direct mail? And on and on. I have personally been involved with many campaigns that produced order-of-magnitude better results that Mr. Ali touts, but then, once again, he doesn't cite any study to support his thesis anyway.
To summarize, it'll take much more than lipstick to make this pig of an article pretty. I suggest that Mr. Ali stick to doing what he knows - having users click on old TV spots to then go to an advertiser's site that they aren't expecting, and leave the targeting to the rest of us.
John, there is plenty of room for debate...but one must at minimum understand what debate consists of. Instead of ad hominem attacks, it would be better to challenge the assertions with your light instead of yoru heat.
Be that as it may, agreeing that humans are not rational and then choosing a technique that discounts this factor is an interesting position.
Rather than challenge the Vidsense model, which the BT article did not address, you showed a less than profound understanding of yoru own craft. The underlying statictical, algorithmic methodologies underpinned the financial collapse....not 80 years ago.
But we do agree on one extremely important point. People must place bets on where they spend their money. Your success depends upon them choosing your way,...I have no dog in that hunt because we have chosen not to spend our money (we are marketers as well as media owners) on yur folly and decided not to layer in BT with selling our media.
Ultimately, the article was not about ME OR MY COMPANY. Because you have chosen to make it appear so, you have strayed from a cogent discussion path that could benefit a lot of people.
John,
Interesting that so far yours is the only dissenting voice to make an appearance here. I also note a decidedly desperate tone to your remarks.
At least you have the good sense to acknowledge that nobody demands more advertising, relevant or otherwise.
Your antihesis takes a hit, however, when you suggest that we are not already paying full price for "...the vast majority of content you (we) enjoy online, the TV programs you (we) enjoy, the radio (we) you listen to, etc. Last time I looked, my cable/broadband bill was about $200/month, not to mention my anti-spamware costs, spy-ware prevention costs and Norton Security fees...oh, and the new computer I purchase every few years. Heck, a fella could go broke enjoying all this free media!
You also inadvertantly cite a 0.02% CTR (I think Mr. Ali was referring to the industry-average click-through rate for display ads), when his article cites a 0.2% figure. Talk about splitting hairs! With all due respect to you both, a recent comScore whitepaper chops Mr. Ali's number in half (though it improves your figure by a factor of 5X), to wit: "With online display ads yielding click-thru rates of less than 0.1 percent, advertisers can no longer rely on click-thrus to gauge online ad performance. "
At the risk of rubbing salt in the wound, with average CTRs for display ads at 0.1% (if we are to believe comScore), any discussion of performance is laughable. Although I must say that I feel very secure in the knowledge that we can now measure statistical zero so precisely.
Recent directives from some major brands indicate that patience is wearing thin. Should we pity the poor agencies who are getting cut out of the billable-hour business and forced into the performance business? Or are they only reaping what they've sown? Perhaps BT technology can save them. After all, if it can help them get that number back to 0.2% they can rightly claim a 100% improvement in performance! The only trouble is, lots of these agencies were already using BT to get where they are. Oh what a curious web we weave...
With all due respect, all of this smoke and mirrors is a waste of good lipstick, and does pigs the worldover a tremendous disservice.
Jaffer, point well taken on my reference to your business. No blatant disrespect was intended - rather just an attempt, apparently flawed, to point out that all marketing methodologies have flaws, and in my view it's futile to disparage one in favor of another without objectively noting the strengths and weaknesses of each.
The simple truth is that marketers and their agencies are rapidly and increasingly becoming more willing to "vote with their feet." The channels, techniques, etc. in use today wouldn't exist if there weren't a market for them. And, inevitably, those that don't address marketers' needs while respecting individuals will fall by the wayside. Remember pop-ups? Drive-by downloads? Those and others had their brief day in the sun, but quickly faded to oblivion because they didn't strike the balance I mention above. While they might exist in some small form now, for the most part they have gone the way of the dinosaur.
So, to the extent that BT, or search, or email, or social, or video, or anything else really doesn't help marketers with their mission, and don't provide a better solution than the alternatives, they too will die out.
In agreeing with you on the irrationality of humans, in no way am I suggesting that BT counteracts that. (And, on a side note, if I understand your comment, I don't believe that any algorithms underpinned the recent collapse. In my genuinely humble and layman's opinion, it was greed, plain and simple.) Rather, my point remains that marketers have money to invest in order to help grow their businesses. And like the market, there are a myriad of instruments in which they invest those monies, none of which is perfect or without risk. So the wise marketer/investor will evaluate the options, and probably diversify her investments according to their analysis. Ultimately, the returns speak for themselves.
So to wholly disregard any given technique/investment as bad for all, without taking into account the many variables that exist in the marketing economy is foolish, as far as I'm concerned. One invests, one learns, one makes modifications, and repeats the cycle. This is the crux of the exception I take to your recent writings on BT - that they paint with a broad brush that ends up oversimplifying a complicated environment.
I agree with you that this dialogue can hopefully help others wrestling with their decisions, and welcome further interactions.
Michael, I can't make out several of the points you're trying to make, but for those I can, here goes...
First, as far as mine being the only dissenting voice on this thread - congratulations that the sample size of 8 or 9 is in your favor. Bravo. And your skill in overestimating quorum is much greater than your skill in discernment - there is no desperation on this end, no inadvertent stat mentions, and no wound in which to rub salt.
Your point about the expensive free media neglects to note how much the bill would be if advertising weren't there to defray the costs. Certainly, as choices have proliferated, the cost of subscriptions and devices to access them have increased, but I'd be genuinely surprised if you wouldn't agree that the subscription fees would be higher yet if there weren't ads to offset the cost. For instance, how does the average cost of a given cable channel (cable cost divided by number of channels provided) compare to a monthly fee for just HBO or some other premium channel? Now picture if all of those channels were the same cost. They're much lower because they allow advertising. So while some things aren't truly free, a rational person can't argue that advertising revenues don't make them more affordable. That was the only point I was trying to make.
The delight you take in the .01, or .02, or whatever figure you want to use belies a mistaken notion that CTR even matters that much to marketers any longer when measuring performance. The marketers I work with measure performance on business impact, ideally short-term and longer-term (sadly, not enough on the latter). I don't think anyone breaks their arm patting themselves on the back for CTRs, direct mail response rates, telemarketing answer rates, etc. This is the crux - there is waste in all forms of advertising. And the mission of BT, other algorithmic measures, indeed any new technique, is to help reduce waste. Clearly, with any of these stats, there's a long way to go. But that doesn't mean that when progress is achieved, even to a small extent, it should be dismissed because it didn't change things 180 degrees.
I'm a fan of on-demand - I think it's a great development for all involved. But I still maintain that for the subset of marketers who can effectively utilize it, once they tap out that channel, where do they go for the rest of the budget? They choose the next best thing. And the next best thing after that, and so on as long as budget allows and business goals dictate. None of these techniques on their own can fulfill an entire marketing plan. My challenge to you is to take a budget - say $30 million - and a goal of growing revenue by 5% over the next two quarters (make up any numbers you want along this line) - what do you as the CMO do with that money? How much fully voluntary opt-in, on-demand activity can you buy?
Thanks for keeping the debate alive!
John,
I thought we were debating the value of BT. Increasing revenues is a whole different pig altogether.
John,
Your words: "The delight you take in the .01, or .02, or whatever figure you want to use belies a mistaken notion that CTR even matters that much to marketers any longer when measuring performance."
comScores words: "With online display ads yielding click-thru rates of less than 0.1 percent, advertisers can no longer rely on click-thrus to gauge online ad performance.
My question: Which is it, not worth discussing as you claim, or a virtual failure as comScore claims?
Maybe you guys are talking to different advertisers?
John,
It occurs to me that given your disdain for click-thru as a meaningful metric (which I agree with, BTW), you might consider changing the name of your company.
Michael, we have been debating the merits of BT, but that technique, along with others, are generally only employed to grow revenue for businesses so, to me, it always goes back to that challenge.
I think comScore and I are saying the same thing about CTR - that its value as a standalone measure has greatly diminished. I should say, however, that I am not one of those who think it should be ignored - I think it should be evaluated in combination with other measures, including both branding- and direct-response-related. To ignore it completely, as many now advocate, is to neglect one of the key levers that marketers have to improve their marketing efficacy.
In terms of our company's name, it's a holdover from our humble beginnings as the first CPC ad network, way back when. I'll be sure to look you if we ever change it so you can provide branding advice. :)
Thanks, John. I can always use the business.
BT is not a value proposition for consumers.
There is a way to get very good quality data from consumers, willingly and virtually ubiquitously but it has to be part of a larger 'web of trust' system which defaults to anonymity. Advertisers would be able to get their message to a willing consumer, but the consumer would be guaranteed privacy by a neutral third party. The consumer would benefit from easier, safer purchasing and general web surfing.
Google's dumb idea to match browser to credit score will kill the golden goose for advertisers and ultimately google's business model.
That isn't to say that consumers/purchasers won't like the whole process streamlined to a single click - signal intention to purchase/lease/hire/borrow including credit application, identity check (with the consumer), delivery details, payment methods etc leaving perhaps a single authorisation/sing process.
The successful way to achieve these value propositions for consumers and marketers is diametrically opposite to the methods pursued in recent attempts.
Earning the trust of the consumer will provide further opportunities to personalise the marketers message and perhaps gain a peek into the consumer's current mindset. A system where they weren't required to trust individual advertisers could be more effective.
It is still early days for the internet- the dark ages. It isn't going to go away, and mobiles will integrate into a new version of the internet which has infinite touch-points to the real world. It's time to wake up and take a long-term sustainable business view of the web and that should probably start by taking a consumer centric view.
There is a difference between tracking or spying on consumers and being responsive to their mindset.
Micro-payments will come to pass and there will be an increase in paid content, perhaps lead by newspapers - and their advertising customers. In the process we'll see the physical paper integrate into the online and the physical marketplace follow very quickly too. We'll see the end of cards and all the excess paraphernalia which just complicates or makes the purchasing process unsafe - as we move to a one or two click world.
We all need to remember the wallet is - the wallet, not quite infinite as we recently learned. All the more reason to take a long term view to create trust, it can't hurt your message and will likely guarantee you a share.
I imagine there might even be paid search engines where we can choose to have no advertising in the results, or perhaps an instant comparison of different advertised brands of the product in which we are interested. It has as much chance as the newspapers have of charging for online content.
One thing is for certain we will see a 'web of trust' created and marketers surely need to be part of it.
Strangely enough I see newspapers as having a one shot chance at playing a leading part of a very bright future for advertising/purchasing. TV could benefit too, and who knows, maybe even the RIAA.
The choice is get on board or be dragged kicking and screaming into the 21st century we all know is coming, or even worse - become extinct.
We live in interesting times.
I know I'm a bit late on this one but WOW, this article is simply moronic.
'BT' will not destroy online advertising, BT as you refer to it as will destroy traditional marketing models though. Traditional advertising principles, a model built on waste, do not translate well online for a number of reasons.
Consider a career change.