Have a Laugh with Online Tracking
You know that online marketing firms are anonymously targeting ads at you based on what they think they know about you from your online habits and how those fit into narrow profiles.
If you want to find out what some of these companies think they know about you (and be amused or surprised by the results), try the following. This works for online ads with the "AdChoices" logo on the top right (
- When you see an ad with the AdChoices logo on it, click on the AdChoices logo.
- You'll see a block of text from AdRoll that describes what they're doing when it comes to racking. Read it.
- Click on the "More Information and opt-out options" link. This will take you to a page managed by Evidon, a company that offers webmasters and advertisers a way of giving users options to opt-out of tracking and retargeting. The Evidon page will list a number of online advertising companies that have chosen to work with Evidon for opt-out maagement. You might see a different list, but my list shows AdRoll, BlueKai and eXelate.
- For BlueKai and eXelate, you can click on "Manage" and it will display a shadowbox with a lot of demographic markets and marketing subgroups. The ones that are selected or marked are ones that they've put you in based on the average profile of the average visitor to websites with whom they work that you visit, or other assumptions they've made about you based on the location of the IP addresses you frequently use, or on marketing profile info you might have actively provided to Facebook or anyone else that may have asked you specific questions about your tastes and preferences. You can modify these to tune the information that they have on you so that the ads that you receive are more aligned with what you care for -- which might not actually be what you are, but what marketers think people in the groups that you indicate actually care about.
This in and of itself is interesting. But you might now be creeped out and wondering how the hell they either know this much about you or else could have it oh, so wrong.
To understand it, it's good to know what these companies do:
AdRoll is known as a "retargeting" company -- they provide a platform that advertisers can sign up for. The way retargeting works is a bit convoluted, but I'll try to do my best to explain.
Retargeting for conversion
Online businesses that look to make money via something other than advertising do so by having some kind of process in which users actively give them money -- through subscriptions for service, virtual transactions, or actual purchase of physical goods.
When you visit a site and you go through the process of giving them money, you have converted from a window shopper into a shopper -- otherwise known as a "conversion". Retargeting is a process of generating conversions by putting up ads on other sites that remind you of the site you just visited.
Say you are in the market for window shades. So you spend an afternoon going to TheOnlineBlindStore to look at their patterns, BlindsForMore to do price comparison, and BlindsByYesterday to see if there is a difference in taxes or shipping that is advantageous for you. All three of these websites want you to buy from them. Likely, only one of them will get your money. But not today, since you want to discuss the options for the blinds with your spouse, even though it is a futile endeavor because you're hitched to someone that can't tell the difference between Snow in Berlin pattern and Chalk in Flock pattern.
Say TheOnlineBlindStore uses retargeting via AdRoll and the other two don't. That means they set a cookie on your browser with the magic AdRoll ID for TheOnlineBlindStore that identifies you as person who has not yet bought blinds from them.
So tomorrow, when you're looking at Facebook posts, or clicking on a HuffPo article, or reading that funny BuzzFeed thing about the "21 links you must click on before your cat eats you", you might see and ad for TheOnlineBlindStore. They might even offer a 20% off shipping discount, or a special "free consultation" or some other sweetener to get you to click on the ad and go to their store and buy the darn blinds. When you do, then they have succeeded with retargeting.
Retargeting for referral dollars
Online businesses that look to make money primarily via something other than you actually giving them money (i.e. advertising) use retargeting as part of a more complex network of referrals and as a way of getting higher rates for their ads.
TheOnlineBlindStore might be willing to pay a few cents for every thousand people who see their ad under the assumption that 900 of those people really don't care about blinds, of the 100 who do care only about 10 are actually in the market for them, and of those only 1 will actually go through and buy blinds on the website.
What if someone who could tell TheOnlineBlindStore that they would bring them those 10 people who are in the market for blinds and actually get them to buy their blinds. Would they be willing to pay a 2% or 3% commission to whomever gets them to by the blinds? Likely they would -- when running a marketing group with a marketing budget, it's usually easiest to justify your existence if you can prove to the finance people that for X amount of investment you're getting Y type of directly attributable return. That kind of hard proof is hard to conjure up for non-targeted advertising, because it's harder to know if the 2% uptick in sales that you saw last month was because of your $10k spend on radio ads or because it's summertime and blinds matter when the sun comes out early. Justifying your choices constantly to the Money People sucks, and soft answers of "our ads work half the time, but we don't know which half" usually get you eyerolls. Marketers don't like eye rolls.
So TheOnlineBlindStore sets up an affiliate program where they offer 2% commission to any members of the program who can successfully refer an online customer to the site who completes a purchase.
Say you like going to a site called Greenist.com, a website full of "Green and Eco Friendly Design Tips", which belongs to a large multi-website network of lifestyle blogs, Hawker Media. When you visit Greenist (which whom you probably have a non-passing relationship such as a registration or an account), they generally already know something about you. This could be because you told them, or because they assume things about you. Likely, if you're visiting Greenist.com you're 25- to 40-something, 55% chance that you're female, probably with some degree of college, and likely to be either a homeowner or in some kind of redesign process. Now, you might in fact be a 65-year-old male who started out building cob houses back when the brown acid was in style but now drives a Suburban to the vegan coop and the gun shop, but you're an outlier. Your ads will always be weird, but you're likely somewhat fun to hang with.
Say Greenist is a member of TheOnlineBlindStore's affiliate program. Because of their core demographic, they know that running ads on their site for TheOnlineBlindStore's new line of organic unbleached free trade cotton-and-rattan automated shades is likely to generate a click. But they want the conversion! So they'll use retargeting to put ads for TheOnlineBlindStore for you on every other site that is part of Hawker Media's empire -- knowing that if you actually click on the ad and buy those blinds (they are gorgeous), they'll get a fat commission check.
Retargeting can be very, very aggressive. It can also be very successful, particularly if the folks showing the ads are good about matching ads with customers.
BlueKai is a really large data aggregator and profile generator. Basically, they stitch together data from bricks-and-mortar loyalty programs (that 10-for-1 special at FoodCo when you use your shopper card isn't actually free, boss), large scale online profile aggregators (gigantic or otherwise generally large enterprises, forums and chatrooms existing basically only only, kids), and other huge data sets to create fairly detailed marketing profiles consistent with specific usage patterns and demographic profiles. Within this process, they have little interest in anything that actually links anything specifically to you. They don't really care about the individual because individuals are messy and weird and have their own special little variations that make things difficult to gauge. Large data sets are the key here -- values in aggregate paint a fairly accurate picture of an archetypal strawman, and marketing is all about an archetypal strawman, not your own eccentricities and quirks.
Once these profiles exist (and they constantly get refined), they can assign individuals using anonymous tracking ids based on whatever assumptions they can make about those individuals. For example, BlueKai thinks I'm a 50-something Aspiring Urban Early Adopter Female Head of Household who is not Living Well right now.
There's something to be said about the names of the categories -- they're deliberately precious in their naming and almost get weirdly euphemistic -- but they do reflect that they have enough data to move from very coarse age-gender-and-race-based assumptions to ones more reflective of the variety of the human condition (as seen from the side of peddlers of goods).
Because they surmise from your that you're likely a 35-40yo male who likes modern design, lives in a densely populated metropolitan area, visits left-wing and sports blogs, is politically engaged and reads websites in multiple languages, they'll then put you in a demographic profile that advertisers have linked with being someone likely to go on frequent Atlantis Cruises. No, I'm not at all speaking from personal experience.
Much like BlueKai, eXelate builds profiles and analytics based on user online activity. Marketers use this data both to make guesses as to what kind of people are more or less likely to perform x-or-y online activity and tho choose who to target with an ad. Online portals with ad inventory use this data to determine what kinds of people actually visit their sites, which allows them to choose ad types or advertisers that are targeting the kinds of viewers they get on their sites.
All in all, once you start delving into this world you start learning a couple things:
- Using online tracking, it is possible to learn a lot about you
- Using online tracking, it is possible to get it completely wrong about you
Opting out of online tracking won't make the ads go away, but it makes it so that it becomes more difficult for someone to put you in a bucket. They'll still try, and the less info they have the more wrong they'll be. Now whether or not it is desirable for you to be in the right bucket or the wrong bucket once They get access to your profile and They decide to start using it for Evil is really based on you and your circumstances (if you're associated with a profile that corresponding to a 20-something urban upwardly-mobile lesbian, in the USA it makes you part of a prized marketing demographic. In Russia, it could lead to your arrest).