Over the past three weeks, I've been followed. By advertising. Like many, week after week, I land on dozens of sites. Some visits originate from my set of bookmarks, others from the usual click hopping that defines internet serendipity.
In numerous instances, I get the same ad in different formats. The advertiser is called Litmus. I'm testing it. Since it sounds like a good product, I'm happy to link to it. Litmus is owned by Salted Services Inc., a Cambridge, Massachusetts company founded in 2005. It specializes in mass emailing analytics and optimization. You send it a test email, it previews your layout in numerous mail readers, flags any rendering issue, measures its ability to go through spam filters (a publisher's nightmare for legit newsletters) and provides incredible analytics. (In the Monday Note’s case, these analytics are extremely encouraging and a good incentive for Jean-Louis Gassée and I to continue ruining our weekends.)
I've been testing Litmus for three weeks now, on advice from my friend Kim Gjerstad, a great WordPress and emailing specialist. Now, Litmus wants me as a regular customer, and they are stalking me all over the web. Fair enough. How do they do it? They -- in fact, the digital marketing firm they hired -- installed tracking devices in my computer. Since I subscribed to their service and since I'm professionally transparent on the internet, Litmus has been able to reconstruct my complete profile. Therefore, each time I visit one of its thousands of affiliates, the site will identify me as a potential Litmus customer and serve me the right ad. (As I write this, I discover I'm also targeted by Litmus competitors -- some internet marketing arms merchant is making money on all sides).
After a while, I realized how saturated I was with Litmus ads when my synapses (slowly, I admit) finally added-up the number of pages I saw carrying the company's rainbow logo. At the same time (but unrelated), I noticed how my advertising environment surreptitiously changed, depending upon the "freshness" of the browser I used. On one, I never delete cookies; on another I regularly flush tracking devices.
Tracking internet users is nothing new. But I decided to give it a closer look as I was doing research for my group of journalism students at Sciences-Po in Paris. I directed them to the extraordinary Wall Street Journal's What They Know series, a compendium of about 20 articles, produced over the last 12 months by a team of reporters. By the way, to nail a point further down: their coverage exemplifies the superiority of traditional media outlets when it comes to diving deep down into a subject that will require months of work, the implementation of complex methodologies, the hiring of data mining specialists, etc.
Journalistically speaking, it's Hermès saddlery compared to the Huffington Post’s leatherette.
In addition, the WSJ journalists didn't spare their own boutique: they revealed the Journal was also involved in serious tracking...
Just to frame the discussion, here are some of the WSJ findings:
- A group of 50 sites that represents 40% of the US page views installed 3180 tracking files on the test computer used by the Journal.
- On average, each of those sites installed 64 files. Dictionary.com puts 168 trackers, MSN more a hundred. The Wall Street Journal (which doesn't rank among the top 50) installed 60 tracking codes. Only Wikipedia installed none.
- Out of the 3180 trackers detected, 2224 came from131 companies that are essentially involved in consumer databases and profiling.
- Not surprisingly, Google, Microsoft and Quantcast lead the pack.
- These trackers are increasingly sophisticated and powerful: a single visit to a US bank leads to the injection of 5000 (!!) lines of code into the user's computer.
- The notion of anonymity is a joke, see what one of the articles says:
Some of the tracking files identified by the Journal were so detailed that they verged on being anonymous in name only. They enabled data-gathering companies to build personal profiles that could include age, gender, race, zip code, income, marital status and health concerns, along with recent purchases and favorite TV shows and movies.
In other words, once you get such granularity, connecting to a name, address and email is almost trivial.
Below is a short abstract of a tracking device obtained by the Wall Street Journal. Related article here.
Where am I heading to with this? Two things.
1/ Excessive tracking will inevitably backfire. Internet-boosted profiling will directly impact users’ daily lives in a broad way. Take insurance or banking. Zillions trackers have determined with a great deal of accuracy that Mr. Smith, living in Kalamazoo, Michigan is: (a) slightly overweight; he tried numerous diets and pills, bought home training equipment, and visits weight-loss forums -- anonymously he thinks; (b) he has little debt, but he is in a precarious jobs situation; he often visits jobs post sites; he works in the heavy machinery sector, one that is listed as fragile; (c) he lives alone (Smith is tagged as divorced, with grown-ups kids); (d) then he tends to booze and smoke a bit (all of this known thanks to a look at his daily purchases, courtesy XYZStore where he's an identified coupon-redeemer), etc, etc.
What fate awaits Mr. Smith? He'll be struck by an insurance rate hike, and he won't be eligible for a loan to remodel his house. For him, data gathering translates into a tangible cost.
As for Mrs. Jackson, a lawyer in Northern California, she is in a exactly symmetrical situation with a good job, great health and decent credit. Her insurance company and her bank see her as a low-risk, valuable customer and she will accordingly be pampered with low rates.
The result will be the widening of a data-driven gap between the middle class and the upper crust: Mr. Smith will see his income erode as Mrs. Jackson will become much more relaxed (at least financially speaking).
This little tale is by no mean fiction: a firm such as data-collector Acxiom says it collects 3 billion pieces of information per day on each and everyone one of us.
2/ The value created in the processes I described is unevenly distributed between the advertisers, the digital intermediary and... the vector for the ads. And here I get back to my beat, which is the morphing of media business models. According to Dow Jones VentureSource, between 2007 and 2010, venture capital funds have invested $4.7 billion in 356 online-advertising firms, a 29% growth. Most of these firms are in the tracking and data-collection business. For the same period, according to the Newspaper Association of America, the amount of advertising spent in all US newspapers went from $3.166 billion in 2007 to $3.042 billion in 2010. On the one hand, more cash to fund the increasing ad sophistication; on the other, an actual decline in ad spending for the news sector -- which continues to suffer from eroding CPMs (Cost per thousand page views) and limitless inventories.
It looks like there is a slight imbalance, here. Can it last? I don't think so.
Oh, by the way, should I decide to buy the Litmus service, it will be on the sole basis of the test I ran. The word of mouth about it is quite good and it wasn't necessary to saturate my window on the world wide web to convince me.