Last week you wanted to buy an air ticket. You looked at the price on the airline’s website, searched the net for a better deal, then returned to the airline website. You were surprised to find the price had increased. So you booked quickly, before it went up again.
You were conned. When you first visited the airline website it made a note of your computer’s IP address (1), or placed a cookie (2) on your web browser, allowing it to track your movements on the web, and identify you as a likely customer with a strong desire to travel. When you came back to the airline site, it recognised you, and merely had to raise the price to convince you to conclude the transaction. If the flight had been almost empty, the price would have gone down.
Regular visitors to Amazon websites feel as if they were kids in a toy store that knows exactly what they like. A shop that knows your tastes better than your friends do is difficult to resist. The mechanism, “collaborative filtering”, is not new. First developed at MIT (Massachusetts Institute of Technology) in 1995, it uses algorithms to group together people with a similar browsing history and consumer profile. When you click on the albums you want to buy, you may think you are the only person who likes Serge Gainsbourg, and the operas of Rameau, and Metallica, but you would be wrong. Thousands of people have probably bought those albums together before. When these data are logged and analysed, the retailer can calculate your preferences. Then all it has to do is suggest items that your “clones” have already bought.
Amazon’s average annual sales in the US rose from $160 per customer to over $240 after it adopted this system. Back in 2006 more than 30% of its sales already came from recommendations (3). The system is so efficient that the website will even ask if you are buying something for yourself or as a gift, to make sure the purchase does not corrupt your consumer profile.
This is the world of personalised marketing, which gets you to part with your cash by knowing what you want. In the 1980s, marketing experts used consumer surveys to identify market segments, such as housewives over 50, or professionals under 35 earning over $32,000 who played tennis at least twice a month. It was like fishing with a drift net after sonar had identified a shoal of the right species. But retailers are no longer interested in a market segment; they’re interested in you: shooting fish in a barrel.
Help from Facebook users
They benefit from your regular and often unwitting cooperation. More than a billion Facebook users freely hand over information that a few years ago no reasonable person would have given a salesman: socio-demographic profile (age, gender, education, city of residence), musical tastes, friends, photos, plans, dreams and aspirations. There is also your browsing history, and your online behaviour — the tracks you leave when you browse the web or use an application on your phone or tablet. Everything you look at, read, listen to, download or buy, is recorded.
The last time you downloaded an application to your computer or smartphone, did you bother to read the terms and conditions? If so, congratulations: you are one of the 3% who did. If not, you agreed to be spied on, and for your information to be passed to other companies. That is what you agree to every time you click “accept” in your haste to download an application more quickly. It may be morally questionable, but it is legal.
Free applications and services like Google Maps, Hotmail, Facebook and Instagram are all Trojan horses providing retailers with the information to adapt their offers. This is done by search engine optimisation, which aims to show the right ad to the right person, mainly via search engines. It is boosting Google’s profits: in 2012 the multinational’s free services brought in more than $32bn in ad revenue.
Risks for advertisers
‘Such targeting does carry risks for those selling advertising space, because return on investment is measured more rigidly. Ten years ago, the price of an ad was based almost entirely on the number of customers it would reach, whether it would interest them or not. Today more than 20% of advertising industry revenue is based on click-through rates (4): the advertiser doesn’t pay unless the internet user clicks on the ad, which is why it is so important to know the customer.
Another new approach is affiliate marketing, which involves an advertiser paying, in full or in part, according to how many sales an ad generates. So if you click on an ad and buy the product, the advertiser pays a commission to the website, search engine or social network that made contact with you. You can see this at work on some airline websites, when you buy a flight and the website suggests booking a hotel. This approach, and pay-per-click, are replacing traditional ads paid for on the basis of online display — the number of people who see them. But an approach that focuses on an ad’s immediate effectiveness has been resisted by some professionals who do not want to see their role reduced to making instant sales.
Now suppose you bought the product two weeks later from a shop. Is any commission due? The answer to that question is yes, because it is getting easier to follow a consumer profile in the virtual and real worlds, through such tools as Google Wallet. This apparently simple application is meant to take personalised marketing to a new level. You download it to your phone and key in your personal information and credit card number or bank account details. When you get to the till, the shop picks up your phone signal, and once you have authorised the transaction, it is complete (5). The retailer doesn’t have access to your financial details, they are kept by Google which debits your account to pay the retailer, and so learns more about your buying habits.
Google may not yet be selling ads on this basis, but there is no doubt that it understands the potential of its technology. Bringing together information on what ads have been viewed and what sales they led to in a shop or online makes life simpler for the consumer while increasing the chances of a purchase.
Jacques Nantel is a professor of marketing at HEC Montréal.
Ariane Krol is a journalist with the Montreal daily La Presse.
(1) An Internet Protocol address is a unique numerical label assigned to every computer connected to the internet.
(2) A small information file sent to your computer when you visit a website, containing data on pages viewed, the amount of time spent there, etc. The information is stored on your computer, but it is possible to erase it, or set your computer to limit cookies.
(4) Interactive Advertising Bureau and PricewaterhouseCoopers, “IAB Internet Advertising Revenue Report: 2012 Full Year Results”, April 2012.
(5) Microsoft tried to introduce a similar system for online purchases more than 10 years ago, called Microsoft Passport.
This article appears in the excellent Le Monde Diplomatique, whose English language edition can be found at mondediplo.com. This full text appears by agreement with Le Monde Diplomatique. CounterPunch features two or three articles from LMD every month.