What follows is an anecdotal, non-scientific set of personal observations about internet advertising, which forms the core of my queasiness about Facebook’s business model, and those of a variety of other businesses that rely on advertising revenue. The adage is that half of all marketing dollars are wasted, but nobody knows which half. The promise of internet advertising has been to: reduce the amount spent on advertisements with little to no probability of success, and to use superior data analysis and feedback-testing to improve the probability that an ad displayed to a customer will result in a sale. I’ll touch today on a couple of developments that have already lived up to their promise in these regards, but in my next post I’ll also share some reasons why I’m pretty skeptical that the ‘social graph’ will be as much of a game-changer as conventional wisdom seems to believe.
First, some positive developments. The brilliance of pay-per-click advertising is in its name: the advertiser doesn’t pay anything unless you follow their link, at which point you’re already somewhat of a ‘warm’ lead. Google takes this two steps further, first by serving up ads alongside your search results for specific keywords (so there is a chance to promote a product at the moment when you’re particularly open to it) , and second by embedding advertisements in content you read on other sites that ‘distribute’ Google’s AdWords. If I tell Google I sell cat sweaters and provide some other parameters to them, Google is brilliant enough to place my ads on a bunch of cat fanciers’ websites without any further effort from me. This is orders of magnitude better for me, and probably for the sites on which my ads are displayed, than if I had to track down each of those websites, analyze their likely value to me, and negotiate a deal with them.
Google, too, knows a lot about me. Presumably its algorithms have crawled through most of my e-mail and search traffic for the last five or more years, and it has potentially gleaned some insight about my tastes and preferences from the blogs I read on Google Reader or the products I’ve occasionally purchased through Google Checkout. So it is extraordinarily good at delivering ad content to me that I didn’t even realize it knew I wanted. When I prepared to move across town several months ago, I was perplexed by how many websites (not to mention banner ads on my Google Reader feed) were for moving companies — until it dawned on me that this was exactly the sort of thing Google should be best at: figuring out from my e-mail traffic and search queries that I was likely to need a moving company. Sure, they guess incorrectly from time to time, but they’re able to take much better-educated guesses than mass market advertisers. (For example, the ones whose commercials for Plan B and acne cream I sit through while watching Jersey Shore…)
Google is not alone in its (at times, creepy) excellence in figuring out how to target ads. The Wall Street Journal has had some excellent coverage of what companies know about us based on our internet habits and how some of them attempt to monetize this information. In theory, the ‘social graph’ should support the next evolution of targeted advertising: people voluntarily disclose tons of data about their preferences; and even if someone doesn’t, a couple of her friends probably do, which should make for a pretty good starting assumption about what she likes. Yet, I’ve used Facebook essentially since its launch at a handful of colleges, and I’ve never clicked on a single advertisement, nor even thought about doing so. Is this just because Facebook is still a nascent platform for advertising, or might it be easier to monetize this data in theory than in practice? I’ll discuss my thoughts on why it may be the latter, next.