Reading Facebook's tealeaves

With Facebook’s disappointing appearance on the U.S. stock market, it’s as good a time as any to look at how they plan to make money. This is interesting if you are an investor for obvious reasons, but it should also concern you as a user: it reflects how safe Facebook is as a social investment for the long term and how desperate they need to be to make money from you.

A billion deeply engaged users feels like something that should be easy to make money from. The problem gripping all social services is that users are engaged with each other rather than the service they are using. This is one reason it’s hard for Facebook to raise revenue: its users don’t really care about Facebook of itself.

Facebook’s first quarter of 2012 saw increasing revenue. Notably, this increase in revenue was eclipsed by the increase in costs. While it could be a one off problem, costs increasing faster than revenue is a warning sign. The next few quarters will be interesting.

While the numbers tell their own story, more interesting is thinking of Facebook’s chosen business model and how it works for them. Facebook have chosen to become an advertising company, which has served Google well over the years. The question is whether Facebook is as suitable an advertising platform as Google.

Firstly, it’s clear that Facebook’s users find it compelling. Unfortunately, it’s clear that advertisers don’t find their side of the service equally compelling. Facebook’s quest is to narrow this gap.

Google’s strength in advertising comes directly from the fact that it’s able to place adverts in front of you at precisely the right time: when you are already searching for something. There’s a far better chance you’ll click an advert if it’s likely to lead you toward a goal you are already searching for.

In contrast, Facebook’s current adverts appear when the visitor is already deeply involved in a task unrelated to the adverts shown. Compared to Google, the adverts are almost always an irrelevance to the task at hand. The click through rate must be orders of magnitude lower. While this is a similar story to most advertiser-funded sites on the web, Facebook have an apparent ace up their sleeve: the almost unimaginable amount of data they have on your life.

The core question, then, is whether this information is of any use in serving up a relevant advert at any particular point in time. Can Facebook analyse that data in such a way that it indicates what you are likely to be tempted by at that moment?

Loyalty cards, particularly for supermarkets, offer up a wealth of data on buying habits. Stores are able to run fairly simple analyses on the data exposed by them to predict what’s worthwhile sending you a coupon for in the post. The analyses are simpler becase they have your actual buying habits so the inference required is fairly minimal. These cards have apparently been very successful for businesses, judging by their proliferation.

Facebook’s aim is similar to these cards: inferring possible future purchases from current information. Facebook’s problem is that the data they have is far less directly connected to purchasing decisions than loyalty cards provide. What does the fact that you’ve uploaded photos of your sister’s wedding suggest you might purchase? Not that much, I would wager.

On the other hand, some of the data Facebook has is very valuable. Facebook Places is a case in point. Places allows Facebook to offer hyper-targeted advertising. The bar I regularly visit could advertise it’s upcoming events. A restaurant could offer vouchers. Businesses currently approximate this with Facebook Pages — are there ways Facebook could offer more useful things here? I can see opportunities for both Facebook’s users, its advertisers and Facebook itself. The opportunity to provide an intensely personalised and useful experience seems to be there, providing enough benefit for Facebook’s users to outweigh the privacy costs for many.

What seems to be happening, however, is that Facebook are turning to ever more invasive and questionable methods. Misappropriating a Like an an opportunity to sell endorsement is a PR disaster waiting to happen. I’m very pleased to see this is something appearing in the more mainstream media, and also interested in people’s views on being used in this way.

These missteps are clumsy, and also imply Facebook’s inability to draw value from their data set because they have resorted to rather blunt, ambiguous signals. This lack of subtle analysis pervades Facebook’s advertising. For example, it’s fairly clear from our co-incidence in discussion and photographs that I’m in a relationship — and who with. However, I chose not to tell Facebook this explicitly, so I almost always have adverts for dating sites on my profile. From this example and the many others like it, I can only infer that the vast amounts of money Facebook is surely pouring into research isn’t paying dividends yet.

What’s the conclusion? If you’re an investor, I’d be very wary of investing. If you’re a user, prepare to put up with ever more invasive and creepy re-purposing of the information you put into Facebook.

For further reading, I’d particularly recommend The Facebook Fallacy from Technology Review and After Facebook Fails from Doc Searls, which influenced this piece. As I finished writing this, I noticed Rian van der Merwe has collected a great set of further reading too.

.:.