As the dust settles around the Facebook IPO, the important question becomes not what went wrong, but rather, how will Facebook create value to justify its lofty valuation?
Internet platform businesses can scale very rapidly because adding customers is virtually costless. While there are several possible business models for Facebook to create value, each with its potential risk and reward, I propose that Facebook create the world’s first information market. This market-based model would pay Facebook users in exchange for permission to provide selective information to businesses looking to reach specific types of people.
Facebook can leverage its huge base of “sticky” users and monetize its unique and unprecedented access to the largest base of global data on human behavior. Instead of trying to coax its users into clicking on ads or to try gimmicks that risk their ire, Facebook can create an efficient market in which its users, businesses, and ultimately shareholders benefit. Facebook is in a unique position to capitalize on this disruptive idea.
[#contributor: /contributors/5932413644db296121d69eea]||||||
Facebook has little brand loyalty or aspirational value, unlike Gucci or Apple, but leaving Facebook involves large switching costs. Even if a much better social network platform emerges, it will be costly for people to shift their entire social network seamlessly onto the new platform, which means they're not going anywhere else for the foreseeable future. This stickiness provides Facebook with options for creating relationships with its users and monetizing them. As its user base grows, Facebook’s data becomes its primary asset.
Facebook should pay its users to share data with businesses.Large troves of data are valuable because they make it easy to build predictive models which in turn translate into money. If a model can predict the stock market’s direction better than a coin flip, it can probably make money because its recommendations are actionable. Businesses such as Google and Amazon exploit their data advantages very effectively by using their data to estimate the probability of people's actions when they're presented with certain inputs. The companies act on the basis of such predictions and generate huge profits.
Data mining is not new. Banks and merchants have been doing it for more than two decades because they have access to your transactions that tell them what you buy or have explored. Google knows your intentions through your queries. But Facebook has something even better – your network, your feelings, your opinions, your life history, your videos, and much more. Estimates are that Facebook logs several hundred terabytes daily, corresponding to billions of actions such as “like,” “share” and “recommend.” This data is potentially orders of magnitude more valuable than transactions. What could Facebook learn from this data? Facebook can identify behaviors of groups of people that could be of interest to businesses. Behaviors can be likes, shares, aspirations, intentions, happiness, and so on. Groups could be based on age, gender, region, profession, hobby, passion, status, and so on. No business prior to Facebook has had such a rich laboratory of human behavior.
Unlike transactional data where standard data mining methods can reveal simple patterns such as “people tend to buy beer and diapers together,” Facebook faces a huge challenge in interpreting its morass of unstructured data that includes text, images and videos. This area is still a research frontier, but not a theoretically insurmountable hurdle, and one that Facebook needs to address in order to build powerful predictive models.
But users are on Facebook to socialize, not to buy. Whatever Facebook does with their data should not violate their users’ implicit intent, which is to share data with others, not Facebook. Intent is the cornerstone of responsible data use by business. The more incongruent the use relative to the customer’s implicit intent, the higher the risk for the business. The good news for Facebook is that its users provide data voluntarily and willingly -- to share it. Use of data provided voluntarily is less risky than when the data are required for service.
Broadly, Facebook has two options in how it monetizes its predictive patterns of behavior. It could monetize the user side of its network or the business side that wants to reach its users. Targeting advertising is one obvious route, but there are risks involved because people are not on Facebook to buy things. This route also risks being perceived as a “bait and switch” business if people feel that Facebook’s priorities have shifted from a networking platform to an advertising platform. Facebook could also start charging people for certain services, such as alerts of their choosing, in accordance with the “Freemium” model of LinkedIn and others. But again, the potential here is unclear and the strategy is not without reputational risk.
The most promising possibility for Facebook is to monetize the business side by paying its users to share data with businesses who would pay for the service. This would create information markets where people are compensated at market rates for revealing selective information about themselves. When information markets were first proposed, their authors envisioned a bank being the custodian and providing clearing mechanisms. But the Facebook platform is much better suited for this function by virtue of the massive sticky user base and their implicit trust, manifested by the type and scale of information they reveal. This is great news for Facebook users. Imagine Nike wanting to survey runners over 40 in South Africa, a traditionally costly endeavor that could be completed in hours by Facebook users who get paid for providing their information! Or users whose “Interested in” field includes mortgage refinancing or a brand of clothing, who are then exposed to products they have a high chance of buying, while getting compensated for it.
Traditionally, advertisers know too little about their targets which leads to too much spam and turns off people even to offers they might otherwise accept. An information market would help solve this problem and let users share monetarily in the benefits created through a more efficient matching process. Many years ago, I asked the head of the largest credit card business in the US whether they would be willing to compensate their customers in exchange for the business generated through targeting. Without hesitation, the executive said yes, but the systems for tracking and accounting were not mature enough at the time. Now the time is ripe, and Facebook could create such an information market where all parties are fairly compensated.
Image: joshrussell/Flickr
