The Facebook Segmentation Problem
Facebook is about to go public, so obviously the company is getting a lot of attention and focus on its business and long-term prospects. I am not a frequent Facebook user, but I have enormous respect for fellow entrepreneurs and Mark Zuckerberg has done an outstanding job of building a product that people love to use and a very successful company. So congrats on the upcoming IPO.
The big question, of course, is what is Facebook worth? It is an interesting discussion not only because so many people use Facebook, but because it is going public at an extraordinary valuation. Since Facebook provides a service that is free to users, it has to make money from its customers, advertisers.
This is the interesting long-term question: can Facebook create more value for its customers (advertisers) than competitive solutions. If it can, it will continue to thrive, if not, it is a serious risk for any investor. Its users might be happy, but its customers won't be.
Today, there is a story in the NY Times that points out this potential problem:
On Tuesday, General Motors, the third-largest advertiser in the country, shut down its Facebook budget, about $10 million, saying that those ads were simply not doing enough to sell automobiles.
In our language, the job that Facebook is helping its customers execute is to acquire customers. All advertising is about customer acquisition. You might argue that advertising also provides "branding", but this is just a step in the customer acquisition process. No company needs branding without customers, but all companies need to acquire customers, whether or not they decide to build a brand.
Of course, Google has been extremely successful because they got the customer acquisition job done much, much better than alternatives, and they made it measurable and quantifiable. Google satisfied the unmet customer acquisition needs of all companies, big and small, and they have been rewarded with a very defensible model that is exceptionally profitable.
So will this be the case for Facebook? Perhaps, but another part of the same NY Times story is very revealing:
Facebook currently allows marketers to buy audiences, say women between the ages of 18 and 35 who live in a specific neighborhood. But Facebook is essentially a walled garden, so what users may do on the Web beyond Facebook — where they shop, what they read — isn’t available for advertisers who want to home [sic] in on people’s behavior. [Emphasis added]
Facebook's product - how it generates revenue - is very traditional. It allows advertisers to target users based on demographic segmentation, for example, women between the ages of 18 to 35 who live in San Francisco.
So why is this a problem? Because demographic segmentation has failed companies for decades. We have seen this in every type of market from consumer markets, to business markets, to medical devices, to software, and services. Over the long-term demographic segmentation will be viewed as an ancient relic for customer acquisition because demographics create phantom segments. They aren't real.
This makes perfect sense if you define markets based on the customer's job-to-be-done and not a product. Let's look at an example.
Suppose you wanted to acquire customers for a new music service. The underlying job your service satisfies is to listen to music. The job is stable and will never change, but the products and services will continue to evolve as they have in the past from records, to CDs, to iPod, to Pandora.
So how should you target customers to acquire? The obvious answer is to target customers who are unsatisfied with their ability to listen to music successfully. And let's say the part of listening to music they are unsatisfied with relates to discovering new music. Your potential customers are tired of their music collection and they really want to discover new music better.
So you launch a new ad campaign on Facebook. And you select 18 to 35 year-old women in San Francisco. Will this get you the new customers you want? Probably not. So why is this?
Because the people who are unsatisfied with their ability to execute the job of discovering music don't segment themselves by demographics (e.g. age, gender, income, zip code). They segment themselves by how well they can execute the job.
Could an 18 year-old woman in San Francisco, a 60 year-old man in Detroit, a low-income mom in New York, and a wealthy man in Seattle all be similarly dissatisfied with their ability to execute the same job, e.g. discover new music?
They answer, of course, is yes. But these people would absolutely never fall into the same traditional segmentation. And this is the risk for Facebook - its segmentation is inefficient. Chris Dixon points out that Google's revenue per pageview is estimated to be 100x to 200x Facebook's.
This is because a company can acquire customers on Google based on the job the customer is trying to execute, not based on phantom demographic targeting. In other words, it doesn't matter if I am male, female, high income, low income, or anything else when I type a search term into Google. All that matters is that I search for "music discovery" and when I find your new service, I will be a satisfied customer, regardless of my demographics.
Finally, I found it amusing that the NY Times article points to a possible improvement for advertisers on Facebook: that they could in the future use Facebook's user behavior data to segment. This type of behavoir segmentation or "psychographics" is just as flawed as demographic segmentation. It just came into vogue as computers got more advanced and companies could access and analyze consumer credit card transactions.
But what I purchased in the past tells you nothing about how satisfied I am with my ability to execute an important job. So the only accurate segmentation is based on a customer's job-to-be-done (and specifically on the outcomes in the job). This is how customers segment themselves.
If Facebook only offers companies traditional segmentation, their advertising model will be at serious long-term risk. If it can figure out how to get the customer acquisition job done better than Google and others, it could continue to thrive.