One of Facebook's (NASDAQ:FB) goals during the next 10 years is to understand the world using artificial intelligence. The company relies on understanding what people are talking about on its social network in order to target its advertisements to the users most likely to click on them.
Facebook recently moved one step closer to understanding the world by acquiring Wit.ai, a natural-language platform used by developers to add voice controls to apps and devices. Facebook can add Wit's platform to its mobile-development toolkit, Parse, to attract more developers to its platform. It can also use it on its own flagship product to understand its users better. But perhaps most importantly, Wit will give Facebook another way to gather data about its users.
How you talk to your thermostat
The connected home is upon us. We have smart thermostats, smart refrigerators, smart light bulbs, and smart door locks. And when we have to communicate with these smart devices, we use our smartphones.
Wit's technology allows app makers to use natural voice commands -- "turn up the heat" -- to control their smart devices. This places Facebook's newest technology in a very important layer of each new smart device where it can gather gobs of data on its users.
Likewise, even if a regular smartphone app maker wants to use Wit's platform, Facebook will be able to learn how users interact with the app and their smartphones, which can provide a wealth of data it can feed back into its advertising and app-creating business.
That's why Facebook's first course of action when it bought Wit was to make it completely free for everyone. It wants that data.
Figuring out what people are talking about and looking for
Facebook could do so much more with Wit than the original team ever could. Using Wit's technology, Facebook could process the language of the trillion-plus Facebook updates posted by its users. That would help it understand its users significantly better.
Additionally, Facebook could apply Wit's technology to Graph Search. Facebook's recent update to Graph Search left a lot to be desired, but it did add the ability to search through users' old posts. But Facebook doesn't seem to understand what people are looking for when they search for "karaoke bar" or "good place to eat sushi."
Wit's natural-language-processing technology should help Facebook process its old posts better, so it could provide users better context when it finally understands that users searching for a "fun bar" want nearby recommendations based on their friends' activities -- not a bunch of pictures of drunk people.
All of those videos have data in them, too
Another use for Facebook could be to process the audio of the billion-plus videos Facebook users watch every day. Video is especially valuable because it requires active participation from viewers, which means Facebook knows what you saw and heard. With a text post, it's easy for users to glide over a few posts without actually reading them.
Again, Wit's technology could help Facebook gather more data from those videos to find out what its users are talking about in them, and what the audience is actually listening to.
None of those will provide an instant boost to ad revenue, though
While the amount of data Facebook can acquire by using Wit in different ways is fairly substantial, it will still be a long process to gather and process that data for its advertising business. It's not like everyone in the world has a connected home, and a trillion posts is a lot of data to go through -- even for a huge computer network.
In the near term, Facebook's most likely to make money off of Wit by attracting more developers to its mobile-development platform. Facebook makes money on the backend by offering mobile developers the ability to monetize their apps either through in-app payments or through advertisements.
In the trailing 12 months, Facebook's payments business generated $958 million. Growth in the segment has slowed into the low double digits to high single digits, however, and payments revenue from desktop games declined for the first time in the third quarter last year. Attracting new mobile developers is key to spurring the segment past $1 billion in annual revenue.
The bigger opportunity may be in advertisements. Facebook opened its ad network last year, but it's unclear if the additional inventory will have an impact on its average ad price. Considering Facebook was able to increase average ad prices substantially during the last year, including a 274% price per ad increase in the third quarter, I'm betting the trend continues, which will only attract more developers to Facebook's platform.