This is the story of online advertising:

At first, Mark Zuckerberg and his staff didn't know what their business model would be, but they had a strong suspicion that gathering hundreds of millions of eyeballs on a single site might be worth something someday.

Eventually and inevitably, Facebook introduced advertising to its site. The strategy was clear: Get everybody to gather at a single location (Facebook.com), harvest social signals at that location, and sell ads that would be displayed at that location. Everyone and everything in one place.

Facebook made a lot of money, went public, then made a lot more money. (The company this week reported Q1 revenue of $2.5 billion.)

Everything was fine in Facebookland, except for one small problem: Most users were clearly migrating from desktop to mobile, and nobody was making significant ad revenue in mobile.

Except Google.

How did Google make money in mobile? Instead of harvesting personal data in one place and displaying personalized advertising in that same place, it did both in multiple places.

Ever since co-founder Larry Page took over the CEO spot from Eric Schmidt in January 2011, the secret sauce for Google as a business was the controversial unification of the Google privacy policy in January 2012.

That simple policy enabled Google to harvest user signals and personal data from one service -- say, search -- and use that information to personalize advertising on another service -- say, YouTube.

That change ushered in the future of online advertising, as both Twitter and Facebook would later discover.

Facebook's new game: Harvest personal data from multiple apps.

For Facebook, that discovery started with an unwelcome shift in user behavior. It turns out that many Facebook users, especially younger users, got tired of interacting on a single site with their friends, extended families and co-workers -- and everyone else. So in combination with the flight from desktop to mobile, they increasingly embraced little apps for social interaction: Snapchat, Instagram, WhatsApp -- you name it. On these little apps, people could have unfettered interaction with small tribes of friends without the public-pronouncement feeling of Facebook.

So Facebook decided to start buying the apps and services that young users were fleeing to. First, it tried and failed to buy Snapchat. After that move proved unsuccessful, Zuckerberg built his own service, called Poke, but it didn't work out so well.

Then, about two years ago, Facebook bought Instagram. And a couple of months ago the company announced its intention to buy WhatsApp for about $19 billion. This week, Facebook announced the acquisition of the fitness app Moves. (It's marketed as a fitness app, but in fact its purpose is to track exactly where you are at all times and give you incentives to label those places to provide better location data.)

Whenever Facebook acquires such companies, it always says that it's not going to change anything, and that it's not going to fold them into Facebook proper. People tend to roll their eyes at these pronouncements, but I believe Facebook and I'll tell you why in a minute.

These acquisitions were part of a larger "multiple app" strategy by Facebook that has been complemented by homegrown apps like Messenger, Facebook Camera and Paper.

From a data-harvesting perspective, these various apps fall into the categories of "who you know," "where you go" and "what you like." These are the same categories of personal data that Facebook gathers to provide increasingly relevant and customized advertising at Facebook.com.

This week, we learned that Facebook plans to unveil a new mobile ad network at its F8 conference next week. The purpose of the network is for Facebook to sell ads outside of Facebook.com and outside the Facebook mobile app.

If Facebook's direction or strategy isn't clear, let me spell it out: Harvest personal data from multiple apps, then sell personalized advertising in multiple locations.

It's no longer about keeping users in one place:

Here's an oversimplified example: An ad for a Starbucks promotion presented to you in a mobile game (sold through Facebook's upcoming ad network) might be based on knowledge that you spend a ton of time at Starbucks -- information harvested from the Moves app.

As you can see, there's no Facebook -- no social network -- involved in this series of events. But Facebook gets paid anyway.

Twitter is embracing a "mini" version of this approach.

Twitter outgrows Twitter

Twitter acquired Vine in October 2012. As Facebook did with its social acquisitions, Twitter kept Vine separate and didn't brand it as a Twitter product.

And last year Twitter acquired a mobile ad network of its own called MoPub.

Then Twitter this month acquired a company called Gnip, which specializes in the harvesting of social signals for advertising and marketing.

The control and ownership of Vine, MoPub and Gnip demonstrate a shift in thinking. Twitter is no longer a social microblogging service, but instead a personalized advertising company that harvests user signals from wherever and then displays personalized ads wherever.

Which brings us full circle back to Google.

Google outgrows Google+

Google this week announced the departure of the guy who, in Larry Page's words, "built Google+ from nothing." But Vic Gundotra's departure appears to be part of a larger de-emphasis of Google+ as a social network that unifies everything.

Wherever the users go, there you are:

The chatter in Silicon Valley is that Google will keep Google+ going, but invest more heavily in Google+ as a platform, rather than as an all-purpose destination social network.

In other words, with Facebook outgrowing Facebook and Twitter outgrowing Twitter, there's no need for Google to drive so many forced integrations with other Google properties.

Google has realized the same thing Facebook and Twitter have realized: There's no need for unity. Ubiquity and diversity -- what Google has succeeded with all along -- is more powerful.

The new model is to harvest social signals from wherever and sell personalized ads wherever.

Of course, none of the social networks are going anywhere. They're still important both to their companies and to their users, and they still play an essential role in user identity and data harvesting and, for Twitter and Facebook at least, as places to display advertising.

What's important for these companies from a future-facing business perspective is to have multiple mobile apps that harvest multiple dimensions of personal data that can be applied to highly customized and personalized mobile advertising at multiple locations.

The social networks are falling apart -- they're breaking up into multiple sites and apps that do in a scattered way what used to happen centrally.

If you're in the personalized advertising business, why restrict yourself to a single social network? Users don't.

More on social networking:

• Facebook to test new mobile payment service
• Facebook embedded posts good for business
• Facebook rethinks its 'hackathons' with an eye toward mobile
• Facebook gets its mobile game on
• The social network wars are over. The winner: email!

what can be more valuable than advertising? is advertising a terminal business model?

Apple, Alibaba, and Amazon have demonstrated that directly selling goods to consumers can be more lucrative in the long run.