Pictures and Posts: The Future of Online Advertising is You

User generated and content based. Those are the buzz words that describe the future of online advertising.

User generated is easy enough to  understand: it refers to the photos, videos, likes, friends and other inclinations we all signal in our desire to communicate meaning.

And somewhere, someone is digitzing and correlating that data with other information in order to generate a 'map' like the one here so that the interrelationship between these various disparate interests can be converted into a commercially viable tactic to get you to buy something else.

Content-based is a somewhat more amorphous concept. It infers that content is not really involved but is applied in order to generate an outcome of value to an entity with an interest in it.

If this all sounds a trifle indirect, well, welcome to the world we've created through our embrace of media fragmentation. We now deal in brand affinities as much as loyalties. It's an odd bargain: tech has shattered our once stable binary input-output world so all of the associated software, hardware, media, advertising, entertainment, commercial, merchandising stakeholders are now asking us to help them make sense of it via our texts, tweets, posts, blogs, likes, friendings, snapchats, instagrams and youtubes. They  ostensibly ask our permission but what they really want is an explanation. If it really has value, though, they will ultimately need to share the wealth such meaning generates. JL

Derrick Harris reports in GigaOm:

Artificial intelligence techniques such as natural-language processing and computer vision will someday revolutionize our world. First, they’ll probably help retailers sell us more stuff.
Advances in artificial intelligence, particularly around computer vision and natural-language processing, promise some truly life-changing capabilities. But before they change medicine and make driverless cars the norm, they’ll probably find their way into our lives via targeted online ads. On a web dominated by social media and littered with the corpses of unread banner ads, it’s just too natural a fit.
Now, I am no expert in strategies behind targeted marketing or the technologies that run ad platforms, but I’m sure of this: Knowing a consumer’s intent is the key to getting more clicks. Search proved to be a highly lucrative business for Google in part because it was such a great medium for advertisers. Consumers tell search engines exactly what things they’re looking for, and companies selling those things can step into the paid slots like knights in shining armor.
Across the web, though, current targeting techniques don’t know a whole lot about what website visitors want. Cookie-based methods have little intelligence into whether users were ever interested (commercially, at least) in a site they visited, or whether they were and already made a purchase. Countless data scientist hours have gone into trying to infer intent based on behavior data, but it’s not clear the results have been worth all the effort.

The Ditto Labs streaming demo in action.

It seems startups and investors are catching on, and are betting they can learn more about what consumers really want by looking at the content they generate. On Wednesday, a logo-recognition startup called Ditto Labs announced it has raised a $2.2 million funding round (check out its streaming firehose of Twitter, Instagram and Tumblr images here), and a social media–focused NLP company called Attensity announced it has raised a $90 million growth round of equity financing.
Companies are also trying to glean data from customers’ reading habits. A few weeks ago, a startup called Snap Skout launched a product that serves up local deals related to the content on the websites people are reading. We’ve covered a company called Cortica that helps users serve ads based on the images contained on a web page.
Over the past year, there has been a spate of acquisitions in the computer vision, NLP and general AI space, including by companies such as Yahoo and Pinterest. Facebook and Google both have deep learning teams working on building systems that will master object and facial recognition, as well as the understanding of language.
This is not to mention the general-purpose companies, such as AlchemyAPI, Expect Labs or IBM (with Watson) trying to democratize AI via API calls. I’m sure there are dozens of other companies working on similar things to all of the companies mentioned (please link to them in the comments if you’re so inclined).

The approaches are different, but the underlying rationales are the same. When someone posts a message online, they’re probably saying something about a topic or product they care about. It might include their feelings, it might include language indicating they’re in the market to buy something or getting ready to go somewhere. When someone posts a picture on Instagram, Pinterest or anywhere, even without accompanying text, it can contain deep information about what they’re doing, who they’re with, what kind of clothes they wear and what things interest them.
Even if this information doesn’t trigger an automatic ad, it can feed broader efforts around trend detection and consumer segmenting. Companies such as Facebook and Pinterest, which already have demographic data on users, can now know a lot more that will be a lot more valuable to advertisers. So REI can target users who were recently talking about going hiking. Jack Daniels can learn that people who drink its whiskey also like Metallica. Macy’s can see, sooner, which fashions are all the rage among women aged 18 to 24 in the various cities it serves.

Snap Skout in action.

At the very least, when someone is reading content on a website, knowing what’s really on the page is probably better than knowing where someone has been online or being able to identify some keywords. Display ads don’t work, and we’ve all seen those keyword-triggered ads that result in horribly inappropriate juxtapositions. On the other hand, if someone is reading a story about a new restaurant opening, hit them up with deals for new restaurants. If someone is reading about the hard life of a capybara, maybe an ad for an Amazon adventure or the San Diego Zoo will trigger a click.
This type of user-generated and content-based advertising isn’t going to save the starving children. It certainly won’t quell concerns over online privacy. But it is coming and, who knows, it might even get you to click.