Were you unable to attend Transform 2022? Check out all of the summit sessions in our on-demand library now! Watch here.

Imagine if great content found you because it’s smart enough to know who really wants to see it.

Over the next ten years, content embedded with machine-learning technology will start to be smart enough to seek out its own audience. This will solve a growing frustration for consumers today.

About three years ago, several companies independently started to work on this idea of smart content that is distributed and consumed in personalized feeds, each with its own different content type — user-generated vs. premium — and format — stories, photos, or videos. As an example, Facebook made its posts smarter by targeting them to audiences with specific interests; the Newsfeed now promotes posts based on you, no longer serving as just a timeline built on when someone you follow posts an update. Twitter recently introduced an algorithmic timeline to prioritize tweets according to their importance to you, significantly moving away from a timeline feed based on chronological order.

shot of a laptop computer as the screen glows in the darkness

Machine learning will be used to place good content that you care about in front of you when you’re online — on whatever device or social network you happen to be using. This will profoundly change our experiences with videos and stories, ushering in a new era of content discovery.

This new era is beginning already, driven by the type of learning technology that helps a Nest thermostat adjust to different family members or tells Uber where to send cars just before people need them.

Three eras — so far

If you think about it, there have been three eras of content discovery so far. The first era was programmed content by humans. Consumers would have to trust that a programmer on TV, an editor in print, or a DJ on radio would serve up good content built around a theme. So we’d turn to ESPN for sports, buy Fortune for business features, or tune into a jazz station and just hope something we’d like might be there. That era dominated media until the Internet showed up.

The second era, still going strong, is consumers searching for content using machine search engines. The 800-pound gorilla has been Google. Its whole business was and is based on helping us find what we’re looking for — but first we have to know what we’re looking for. Search-driven content is great when we know what we want, but doesn’t help us discover something we don’t know about.

We’re now deep into the third era: the social era of user-generated content. We discover content through friends or strangers we follow who post stuff they like. That sometimes leads us to premium content we want to see, and serves up everything from clickbait to videos of third-graders’ piano recitals.

But the fourth era is coming. In this next era — machine-learning, smart content — human-curated content itself will start to understand who we are and what we enjoy, and will seek us out. We won’t have to rely only on programmers and editors. We won’t have to know what we want so we can search for it. We won’t have to hope friends send us something good. Our online behavior will tell the technology what content to send our way and what to screen out. It’s a little like pairing a Netflix, Amazon, or Apple Beats recommendation engine with human-curated content.

Mode op-ed 3This technology will help more great content get created. Right now, content makers are struggling with major shifts in media — including the lack of a captive audience, the disaggregation of content from media properties, and recent moves in social media to serve content on their own platforms instead of linking out to the original host. Content driven by machine learning will help content creators leverage social networks to find the right audience, and then deliver the content to consumers anywhere — which will help them find engaged users and get paid for their work.

Even better, the technology will report back to creators about what content works with what audience and how to reach that same audience again with new content. This will be interesting to brands, which will be able to buy advertising up front based on a promised audience size and type — somewhat the way ads are bought on network TV, but with much finer and real-time targeting.

Machine learning is driving change

Machine-learning content — combined with human curation — is the new heart of platforms with smart social feeds. There are many companies now building competing solutions: Facebook with their new Newsfeed in which videos are now played instead of linked; Snapchat, which has added Discover for content from top media owners and city/event stories; and Mode.com, which allows professional creators to build stories and upload videos for personalized streaming in feeds. As a consumer of smart content on these platforms, every time you view or like or share content with machine learning, the technology will take note and begin to form an idea of who you are and what content you like. The more it learns about you, the better it will get at identifying you on any platform and sending tailored content your way.

Importantly, these algorithms can be applied to increase views of original content and also tell content creators, curators and brands what their audience actually wants. The content engines would work like a positive expanding spiral. Good content will find its audience and then report back so that more good content for that audience can be created.

Automated advertising will find the right people

All of the machine learning from all of the content propagated should add up to incredible overall insights about user interests. A technology engine will measure actual engagement and viral lift so it can start to predict what’s likely to generate 100,000 views and what has a chance to rocket to 1 million. The right technology could tell us, for instance, what most appeals to 33-year-old men with one child, a dog, and a Japanese car.

And of course, if the technology can tell us those things, then it will serve the needs of advertisers, which means advertisers will have more confidence that if they place native ads or sponsor videos and branded content, their message will actually get delivered to the audience they’re seeking. This is fueling a rapid growth of native and video advertising, a market predicted to grow from $5 billion in 2015 to over $20 billion in 2018, according to eMarketer.

So how will smart content work in the future? Today so much great content is created that simply never finds real users. In the future, each video or article will have its own mini learning engine that will continue to seek users who will love it and want to share it. So instead of driving users to a few large locations, smart content will come to us wherever we are — in apps, in our TVs, on our phones, on new devices, in our cars, at work, and at home.

The beauty of machine learning, as you might imagine, is that it will learn. It doesn’t have to be reprogrammed every time the media universe shifts. And it will no doubt shift many, many times in coming years. We are at the very start of this transformation, with the first parts of this vision starting to show up in platforms and apps now.

The future machine-learning era of human-curated content promises to cut through the next generation of media evolution and will make us all happier content consumers.


Samir Arora is CEO of Mode Media. Mode Media is a company in Silicon Valley that builds products that help people, create, discover and distribute native video and stories. A private technology media company, Mode says it has more than 140M monthly unique users and 45 Million monthly viewers on its new Mode.com video streaming platform, launched in 2015.

VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.