How one West Coast tech company is turning traditional broadcast publishers into hyper-personalised publishers
Boomtrain is possibly one of the craziest named companies in tech. It’s also a company with a super serious mission: to personalise the publisher / reader experience and to turn the interaction between publisher and reader into a 'conversation', rather than a ‘megaphone-style’ one-size fits all experience of the past.
Powered by AI (artificial intelligence), Boomtrain’s technology enables publishers to create customised content experiences across all devices (mobile, desktop and email) off the back of each interaction a reader has with a publisher’s website. The ultimate goal is to create a level of content personalisation for publishers akin to Facebook’s feed-personalisation.
Whilst over in San Francisco earlier in the summer, Hackers.Media sat down with founder Nick Edwards to discuss Boomtrain and publisher personalisation.
Q: Vision. Where did your vision for Boomtrain come from?
A: We started Boomtrain as computer science students intrigued by machine learning and how things like AWS (Amazon Web Services) give you access to masses of data cheaply. AI has been talked about for a while, but historically most AI has been based in laboratories with little in practice. Boomtrain is very much about applied AI, providing technologies to marketers to make the most out of AI.
The publishing world is going through massive transition. A teenager in their bedroom has the same potential reach as a billion dollar company like Time Inc. The question is how can publishers build long term value in their business? And we’d say that the answer isn’t so much about reach any more, but more about relevancy - relevancy is the thing that will gives publishers a competitive advantage. In a world where people are used to a hyper-personalised content experience via their Facebook feed, consumers expect a personalised content experience wherever they consume content. And this is the challenge that Boomtrain is focused on - providing a similar personalised content experience as Facebook but for editorial publishers. Our goal is to help publishers understand and speak to their audience at scale.
Q: Practical. How does Boomtrain work behind the scenes, what’s the end experience for the user?
A: All machine learning / AI is fuelled by data, and Boomtrain is no different. For our technology to deliver content personalisation, we need to have a machine-based understanding of the user, and the content they’re consuming: what someone’s reading, how they’re interacting with the content, what they’re liking, commenting on etc. With all these data points, Boomtrain’s technology then creates a customised content experience per publisher channel.
To address potential negative effects of content personalisation, like the phenomenon of ‘echo chambers’, Boomtrain’s algorithm includes the idea of ‘serendipity’, including a sense of randomness in our recommendations, to enable people to discover things that are brand new to them. It’s also important to remember that personalisation isn’t one-size-fits-all, each person’s content experience is different and will ultimately change over time as the machine learns an individual’s behaviour.
Q: Proof. How does a publisher know the AI is working?
A: Frequency is one of the key metrics our publisher partners use as a proxy for success. Are users coming back to the site more frequently, clicking more, reading more articles, spending more time on site. In essence successful use of Boomtrain’s technology should result in greater engagement with your quality content.
Core to Boomtrain’s IP is that our algorithm is continually learning and constantly optimising. There’s 100’s of different approaches for AI optimisation, different for Netflix, different for Facebook, Amazon etc. These big tech companies build AI engines for their own specific use. What Boomtrain has done is create an AI engine that customises itself for each specific use of each specific publisher.
Q: Publisher demand. How are publishers using your tech the best?
A: Publishers using Boomtrain for email / push notifications are some of those getting the most out of our technology, by delivering the right selection of articles, at the right time with the right subject line. A good example is GameSpot (from CBSi). GameSpot is one of the biggest gaming sites in the world, with a weekly digest newsletter going out to millions. The challenge GameSpot has is that gaming isn’t a generic subject matter but hugely diverse, as every gamer has their different passion point; Xbox gamers, PC gamers, Mine Craft gamers etc etc. What Boomtrain does is help Gamespot learn the passion points of each of their readers, and create a customised newsletter for each - at a huge scale, something that a human editor could never do.
Q: The future of personal publishing. What's next for publisher personalisation and Boomtrain?
A: In essence Boomtrain acts as a ‘mini-assistant editor’, helping editors do their job better by serving up the most appropriate content at the right time. There’s a ton of ways personalisation could move next, but a couple of ideas that we’re toying with are using AI to provide a better ‘content trend feedback loop’, understanding what content’s resonating the best with different clusters of users, and using AI to improve 'chatbot publishing’. With all these innovations it’s a ‘man-machine collaboration’, in that the innovation generally starts with an idea from the editor which we then apply AI to improve and scale. It’s all about the machine helping man; Boomtrain provide publishers with the machine, so editors can do what they’re best at - creating great content.