Prismatic Pivots To B2B, Packaging Its Content Interest Graph Into APIs For Different Verticals

Next Story

Microsoft Apparently Wants Cortana To Be Fluent In British Humor

The market for news reading apps got a little more crowded earlier this year when Apple dropped into the scene with Apple News. But as of today, it’s thinning out a little bit once more.

Prismatic — a company that developed a way to suggest things to you to read based on what your friends and connections were liking and reading, using some clever algorithms to surface unexpected but relevant content — is pivoting into B2B services. Specifically, it is repositioning its technology around a set of APIs that can be used to track and surface content relevant to the needs of specific verticals.

It is starting by targeting three areas — publishers, app developers and hedge funds — that are meant not just to build up customers in these specific verticals but also showcase the range of ways that Prismatic can apply the tech it has built.

At the same time, Prismatic’s consumer-facing apps and site will stay alive for now, but they will no longer be updated, Prismatic’s CEO and co-founder Bradford Cross says. Instead, the idea is to develop more targeted products, such as a “Prismatic for finance” that would follow on from the APIs for hedge funds and compete against the likes of Bloomberg or other news aggregations services.

The news comes at a crunch time both for Prismatic and the wider area of news apps. If you recall, we reported rumors in March that tech companies were sniffing around to buy Prismatic, which had been going for a few years but failed to pick up much traction with consumers or even chatty power users.

Cross now describes that M&A activity as long-standing, without addressing why the company never chose to take the acquisition route.

“Companies have been trying to buy us since 2012 for lots of reasons,” he says. “What we have built is strategic for a lot of them, and it’s very common when you are doing a funding round or pivot that people will hear of that and come to you with deals.”

He says Prismatic is pivoting, but not raising money. (To date the company has picked up a very modest $15 million, with investors including Accel Partners, Jim Breyer, Yuri Milner and Javelin’s Alex Gurevich.)

Yet with the pivot more or less complete, those rumors haven’t really gone away. Even as I was working on this story, we were getting anonymous tips that the site was inching to closure and moving to an acquihire situation with a very large startup.

Cross says those most recent tips are “hogwash,” although he did say there have been some layoffs in the lead-up to the pivot, with the company whittling down to 10 employees.

From Recommendation to Refocus

The wider market for news apps, of course, is seeing some considerable pressure but also a bit of consolidation.

Aside from Apple launching its own news app, other large platforms have also been making their news aggregation plays. Facebook — itself a go-to place for people to discover online content — has doubled down with its own Instant Articles. LinkedIn, meanwhile, has been trying to make Pulse ever more relevant to its audience and a way to pick up more users. Google is taking its own route with Newsstand.

And independents like SmartNews and Flipboard are still trying to get more scale to compete against the lot of them. Flipboard recently announced 70 million monthly active users. SmartNews has been raising money and attracting attention but as of March had only 5 million MAUs.

But as we pointed out when we wrote about Prismatic in March, news aggregation is a tough business.

“It’s not clear there’s enough of a news junkie audience to sustain [those] companies by consumer Internet standards,” a source told us. “It’s also not clear that even a beautiful, relevant product can tear people away from social networks as the primary distribution channel for that kind of content. People turn to Twitter/Facebook for reasons stronger and stickier than news.”

By refocusing on B2B, Prismatic is hoping to step away from that race and at the same time provide tools to smaller companies and those who are not pouring millions of dollars of investment into data science, to help put them more on par with whose who do.

“Companies like Google, Facebook, Amazon, Netflix, and LinkedIn have been driving revenue and engagement with machine learning for years, from recommendations and personalization, to ads and e-commerce,” Cross says.

Publishers will be able to match what kind of content is tracking with which users in order to sell ads more accurately against that data and provide better ad pricing optimization. Developers can use Prismatic’s “interest graph” which it originally developed for its consumer product to surface relevant content or other information to keep people engaged in their apps.

And hedge funds will be able to use the APIs to surface breaking news across the web relevant to a topic or business that they are following, not unlike the kind of technology that, say, Dataminr provides around Twitter but covering the wider web.

Whether Prismatic will find more success with the new direction — or whether one of those acquisition rumors finally proves to be true — remains to be seen. There are already some companies that have signed on to use the tools who are named as reference customers, including Purch (publishers of Tom’s Hardware and AnandTech) and SimpleRelevance, so that is a start.

More generally you can see some of the logic behind the decision not just for Prismatic the business but also in what it can bring to the table for potential customers. Right now, regardless of what industry you are in, the name of the game is technology, and how you can use it to supercharge the work you produce. And what that often boils down to data and how you can harness it better. It’s among companies like Dataminir and Scaled Inference in trying to create ways of helping businesses do that even if they are not Google or Palantir.

“We’re levelling the playing field by enabling publishers [and developers] with the same technology, without the need to build it in house. We see the bigger opportunity building machine learning models for common problems that developers can just use as off-the-shelf solutions. For example, use an API to classify web pages by topics, rather than using tools to build your own machine learning models.”

Prismatic gets a percentage of the revenue increase that it generates for publishers, while developers pay a flat monthly fee based on usage, Cross says. It’s still working out pricing for hedge fund API usage. (As a point of comparison, the revenue model for the old Prismatic was the same as the current model for publishers. “We just meant for it to be more like Adsense/Adwords where we ran all this in our own consumer products too,” Cross notes.)