Sunday, April 07, 2013

Visual Revenue Lives Up to its Name

I see from the WSJ that Visual Revenue has been acquired. Good for them. Met CEO, Dennis Mortensen, at a conference in Paris last year.
CEO Visual Revenue Dennis Mortensen
Outbrain, a content-recommendation startup based in New York, is acquiring Visual Revenue, another startup that we have mentioned on this blog before. VR gives editors of news sites tools and data to help with decision-making when setting up page layouts online.
Here’s how Visual Revenue works: When an editor is deciding on the layout for a page, Visual Revenue uses data to predict how well a story will do given its placement, includes a number of automated tools like switching between headlines until it finds the most effective one, and tracks performance.
Outbrain powers many of the recommendation engines for publishers on the Web, which usually show up as a “see also” array of other pieces of content on pages.  Outbrain reaches about 80% of the online population the U.S., Chief Executive Yaron Galai said.
But as powerful as those algorithms are, there isn’t a replacement for human editors who will recognize the significance of a story without needing the data that algorithms need to operate effectively, he said.
“I don’t think it’s a matter of the technology not being enough, it’s that algorithms are just as good as the data that’s fed into them,” he said.
That’s why the company is making its third and largest acquisition to date with Visual Revenue closing the loop between human editors and the algorithms that help build the most personal page layouts possible.
We caught up with Mr. Galai to find out more about the acquisition. Here’s a recap of the conversation:
WSJ: What does Outbrain do?
Yaron Galai: What we’ve historically done is power the content recommendation of a publisher’s site. What Visual Revenue does is help the edit side of the house better program the areas that it controls using signs and data.
WSJ: Why do editors need tools like this?
YG: Historically, the home page has been programmed by the editors deciding the most important stories. We think that’s a great way to do it, but there’s so much data and science that can help editors make better choices as to what they show and add some personalization. That’s the piece Visual Revenue brings to the table.
The way I think about a publisher is like a plane. Outbrain developed the best autopilot system, and Visual Revenue has developed the best GPS or navigation system. By putting the best of the two together, that makes for the best plane possible.
All the recommendation engines seem deceivingly similar in idea, I think they’re very different in approach based on the domain. Our engine would do a terrible job recommending DVDs and Netflix’s amazing recommendation engine would do a terrible job recommending news articles. We try to focus on one area, just publisher content, and do that better than anyone else.
WSJ: Why not just use algorithms?
YG: I strongly believe there is no replacement for editors. I can’t imagine how that might ever happen. Once stories break and once people start engaging with them and you start accumulating enough data, then algorithms can come in and say we can optimize. But they are very poor at predicting what’s going to be important. I don’t think it’s a matter of the technology not being enough, it’s that algorithms are just as good as the data that’s fed into them.
WSJ: What’s happening to the team?
YG: We’re taking the whole team of about 25 with us. We’re keeping the product, and over time we’ll start integrating the solutions for publishers. For now, publishers can use Visual Revenue by Outbrain and use the Outbrain stuff for the automated side.

No comments: