FT’s understatement on Newssift.com
by Pete Bell
Pete Bell
Yesterday, the FT gave a statement to paidContent UK about shutting down Newssift.com:
“It’s the nature of digital media start-ups that not all work out, but we still firmly believe in the need to experiment and innovate. Newssift was a good idea, but the timing was unfortunate in that its launch coincided with the advertising downturn.”
So how good was the idea, and how unfortunate was the timing? For fans of classical rhetoric, that FT statement is an understatement.
Newssift.com was the product of FT Search, a spinoff from the Financial Times. It was the brainchild of Robin Johnson, a senior veteran of both the media and search worlds (and a great guy to boot), who took it from idea to launch as its CEO. As Robin put it, the financial world was very well served with quantitative research tools, but not by qualitative ones. He wanted to correct that.
Robin’s vision was to create a vertical news search service that would use text mining tools to help analysts make sense of thousands of news feeds, as the stories broke. Endeca provided the search platform, making possible the faceted search, as well as business intelligence visualizations on top of measurements like sentiment. The platform was fed by several text mining technologies, including Lexalytics, as well as NStein, which has since been acquired. At its launch, the site received tremendous attention, and then continued to generate enthusiastic buzz.
At our Discover user conference, Robin sat on a panel about innovation – here’s a video we showed to introduce him that encapsulates the Newssift vision:
Along the way, Robin and the team shared some great metrics:
The audience loved the product:
“I’ve been doing product development most of my career, and we’ve always slapped a five-box scorer on new products. With the people who are our target market, 85% of them pick the top box, saying they’d like to use it again and they would recommend it to a friend.”
And it was a great audience. For the FT, CPMs were a stunning 10x the industry average:
“Who reads the FT? All the central bankers, the CEOs of the largest companies in the world, the people who travel overseas in business read this paper. It’s a remarkably special audience. It’s somebody with an average household income of $500,000 dollars a year spending half an hour with your product. That’s fantastic.”
Dream product, dream audience, dream team. That gives you sense for just how big a challenge the online media industry is facing with ad-supported models.
When the dust has settled, Newssift.com will be remembered by the business world as a failure of the traditional media business in the face of a bubble made of stainless steel appliances & granite countertops. But it will be remembered by the search community as the best public showcase of text mining to date, and as a vision for vertical search for news that will surely return in some new form.
on March 17, 2010 at 10:59 am
Permalink
As I wrote a year ago commenting the launch of newssift, if to build a site like Newssift you need 25 people and 4 different technologies, the market will continue to perceive text mining as an expensive, uneconomical, academic effort.
http://techcrunch.com/2009/03/18/the-financial-times-launches-its-own-business-news-search-engine-newssift/
Think about building the site in 4 months, with 4 people and 1 core technology better performing than FT at a fraction (really) of the cost. The site would have become profitable in 3 months and the world could have learned a new way of accessing relevant news.
on March 17, 2010 at 9:43 pm
Permalink
Luca, I agree that new semantic technologies are making it less expensive to run sites like Newssift, and that the benefit of those cheaper technologies will accrue to media businesses and beyond. Even still, the very best semantic technologies still have hard limits, which means that for an audience of the highest-end investors, human editorial judgment must complement the semantic technologies (a theme of this blog is the HCIR precept to let people do what they do best and let computers do what they do best.) And beyond that, those thousands of content feeds cost money.
But in the case of Newssift, the cost of operations wasn’t what got them. It was the failure of the traditional ad-based media model. As a venture-backed company and FT spin-off, they couldn’t just be profitable — they had to be profitable at scale. And that takes international ad sales and business development, which is expensive.
I could certainly imagine another vertical news portal running on cheaper semantic technologies, cheaper news feeds, less human editorial judgment, and automated ad sales like Google Adsense. It’s just a different kind of site. And hopefully we’ll see one soon…
on March 26, 2010 at 6:24 am
Permalink
> and a great guy to boot
Indeed.
on May 6, 2010 at 9:10 am
Permalink
> and a great guy to boot
I 3rd that sentiment