Open Text Buying Nstein, TechCrunch Misreports

Enterprise content management (ECM) vendor Open Text's planned acquisition of ECM + text-mining provider Nstein is an interesting bit of news, perhaps indicative of market challenges that may result from an over-narrow focus on a single, struggling market (publishing) and of the opportunities that text analytics -- software that automates tagging and classifying content, among other functions, to boost findability -- is perceived as presenting.

The story is significant in the text-analytics/publishing/content-management worlds. It gained the attention of TechCrunch, a much-visited site that "was founded on June 11, 2005, as a weblog dedicated to obsessively profiling and reviewing new Internet products and companies." In this instance, however, TechCrunch's "obsession" did not extend to thorough reporting, which is an interesting twist. TechCrunch got four points fair-to-middling wrong in one short, three-paragraph, four-sentence article, three of them in the 12-word headline alone.

I'll interweave my own analysis of the planned acquisition with a look at TechCrunch's reporting.Start with TechCrunch's headline:

Compare to the headline of the press release that ran yesterday on PR Newswire:

Note the difference in tense: "buys up" (TechCrunch) versus "to acquire" (release). The deal has not yet closed. That's a minor TechCrunch misreporting, but add to it TechCrunch's characterization of Nstein as a "startup." Nstein was founded in 2000. The company has been publicly traded (CDNX: EIN.V) for years. Nstein is not a startup.

(Time for a disclosure: Nstein was one of seven sponsors that funded an editorially independent text-analytics report I wrote last year. Nstein also paid me to speak on text-mining and semantics at three marketing seminars and in one webinar in 2009 and I own a modest number of shares in the company. I believe I have not otherwise written about the company since a 2007 comment to analyst Curt Monash's blog.)

TechCrunch's characterization of Nstein's business as "content analysis" is half of a truth. Nstein sells text-mining software that can and is used to analyze content in the form of published and publishable text, but their business focus is, and has been for several years, content management solutions. This is easily verified via a visit to Nstein's Web site. I've been told that 90% of the company's business is media and publishing, which is not hard to believe.

Nstein's focus on that largely struggling market has limited growth as compared to solution providers with roughly comparable text-analytics revenues such as Attensity, Clarabridge, NetBase, and Temis. The latter two companies also have significant media & publishing sales but overall are much more diversified.

Here's where we get to a fourth bit of TechCrunch misreporting. Add it to the three smallish TechCrunch inaccuracies I've offered and cumulatively you get something bigger, an indicator of lazy journalism. TechCrunch's present-tense statement, "For example, Nstein powers the backend of The Financial Times's semantic search engine, called Newssift," is icing on the cake. Fact is that the FT killed Newssift back in December. If you go to newssift.com, you'll see that there's nothing there.

Newssift was a nice site, combining Nstein entity extraction with Lexalytics sentiment analysis and Reel Two classification, all delivered via Endeca faceted search. (Want to learn more? Leslie Barrett, former Director of Language Technology at the Financial Times, worked on Newssift and will be speaking at the April 13 Sentiment Analysis Symposium.) Four software licenses + expensive content feeds… and no discernible revenue source. No wonder that the Financial Times pulled the plug.

Newssift failed, and now publishing focused Nstein is being acquired. It's difficult to make a go as an IT solution provider in the publishing market nowadays although I do note the existence of thriving counterexamples such as Mark Logic.

Nstein has had mixed financial results over the last few years. Per the blog-article comment of mine I cited above, the company had been losing money. CEO Luc Filiatreault turned things around prior to the current economic crisis. 2008 revenue was US$ 25.7 million, up 41% from the year before. (Early 2008 acquisition of digital asset management provider Picdar did add to the top line. Nstein had grown via other, earlier acquisitions, for instance of Web content management publisher Eurocortext in 2006.) Last published figures, however, show a 17% quarterly revenue decline, year-on-year, for 2009 Q3. Nstein folks had told me last year that core text-mining revenue was actually holding its own, that it's the content-management software that was losing ground.

Nstein started to shift focus this last fall, for instance with the November release of Semantic Site Search (S3), which uses Nstein's Text Mining Engine to annotate "concepts, categories, proper names, places, organizations, sentiment, and topics" in source documents in conjunction with a faceted search interface. Just last month, in January, Nstein hired Marten Den Haring, who had worked on collaboration and BI systems at Oracle, as executive vice president.

I had perceived Den Haring's hiring as a sign that Nstein would continue to re-focus products and solutions on markets other than media and publishing and that Nstein was prepping itself for sale with Oracle as the most likely buyer. Nstein's tools would have complemented Oracle's Universal Content Management nicely, but instead they'll fill that complementary role for content-management vendor Open Text.

All of this makes for an interesting story, when you hear the full telling.Enterprise content management vendor Open Text's planned acquisition of ECM + text-mining provider Nstein is significant text-analytics/publishing/content-management news. It gained the attention of TechCrunch, which managed to misreport the story. Here's my analysis of the news with a look at TechCrunch's mis-telling...