Text Analytics Comes of Age

An Accelovation briefing earlier this week was doubly helpful in affirming my take on the maturity of the text-analytics market and in showing me that I might be doing my job wrong.

Regarding the first point: Despite an hour-long presentation by company CEO Jonathan Spier and VP of Products Jens Tellefsen, I don't have have a clue how the company's "business insight discovery technology" works. That's because the company is pitching solutions to business analysts and not to IT geeks like me, a sure sign that the underlying technologies are stable and capable, and Spier and Tellefsen steered our conversation away from tech talk.

That a vendor can center its message on what rather than how is a hallmark of a maturing market.I did take away that Accelovation processing is built around a conventional retrieve-parse/classify/rank-analyze-output pipeline. Business-analyst users access a hosted system to perform semi-structured searches against a repository that currently holds 1.7 billion "solution statements" and 1 billion "needs statements." Use is facilitated by business-domain-specific templates and built-in collaboration and workflow-management capabilities.

That information, however, is almost incidental to the company's value proposition. All that Spier and Tellefsen would say about technology is that where most text-analytics vendors focus on information extraction - on discovering entities and relationships - their intent is to discover business opportunities and generate structured competitive assessments.

Every well-conceived business information technology aims to replace tedious, time-consuming manual work. Bonus points for improved results. We're all, including consultants like me, looking for IT that can deliver on those aims.

A significant part of my job involves studying technology positioning. I look at market opportunities, product capabilities, evolving demand and the competitive landscape in order to develop and assess strategy. Necessary tasks often fit that "tedious" category - iterative Web searches, sifting through often-discursive documentation, transcribing what I find - although fortunately I also get to do higher-value analyses.

I didn't imagine, going into the Accelovation briefing, that the company might offer a product I could actually use, but their solutions just might be worth my pursuing for my own use. Whether I can afford them and whether they deliver the accuracy that I require for my work - the precision and recall - is a another story, something to assess in a deeper evaluation. Nonetheless, at this stage, Accelovation and their solutions-focused pitch have provided strong, supporting evidence that text analytics has come of age.

Seth Grimes is a principal of Alta Plana Corp., a Washington, D.C.-based consultancy specializing in large-scale analytic computing systems. Write to him at [email protected].An Accelovation briefing earlier this week was doubly helpful in affirming my take on the maturity of the text-analytics market and in showing me that I might be doing my job wrong… The company is pitching solutions to business analysts and not to IT geeks like me, a sure sign that the underlying technologies are stable and capable. That a vendor can center its message on "what" rather than "how" is a hallmark of a maturing market.