Thursday, April 17, 2008

Text Analytics Revisited

A week ago I posted about the Attensity text analytics tool that one of my clients was looking to purchase. The post garnered a bit of interest. Most notably, Sid Banerjee of Clarabridge (a competing product) left a long and well thought out post on how he saw the market unfolding and how his company was positioned in it.

I wanted to do a follow-up post as I still think there are some issues with the way these text analytics tools promote themselves. This is not at all a commentary on their value to an organization - I think they are extremely valuable - it's more a comment on where I think the sweet-spot is in promotion of the idea.

My comments...

  • Don't be a tool. Anyone can claim to make a great hammer, but few people wield it expertly. You have to concentrate on selling solutions - which means knowing what the problems and issues are.
  • Don't be vague about what it is you do. Be careful in the promotion of any slogan with 'insight', 'actionable insight', 'transform', intelligence', 'actionable intelligence', etc. You get the idea. These are meaningless at the coal face. Be proud (and promote) real problems client's have solved using your tool. This gives prospective clients a far better idea of how it might apply to them.
  • Don't try and re-write the rules of marketing and branding. Yes these tools can be good, but they are not substitutes for traditional research methods. Believe me, you're going to come off looking like idiots if you get into a debate over the merits of text-analytics versus customer qualitative research. There is a whole different spectrum where text-analytics is useful, and it doesn't dovetail with traditional research as much as you think.
  • Sell as high up as you can. While these products are probably understood best by analysts, you want to sell them to CEOs. Any intelligent CEO is going to do a quick back-of-the envelope equation on the number of feedback points they have, the huge amount of information from these points, and the time and man-power to process it (or the opportunity cost of not-processing it) and realize the tool is worth it. Then tell them to factor 5 year growth in information received and they might just write you a check on the spot.
  • Don't use a technology story. No one cares. Gartner analysts might. But they are just strange. Marketers in particular don't care about technology - at least not in a geekish way.
  • Under sell. Text analytic tools are very powerful for certain problems. At some point, people 'get it'. They see what it can do and their eyes light up. Under-selling it lets them add enthusiasm. You want them to be enthusiastic, not you.
These points are of course based on my limited experience of seeing text-analytics sales presentations. So digest them with that in mind. But I do think they are all valid.

Like Sid pointed out, I think this whole area is really exciting and well timed given the huge amount of information flooding into companies these days. Sorting through this information is no small task.

Digg this
Sphere: Related Content


Justin Langseth said...

Paul, your comments and thoughts are consistent with what we have been experiencing in this market, and your suggestions are spot-on. Welcome to the world of analyst/bloggers who are tracking this space. You may want to take a look at Curt Monash's writings on this space and the companies in it, ones about us and some related companies are at

- Justin Langseth
President & CTO
Clarabridge, Inc.

Paul Soldera said...

Hi Justin, thanks for the comment. Curt's blog was buried deep in my RSS reader, I will have to start looking at it more closely.

It is an interesting area and one that dovetails a lot with work I am doing - mainly around customer analytics for Marketers. Going to try and keep a close eye on the space.


Justin Langseth said...

You may want to consider attending the text analytics conference in June in Boston.

Also we published a new whitepaper today on the customer analytics use cases for text analysis, which you can get from

- Justin

Sid Banerjee said...


Great post -- you've provided a great distillation of the reason to think about, and seek out solutions from providers who've clearly demonstrated the ability to bring business value, not just technology whizbang capability when considering text analytics in the enterprise. If you do make it up to the Text Analytics show in Boston, I'd love to catch up face to face. Justin will be there, presenting, and speaking on a panel at the event.