My Photo

« Attention Variety | Main | What Happend To Technorati Charts? »

June 10, 2007


Alan Wilensky

I would never presume to refute THE leading light of the text mining industry, Matt Hurst, and I agree that all technologies that break new ground go through an adoption / refinement cycle.

That said, the six month study I conducted for France Telecom showed certain weaknesses that many end-user clients of text mining and brand monitoring services found noteworthy enough to mention to me, in the course of my survey. These folks have budgets that are able to encompass standard brand equity consulting from the industry giants, and the new-age text mining and brand monitoring services now emerging.

But more important, the trilogy of articles on my blog point to a more important and vital strategic issue beyond the current weaknesses of the state of the art - this is the fact that the practitioners of record for brand monitoring have missed the fat part of the market, the brand intermediaries, the mid-market, due to the fact that serving them is a more subtle and nuanced undertaking.

There is more to my observations than pointing out that sentiment analysis is weak, provides little fodder for creating decision support matrices, and misses the most subtle and important conversational topics that are the spring from which redress seeking and declarations of the consumer's perceptions of product performance and interaction outcomes gush forth from. ebay is pursuing this work now.

I invite the readers to skim the other articles where I spell this out and propose an outer architecture for detecting these dynamics.

And, I am sure that great scientists, like Dr. Hurst, will evolve the state of the art and make better tools. Certainly, that's what I and Dr. Charles Martin are pitching.

Matthew Hurst


Can you share with us the vendor's that you reviewed? Was the study for english language data or french? Was it a brand based study or product/service?

Oh, and thanks for the sarcasm: refreshing to hear.

Alan Wilensky

Sarcasm, moi ? False modestly, et tu? You, Matt, are the current expert of record in the field. The market most in need is not being served.

You can probably guess that my NDA precludes verbatim elements of the study, but Karthik came to the lab. I have written about Cymfony, and how their acquisition by TNS is a repudiation of targeting text mining in order to mimic the dominant brand monitoring by the big boys.

And, I think that, yes, the strategic alignment of the products/services of the vendors I covered will go through a maturation and revelation cycle.

It was not meant to be a product vendor book, ala Gartner or Forrester. It was an (English Language)survey of the perceived applicability of such services:

1) Are the dashboards really all that sexy?

2) study of the latency of the setup of the CSA campaings (a real sticking point, btw).

3) Are the output of such surveys useful for steering decisions, or just interventions?

4) as I mentioned before, the market is missing the real opportunity. This article:

and this paper:

and this:

explain what we are getting at.

Again, it's not that Sentiment Scoring is bad, it is that the market most in need is not being served.

The comments to this entry are closed.

Twitter Updates

    follow me on Twitter

    March 2016

    Sun Mon Tue Wed Thu Fri Sat
        1 2 3 4 5
    6 7 8 9 10 11 12
    13 14 15 16 17 18 19
    20 21 22 23 24 25 26
    27 28 29 30 31    


    Blog powered by Typepad