A Formula for Selling Content Processing Licenses

January 23, 2014

Do equations sell? Some color:

I know that I received negative feedback when I described the mathematical procedures used for Google’s semantic search inventions. I receive presentations and links to presentations frequently. Few of these contain mathematical expressions. In my forthcoming no-cost discussion of Autonomy from 1996 to 2007, I include one equation. I learned my lesson. Today’s search and content processing truth seekers want training wheels, not higher level math. I find this interesting because as systems become easier to use, the fancy math becomes more important.

Anyway, imagine my surprise when I received a link to a company founded 14 years ago. The outfit does business as Digital Reasoning, and it competes with Palantir (a segment superstar), IBM i2 (the industry leader for relationship analysis), and Recorded Future (backed, in part, by the Google). Dozens of other companies chase revenues in this content processing sector. Today’s New York Times includes a content marketing home run by an outfit called YarcData. You can find this op ed piece by Tim White on page A 23 of the dead tree version of the paper I received this morning (January 23, 2014). Now that’s a search engine optimization Pandas and the Times’s demographic can love.

To the presentation. My link points to Paragon Science at http://slidesha.re/1jpXAGd. I was logged in automatically, so you may have to register to flip through the slide deck.

Navigate to slides 33 and following. Slides 1 to 32 review how text has been parsed for decades. The snappy stuff kicks in on page 33. There are some incomprehensible graphics. These Hollywood style data visualizations are colorful. I, unlike the 20 somethings who devour this approach to information, have a tough time figuring out what I am supposed to glean.

At slide 42, I am introduced to “dynamic cluster analysis.” The approach echoes the methods developed by Dr. Ron Sacks-Davis in the late 1970s and embedded in some of the routines of the 1980 system that a decade later became better known as InQuirion and then TeraText.

At slide 44, the fun begins. Here’s an example which I am sure you will recall from your class in chaos mathematics. If you can’t locate your class notes, you can get a refresher at http://bit.ly/1mKR3G9 courtesy of Cal Tech, home of the easy math classes as I learned during my stint at Halliburton Nuclear Utility Services. The tough math classes were taught at MIT, the outfit that broke new ground in industry sponsored educational methods.

image

As the reader will easily see:

image

On a related point:

image

Applying these and other methods to the Enron email corpus, it is obvious that:

image

Let’s pull up the galloping pony. Whoa, Nellie.

There is considerable confusion about search and content processing. One can observe this by scanning the LinkedIn thread “Semantic Search Technology—Does It Actually Exist?” The thread has been active since January 2010. Let me invoke some Laplacian methods to confirm that the topic has lived and breathed for three years and 13 days. Close enough for horseshoes and refined Bayesian methods.

If the experts in this discussion group cannot provide closure to the fact of existence of a technology, will doing lectures about advanced numerical recipes clarify the confusion and baloney in the search and content marketing sector.

Several observations:

  1. Marketers are growing desperate to find a way to hook their technology to a solution that will close a deal. However, when the experts cannot figure out what’s what, techno-babble may be one path forward in a frantic effort to generate big bucks.
  2. Customers and prospects are even less likely to know what the heck the vendors spouting math are talking about. The math is less important than PowerPoint ready graphics. I understand the need for eye candy. Why not focus on that and skip the “We’re really smart” stuff.
  3. Technologists who develop solutions and then try to find a problem to solve may have the cart before the horse.

Whether it is content marketing in the New York Times or references to math that a tiny percentage of the world’s educated can understand, search and content marketing vendors are struggling to find a way to sell.

Should search and content processing companies follow IBM’s path? Probably not. Should firms in this sector chase the consumer approach. Google just delivers what boosts its ad revenue. No explanations necessary, thank you.

The problem is that there are quite a few companies that have taken venture funding. Now those companies want revenue and their money back. Does math sell? Does a game show win sell? Does a New York Times op ed sell?

My hunch is that we are living in the midst of a case study about how companies under pressure try to generate interest in what is a commoditized utility function. I find the essays, the slide decks, and even the techno-glop fascinating.

For those whose livelihoods and self respect depend on making search and content marketing systems pay the bills, life is going to be darned interesting in 2014.

Stephen E Arnold, January 23, 2014

Comments

One Response to “A Formula for Selling Content Processing Licenses”

  1. iMedia Designs on February 11th, 2014 6:27 am

    That’s great, thanks.

  • Archives

  • Recent Posts

  • Meta