IBM: Watson Needs A New Sales Mentality and Model

We believe that super computing is an opportunity for a limited number of global players in the future. Super computing is one of the areas where a tangible connection can be made to the whole ephemeral concept of Big Data.

Particle physics research, running large scale climate simulations, and weather forecasting are examples of a few areas where customers are deploying super computer configurations and buying new iron. These are long cycle, episodic big ticket sales. This is not what IBM is used to doing.

Personalized medicine and supporting patient diagnoses with simulations and statistical analyses have a long way to go, and it's not clear that this will generate the kind of quick fix revenue growth that investors like.

One of our most widely read posts talks about the IBM Road Map and some of its assertions, which might not be consistent.  The problem is, as it is for most of the Tech Giants, revenue growth.  We noted,

 "The Road Map assumes, on average, 11% a year in constant currency growth contribution from IBM's "growth markets," which means non-North America and developed Europe.  It won't be easy."

 Markets outside North America and developed Europe surely won't be driven by super computing sales. And, furthermore they will likely vary too much quarter-to-quarter to move the consolidated revenue number on a consistent basis.

So, back to super computing. Much of the hype surrounding Watson comes from IBM's own PR that attempts to differentiate itself from other super computing players by talking about "machine-based learning." That is, the machine starts working on a complex business or process and quickly learns where it can produce better results and adjusts itself.  Some of the thought behind this is futuristic, but much of it comes by analogy from Deep Blue/Watson's work on chess and Jeopardy! where this process worked well.

The big difference?  In both these cases, the rule books were rather small and rigidly defined.  They were fixed and would not be improved.  Even though chess has lots of combinations from a move, the whole computational matrix had boundaries defined by the board, numbers of pieces, and rules limiting the kinds of moves a piece could make.

A machine learning a business today is hype.  Today's WSJ article essentially has the customers saying so.

"For example, Watson's basic learning process requires IBM engineers to master the technicalities of a customer's business—and translate those requirements into usable software. The process has been arduous."

The first problem is that most customers don't understand their businesses well enough to document the processes in flow charts or decision trees.  Heuristics are in place that work currently, but many of these reside in people or groups.  All of this has to be documented, processed and checked by the IBM engineers and then written into software.  Only then can the machines try their hand at simulating, modeling and back testing the results.  This process, as opposed to the hyped "machine learning" is time and people intensive.

The sales process for this kind of deployment is not the kind that tech salesmen like, viz. find an upgrade area and make the sale on specs, provide financing, and book it just like you've done for years.  Aftermarket support can be done by relatively lower cost resources doing traditional IT fixes.

The kind of post-installation support, including business consulting as opposed to IT consulting, required at a large medical institution is going to totally different from what IBM is used to providing.

"So far, just a handful of customers are using Watson in their daily business. With the supercomputer's help, health insurer WellPoint Inc. determines if doctors' requested treatments meet company guidelines and a patient's insurance policy. Elizabeth Bigham, a WellPoint vice president, said Watson initially took too long to "learn" WellPoint's policies."

So, the IBM CEO felt confident, based on the enthusiasm of her senior sales exec that Watson would have the fastest path to $1 billion of revenue of any venture in the company's history.  IBM really needs to do a lot less of the tiresome "Smarter Planet" brand building, and a lot more rethinking of its sales processes, organizational structure and sales execution before the path to journey's end on the "Road Map" becomes clear to investors.