From 1992 – when I joined the mechanical engineering degree class in TKM College, till today – I have been a fan of Analytics.In fact, I am pretty sure it is the engineering education that put this fascination in me. And it was my statistics professor Mr. Kalyanaraman who took it to the next level.
Nothing fascinated me more than numbers and making inferences based on them. It was not as if I didn’t realize that texts and pictures and all the so called “unstructured” data was very important – it is just that I always felt that there was plenty more to be done in the “small data” world of numbers, before any one worried about “big data” . I have kept on questioning the value of the insight that will come out of big data for most companies, since they cannot even make decisions based on relatively small and highly structured data from pre-defined sources.
And then, along came IBM Watson and it changed my perspective on analytics and big data completely. Although I work in IBM, I don’t work in the team that works on Watson directly. If I am envious of any one professionally – it is that group of colleagues who get to work on Watson. Watson captured my imagination from the first day I heard about its plans to play Jeopardy on public TV.
Now, god knows how I don’t suffer marketing . I attribute it to the compulsory marketing classes I had to take in B School. And the irony is that IBM has world class marketing. So when IBM trumpeted Watson from the roof tops, my natural instinct was to cringe. But as I thought through the implications – it became clear that Jeopardy was the perfect way for IBM to avoid evolutionary steps, and make a grand leap into the future of analytics. Jeopardy needed everything – ability to consume big data with no structure, ability to understand natural language, truly massively parallel processing, ability to work on commodity hardware, lightning speed, ability to make a decision, ability to learn and many more. And it was a safe test bed to see if technology can stand up to that stress in an environment that is not “life and death” types, but useful enough to make a determination if this has a future.
Right after Watson won Jeopardy against the “human” champions, the IBM team started focusing it on real world problems. And this is where I was most fascinated by the choices of that team and its leader, Manoj Saxena.
IBM has a huge army of smart sales people, who could have very easily sold Watson in some form to a large number of clients across the globe.It would not have been hard at all – my own clients have asked me multiple times how they can use Watson to help their business, without me having to bring it up. As we know, IBM is a publicly traded company with a published roadmap for earnings till 2015. But instead of taking a short term view of cashing in right away, they took a long term view of proving it out thoroughly in the real world with real customers.
Instead of trying to do too many things across all the geographic regions that IBM does business in, they chose to focus on a small number of very specific high value use cases in healthcare, insurance, banking etc. And they partnered with some of the best in class clients in those industries to do so – and in public,not behind closed doors. Now, that is good marketing – the kind I can relate to. Let the customers declare the vision and the success, not the vendor.
The use case that makes me most excited is the cancer treatment one where IBM is teaming with Memorial Sloan-Kettering Cancer Center. Like everything else, there is of course a commercial angle to this – and I can imagine this to earn IBM good revenue. But that revenue could also have come from many other use cases. It is the humanitarian angle that impresses me the most. Cancer research and knowledge can now be spread across the world in very short time once this project succeeds. Doctors outside major research hospitals can reduce a lot of dependence on opinions and guess work and experience, and do a lot more “evidence based” decisions. Of course I don’t expect Watson to ever replace a doctor, but Watson has the potential to be the strongest weapon an oncologist has in the fight against the deadly disease. That is not evolutionary – it is revolutionary. It makes me wonder how many other big problems can be solved by the judicious use of analytics theory and technology.
Please watch this and listen to Dr Norton explain this
And finally, I like the sense of reality the IBM team and the clients who are partnering with them display. They clearly explain what Watson can and cannot do , and how long it will take to get there. Now, I know a lot of my friends like innovations to be out in the market quickly – and I understand where they come from. So on this cancer treatment use case, I pinged a few friends who are doctors in India, who I know from childhood to understand more. It sounded like on an average it takes anywhere from 8 to 12 years according to them for information on diagnosis and treatment to become common knowledge from the time it is published. According to them, they will be thrilled if they can cut that time by a third. So even if Watson can start being of help to cancer patients in couple of years, it will be a big deal, and quite fast in “time to market” .
I am sure the Watson team will have its ups and downs in this journey – but I think it will be well worth the proverbial blood, sweat and tears. I wish them the best. And tonight I will dream of doctors in my Grandom’s village having a pocket Watson with them when they help their patients fight their diseases.