Yesterday I published a post on the Official Microsoft blog (OMB) that focused on the $9bn we spend each year on R&D and how we turn that in to products – it was accompanied by an infographic that showed how some of that process happens.
Now it’s time for a confession – for a few years during my early career at Microsoft, I thought all those billions of dollars of R&D investment went in to Microsoft Research (MSR) and I wondered where it all got spent. It turns out I was wrong. Very wrong. R&D does not = MSR. It’s obvious when I think about it but R = MSR while D = products. Therefore, our R&D investment not only pays for the work of the smart folks in MSR but also the products we release - products such as Windows, Office, Xbox, Bing, Azure and many many more.
I fell in to the stereotype trap of thinking R&D is all about folks in white lab coats working on future research. At the same time, I hadn’t grasped the nuance that development means real shipping products. It feels a bit “duh” now :) However, I confess this as a noble act (cough) as I worry that others may fall in to the same trap and then wonder, what does MSR spend all of that money on? As I noted in the OMB post, we have 850 Ph.D.-level researchers in Microsoft Research and around 40,000 developers in our product teams - that should give an indication of how we balance that $9 billion between research and the development of shipping products. I call it small r and big D.
With that little confession out of the way, we can return to normal programming and I can finish up my reporting on TechFest this week – where our small r folks in Microsoft Research show that their work is anything but small!
The R is a small part of the R&D, but it's still larger than almost any other company. I think what many wonder is how Apple spends 1/8th as much on R&D (for most of the decade it was 1/10th) but has built a larger, more valuable, and faster growing business that most think runs circles around MS in innovation. Everyone understands that MS has smart people in research. The company just doesn't seem to have a good system for getting R ideas into products it can sell.
The company just doesn't seem to have a good system for getting R ideas into products it can sell...
Like the Kinect for instance?
Mark - you've hit on one of the perceptions that I hope to change - that we don't have a good system for tech transfer from research to products. There are occasions when this is very visible (at least to tech observers) in instance such as Kinect. The story Wired UK ran about the development of Kinect gives you a real insight in to the transfer process. There are many many other examples of where research has contributed to products that go unseen - which is perfectly fine as it's stuff that we just hope customers enjoy rather than need to recognize. For tech observers though, perhaps we can offer more examples like the work that MSR does to help with spam in Hotmail and Outlook or 3D aerial views in Bing Maps or facial recognition in Windows Live Photo Gallery or ranking technology that makes multiplayer gaming in Xbox seamless in the way it finds you a player of similar skill from the millions of online players.
In terms of Apple, it's hard to do a real apples to apples comparison (pardon the pun) as our work covers a broader landscape than Apple. That's not a criticism, just an observation of the differences between the two companies and while we're often compared to Apple on the consumer devices front, out work spans from consumer, to businesses of all sizes, gaming and on through work that is seeking to address global challenges around healthcare, energy and climate. That work is less well known but nevertheless important.
hope that helps to explain to some degree
Research and Development are not purely black and white as most people think and assume. They are tightly intertwined and there is no one definition of either that will put them in their own pile. When does an idea get classified as a research idea and not as a development idea? When does an idea gain the reputation of being research? It’s not a very simple phenomenon and it requires one to understand that the essence of research is the depth of analysis one has put into understanding the problem. Research is to gain understanding of what causes the problem and providing a solution that addresses the root cause. Let me take a simple and naive example of divide by zero error. If you fix it in the code with a divide-only-if-denominator-is-not-zero, that solution may be termed as a fix. But if you go into the details of why the denominator is becoming zero and ask questions on the behavior of the data that causes the denominator to become zero, that is research.
My suggestion "Pasteur’s Quadrant: Basic Science and Technological Innovation" by Donald Stokes.
Having said this, there is a lot of research going on in MS. There are a lot of PhDs working on very interesting problems in MS also. While MSR mostly has PhDs, the sharing of thoughts across both organizations is far greater.