In Thursday’s New York Times (and NPR, and virtually every other news organization in the world), we learned that Microsoft was cutting back:

The layoffs span across research, sales, finance and technology roles, the company said.

To be slightly more precise – as some later prints of the article mention --

The layoffs span across research & development, sales, finance and technology roles, the company said.

We run into this in other contexts, too. In Microsoft’s annual balance sheet, we learn that R&D made up $2.2 billion in 2008—and, again, that gets rephrased as Microsoft spending billions on research, both from Microsofties and others (such as this Boston globe article).

There’s an important difference between “Research & Development” and “Research” at Microsoft—and at most other software companies. It’s a natural mistake. At most businesses,”R&D” is a fairly small bit of the budget, and its largely undifferentiated. At a car company, the amount of money that goes into designing next year’s car is far smaller than the big costs: sheet metal, assembly, transportation. Cars are expensive things, even if you have the blueprint in hand. There’s a whole lot of people who are out there welding together car parts. At a car company, too, the R&D are not necessarily distinct (at least to the outside world): there company dedicates a chunk of its budget to both improving underlying technologies and developing future years’ cars.

At Boeing, say, the costs are more balanced. It costs a lot to figure out how to build a 787, and a lot to build it.

In contrast, the manufacturing cost of a copy of Windows is virtually nothing. Stamping a DVD or printing off a license takes all of a few seconds. There are ongoing costs to support that license, and to keep that machine running stably, but the real cost of Windows comes early in—when a team of humans sits there, tapping away at computers, writing new software. Same for Office, or any other product. (Not as much for XBox, which has a substantial hardware costs.) That team of people are called “developers”, and that slew of developers is really what makes up the bulk.

So at Microsoft, R&D is a huge chunk of the budget, and manufacturing is smaller.

Within R&D, we have researchers and developers—and they are meaningfully distinct roles. There are some 700-odd researchers at Microsoft Research; we try to figure out how to do things that no one has done before. Preferably, things that are generally thought to be impossible. There are tens of thousands of developers at Microsoft; they build operating systems and word processors and databases and all sorts of things that we generally do know how to do, but want to do better.

(In addition, there are a number of smaller hybrid “advanced development” groups—Live Labs, Virtual Earth Labs, Office Labs—which sit somewhere in between.)

What this means is that when you see that “R&D” figure—or a reporter short-handedly writing “research”—it’s not just the 700-or-so researchers in the US, China, India, and the UK. It’s also several tens of thousands of others at Microsoft who makes products.

(Thanks to Kevin for some clarifications on this.)