Wednesday, 3 November 2010

Faster, Better, Cheaper

A decade ago, NASA had some well-noted disasters with unmanned spacecraft. The Mars Polar Lander, the Lewis earth-observing satellite, and the Mars Climate Orbiter. Fortunately none cost any lives, but they all proved embarrassing to NASA, which is supposed to be the pinnacle of American scientific and engineering achievement.

What is perverse is that many of the problems could be put down to one phrase: "Faster, Better, Cheaper". This phrase was dreamt up by NASA Administrator Dan Goldin, who took up the post in the early 90's. It is now widely seen as having been a disaster, even in official reports.

So what was the problem? The problem was, in my opinion, simple. Engineers need to be able to measure things. You can measure time, speed, money, weight, distance, and any other number of metrics. In the phrase "Faster, Better, Cheaper", it is easy to measure 'faster'. Has a project been delivered faster than would have been the case under the old system? Cheaper is also easy: has the project cost less than it would under the old system?

Of course the actual metrics used will be more complex that that, but with both 'faster' and 'cheaper' the measurement is possible and obvious.

The devil is in the word 'better'. How do you measure betterness? Could a project that didn't work fully still be called better because of some arbitrary other metric? "Gee, the craft crashed into the moon instead of orbiting, but it was better because we all got more publicity!"

Perversely, 'better' allows you to mask failures, and it does not give engineers direction.

Many engineers say that it is possible only to have two out of the three; you can have faster and cheaper, but you won't get better. Or you can have faster and better, but you can't have cheaper. Then there is another viewpoint, where you can have all three. There is the following quote from that link:
No, it’s not a fact of life. It is possible. There are two cultures. The second culture is the culture that dominates the new information-age industries -- like Microsoft -- which is, you can simultaneously improve cost, schedule and performance.
And hereby lies the problem. The writer talks about cost, schedule and performance. Cost is related to 'cheaper', and schedule to 'faster'. However, performance is just one part of 'better'. A measure of 'better' might be something different from performance, depending on the mission. A 'better' on the Space Shuttle might be measured on the safety rating for the crew, whilst performance might be the maximum payload lifted, or the thrust of the engine, or any other such metric. He has altered 'Faster, Better, Cheaper' to be 'Faster, Better, Performance'.

Additionally, it is a fallacy to say that the high-tech industries such as Microsoft have any relation to the space industry. They do not. A company like Microsoft can afford to take limited risks, whereas in space they cannot. Put simply, if software goes wrong, most of the time it can be updated and fixed (there are exceptions to this; such as firmware updates, but these are relatively rare). A rocket launch or a space mission is a one-off shot; if it fails, it can cost hundreds of millions or even billions of dollars.

By all means, keep faster and cheaper. Space access needs faster and cheaper. But instead of 'better', pick another, narrower metric. For manned systems, perhaps they should use 'faster, cheaper, safer'.

No comments: