The Downtempo Blog

Extremely thought provoking piece thought provoking piece in this month’s IEEE, by Tom DeMarco::

“You can’t control what you can’t measure.” … Implicit in the quote … is that control is an important aspect, maybe the most important, of any software project. But it isn’t. Many projects have proceeded without much control but managed to produce wonderful products such as GoogleEarth or Wikipedia.

To understand control’s real role, you need to distinguish between two drastically different kinds of projects:

Project A will eventually cost about a million dollars and produce value of around $1.1 million.

Project B will eventually cost about a million dollars and produce value of more than $50 million.

What’s immediately apparent is that control is really important for Project A but almost not at all important for Project B. This leads us to the odd conclusion that strict control is something that matters a lot on relatively useless projects and much less on useful projects. It suggests that the more you focus on control, the more likely you’re working on a project that’s striving to deliver something of relatively minor value. (emphasis mine)

As a staunch proponent of metrics for software projects, this passage was quite upsetting.  I am committed to measurement.  I take things apart, I tune them, and I make them better.  The author suggests that my obsession with performance monitoring, continuous integration, code coverage, bug counts, and product backlogs is not only an expensive indulgence, but a warning sign that the projects themselves are of little value.

This stopped me in my tracks.  I thought about it for a few days, and I’ve decided that for a startup I fundamentally disagree.

In an enormous company, he’s got a valid point.  A million dollar project which eeks out a meager 10% return is a waste of resources and no amount of control is going to turn a 10% return into a 10x return.

But things are vastly different for entrepreneurs.  We own our little companies, with shallow pockets.  Everyday we work to mitigate risk.  We are trying to effectively manage short runways.  And there is a constant battle for time, our most precious resource.

If you work for a Fortune 500 company, your runway is essentially endless.  In this environment, I’d agree that you must peel away layers of process and structure to allow the planes to get off the ground more quickly.

Startups are different.  Startups are already implicitly pursuing the $50 million opportunity, but their resources are fiercely constrained.  There’s never enough time, never enough money, and never enough patience from your investors.  An entrepreneur’s time is spent hustling to avoid crashing into the field at the end of a potentially ever shorter runway.  To extend DeMarco’s argument into this environment is to advocate flying blind, with no instruments.

If you’ve only got eight months of cash in the bank, you better have a darn good idea of what features you can build in that amount of time.  And it will be far easier to raise money and increase the length of your runway once you can quantify the elusive path to profitability to your investors.