Cannot Measure Productivity (2003)

While I don’t disagree that in the end, the only meaningful measurement is business value, I’m not sure it’s fair to put the value of John’s implemented features against the value of mine. Neither me, nor John had very much say in which feature we get to implement (assuming we are both, say frontend developers. Might make more sense if John is in devops)<p>Those decisions are usually decided from higher up than developers.<p>Maybe John got the job of fixing the advertising embed code in the html template, because he’s a junior and its basically just copy paste from the documentation and since all revenue come from ads, he is now responsible for the most valuable feature.

&gt; economists are now seeing productivity increases in business due to the computer investments in the nineties. The point is that the improvements lag the investments<p>Aye. And one rarely sees a software team stay intact long enough after shipping software to make a fair evaluation of what they actually delivered. Most of the time the incentive structures cause everyone who didn&#x27;t get a promotion off the back of shipping to bail, and corporate then scavenges the remainder to staff other projects...

The best measure of productivity for a software implementation team is the quality of the product as measured by latent defects and technical debt. The team that remains to support the system will get to spend some or all of their time dealing with these. For a year or a few years or in perpetuity. Their productivity is directly related to how much time they spend unburdening themselves from that hidden cost.

And yet, people talk about productivity all the time because they can feel it.<p>People say: &quot;oh I am much more productive with C# than I am with Python&quot;.<p>Or: &quot;This guy is one of the most productive engineer I know&quot;.<p>But the issue here is that value is subjective. But is that really a problem? &quot;Value&quot; is not better defined than &quot;urgency&quot; or &quot;criticality&quot;, yet we prioritize tasks all the time based on the perceived urgency&#x2F;criticality of task A vs task B. So in the same vein, we could assign a &quot;value score&quot; of project X, that would amount to something like &quot;usefulness * difficulty * quality&quot; (to be discussed), and define productivity as that over time spent.

in software making, what&#x27;s Doable, Sellable, Wanted, and Needed, rarely overlap.. When they do, there&#x27;s a dream come true. But - how to measure that? Is that productivity? or acumen? or luck?<p>Which may hint at Exupery.. How do you measure love?

(2003)