Update: For an interesting counterpoint to the myth of the 10x engineer, check out this blog post by Shanley. My post is more focused on what makes a good developer than the 10x myth.

In the The Mythical Man-Month, Fred Brooks highlights an eye opening disparity in productivity between good and poor programmers (emphasis mine).

Programming managers have long recognized wide productivity variations between good programmers and poor ones. But the actual measured magnitudes have astounded all of us. In one of their studies, Sackman, Erickson, and Grant were measuring performance of a group of experienced programmers. Within just this group the ratios between the best and worst performances averaged about 10:1 on productivity measurements and an amazing 5:1 on program speed and space measurements!

Tortoise and Hare:

Robert Glass cites research that puts this disparity even higher in his book Facts and Fallacies of Software Engineering.

The best programmers are up to 28 times better than the worst programmers, according to “individual differences” research. Given that their pay is never commensurate, they are the biggest bargains in the software field.

In other words, the best developers are generally underpaid and the worst developers overpaid.

But don’t leave your job just yet. This is not to say that there should be a 1 to 1 correlation between productivity and pay. People should be paid by the value they bring and productivity is only part of the value proposition, albeit a big part of it. Even so, we’d expect to see some amount of correlation in pay with such a drastic productivity difference. But in general, we don’t. Why is that?

It’s because most managers don’t believe this productivity disparity despite repeated verification by multiple studies. Why should they let facts get in the way of their beliefs? That would only mean the factonistas have won.

Kidding aside, why is this productivity difference so hard to believe? Allow me to put words in the mouth of a straw-man manager.

Well how in the world can one developer write code 28 times faster than another developer?

This sort of thinking represents a common fallacy when it comes to measuring developer productivity. Productivity is not about the lines of code. A huge steaming pile of code that doesn’t get the job done is not productive. There are many aspects to developer productivity, but they all fall under one main principle (borrowing a term from the finance industry), TCO.

TCO - Total Cost Of Ownership.

In general, I’ve tried to always hire the best developers I can find. But I’ve made mistakes before. Yes, even me.

One situation that comes to mind was with a developer I had hired (under a lot of pressure to staff up I might add) at a former company. I handed off a project to this erstwhile coworker to take over. A few days go by and I don’t hear anything from the guy, so I assume things are humming along nicely.

Fast forward another few days and I swing by to see how it’s going and the developer tells me he doesn’t understand a few requirements and has been spinning his wheels trying to figure it out this whole time.

Good Developers take Ownership so You Don’t Have To

This is one of the first ways that good developers are more productive than average developers. They take ownership of a project. Rather than spend a week spinning wheels because they don’t understand a requirement, a good developer will go and grab the decision maker and squeeze out some clarity.

Likewise, a good developer doesn’t require you to prod them every few moments to make sure they are progressing. If they get overly stuck on a problem, they’ll come to you or their coworkers and resolve the problem.

A developer who can write code fast, but doesn’t take ownership of their projects is not very productive because they end up wasting yourtime.

Good Developers Write Code With Less Bugs

I once worked with a developer who was praised by my boss for being extremely fast at writing code. He sure was fast! He was also fast at introducing bugs into code. His code was sloppy and hard to understand.

The key measure that wasn’t figured into his productivity measurement was the amount of productivity lost by the QA team attempting to reproduce bugs introduced by his code, along with the time spent fixing those bugs by this developer or other developers.

Everyone focused on his time to “completion”, but not on the total cost of ownership of that code. Code is not complete when a developer says it is complete. That is not the time to stop the stopwatch. It’s when QA has had its say that you can put the stopwatch away for the moment.

As I like to say, productivity is not about speed. It’s about velocity. You can be fast, but if you’re going in the wrong direction, you’re not helping anyone.

Good Developers Write Maintainable Code

Hand in hand with writing less bugs is writing understandable maintainable code. As soon as a line of code is laid on the screen, you’re in maintenance mode on that piece of code.

Code that is brittle and difficult to change wastes hours and hours of developer cycles when trying to amend a system with updates and new features. By writing maintainable code, a good developer can make these changes more quickly and also improves the productivity of his or her team members who later have to work on such code.

Good Developers Do More With Less Code

Another hallmark of a good developer is that they know when not to write code. As a friend always tells me

Why build what you can buy? Why buy what you can borrow? Why borrow what you can steal?

With a few exceptions, the NIH (Not Invented Here) syndrome is a pathological productivity killer. I’ve seen developers start out to write their own form validation framework until I point out that there is already one built in to ASP.NET that does the job (It’s not perfect, but it’s better than the one I saw being written).

All of that time spent reinventing the wheel is wasted because someone else has already written that code for you. And in many cases, did a better job as it was their only focus. In such a situation, finding an existing library that gets the job done can provide a huge productivity boost.

The caveat in this case is to be careful to avoid non-extensible and rigid 3rd party libraries, especially for very specialized requirements. You might a lot of time trying to fit a round peg in a square box.

Even when you must invent here, good developers tend to write less (but still readable) code that does more. For example, rather than build a state machine to parse out text from a big string, a good developer might use a regular expression (ok, some will say that a regex is not readable. Still more readable than hundreds of lines of text parsing code).

Back to TCO

Each of these characteristics I’ve listed keeps the total cost of ownership of a good developer low. Please don’t let the term ownership distract you. What I mean here is the cost to the company for having such a developer on the payroll.

By writing less code that does more, and by writing maintainable code that has fewer bugs, a good developer takes pressure off of the QA staff, coworkers, and management, increasing productivity for everyone around. This is why numbers such as 28 times productivity are possible and might even seem low when you look at the big picture.

Hopefully seeing this perspective will convince managers that good developers really are as productive as the studies show. Negotiating a 28x pay increase on the other hand, is an exercise left to the reader.