Note: This blog post was originally written when I was a consultant for OpenSource Connections.
Everybody loves employees and consultants who can get stuff done. Especially if you’ve made a commitment to a client or your boss, it can be very difficult to ask for an extension to that timeline. It’s much easier to ask your team to put in the extra hours to get the job done, even if that means the job won’t be done well. Sometimes leaders will even resort to claiming the product is «done», when it was never truly done at all, and needs a lot more work.
Your development team will likely respond to your calls with heroic effort. And you might actually make the deadline. But if you do it regularly, your reputation as a leader may actually suffer, and your team will probably burn out unless you can pour on enough accolades and spot-bonuses.
At some point, almost all IT leaders, including myself, have done it. That is unfortunately not much of a secret and it’s why IT teams often develop a bad reputation. The bigger secret is that a lot of developers not only regularly take on the hero role, they actually like it.
This creates a dangerous situation, where IT leaders ask for heroic efforts because it is the path of least resistance. And the development teams respond with grumbles and late nights, but also with the misconception of job security and the expectation that they will receive a lot of praise when they come in on Monday bleary-eyed (but with at least part of the job done).
Why is there so much danger in encouraging your developers to be heroes? Do a google search on «cowboy coders» or «programmer heroes» and you’ll see other blog posts that describe the whys and the pitfalls of reliance on hero development. In short, you’ll get inconsistent and buggy code that may not fit your architecture, won’t meet the customer’s needs, is not very flexible, creates technical debt you will complain about for years, and will probably need to be reworked eventually.
So how do you avoid heroes? As a CIO or technical lead, the tools of Agile offer many solutions:
- Sprint planning – restricting the work that is done to the current sprint helps discourage sudden changes of priority. Most tasks can in reality wait until the next sprint to be added, especially if the customer understands that it means the current sprint has to be put on hold or canceled.
- Daily standups and regular customer interaction – Chronic hero programmers often like to hideaway on their work, and so they can be more easily detected when you are interacting with the customer each day. There is less opportunity for the developer to hide away and not give status updates, and the standups provide the customer more immediate feedback on everything that is being held up because they sent the hero on a distracting quest.
- Team estimation – When the team estimates tasks for a sprint together, before they are assigned to people, that encourages open discussion and debate on how long something will take.
- Continuous Integration – Maintaining rules about daily code check-ins and regular integration and unit testing prevents the hero from going too far away from the mainstream code base, and gives visible signs of progress. When done properly, it also helps to reduce the bugs delivered in the rushed code.
In addition, in our OpenApproach implementation of Agile and Scrum, we use range estimation of tasks. Asking people to give you a single-point estimate of how long something will take encourages them to be overly optimistic. «Oh, I can get that done in about 2 hours.»
Really? What if it’s not as simple as you or the developer thought? Then the only solution is to put in a heroic effort if the timeline must be met.
A simple range estimate will provide a much more realistic view, and encourages a more thorough thought process: «It should only take about 2 hours of actual coding, but I’ll have to switch projects and get that site running again on my machine. Plus, sometimes those java script issues can be really though to debug, so it could take as long as 8 hours of work.»
As an Agile CIO, it’s your job to not only be understanding of your developers when they give you a range estimate, but encourage it. Don’t settle for the heroic answer of «I’ll just put in some extra time this weekend.» Get to the root cause of the problems, and make sure the customer or your boss is aware of the delays earlier rather than later. It will make all of your lives easier, and help you to avoid the pitfalls of programmer heroes.
Recent Comments