Time Management: Team Efficiency

Tom Limoncelli taught the second half of his Time Management series this afternoon, Advanced Time Management: Team Efficiency. This session focused less on organizing yourself, and more on keeping teams efficient. There are ways of making sysadmin or other technical teams work more efficiently, and Tom let us in on those.

Tom opened the session with a question: what team problems do you want to see addressed? There were a lot of things called out, but many of them familiar:

• Decisions that don’t stay decided.
• Getting younger sysadmins to think.
• Getting older sysadmins to stop thinking and do.
• Keeping process efficiency as teams grow.
• Dealing with application teams (which include non-technical business users)
• Sharing tribal knowledge.

Tom has written a book on this, so this was a very full session.

Did you know that sysadmins are naturally collaborative? There has been a study on it, peer reviewed even! The datacenter dragon stereotype is not strictly true! Proof! But it also means that we work better when we’re communicating better.

Better meetings mean people actually want to go to them, and things actually get decided. No one wants a talking-group.

Email is easy to overload on. As anyone who has read the Limoncelli Test (or attended the class) knows, an interrupt-driven sysadmin is not an efficient sysadmin, and constantly checking email is interruption. So it pays to limit it. He gave us ways to do it.

Another way to help with the email problem is to get better about sending it. Subjective experience has taught Tom that most people read the Subject: line, probably the first paragraph, and almost nothing after that. You have to hook your reader in those two areas if you want them to read it. And if you do it often enough you’ll build up email karma and people will be more likely to read your longer screeds just because it’s you writing it.

After the break Tom moved into ways to improve task selection efficiency. For this he drew a two axis chart, easy to hard, and rare to common. The four quadrants give you the recommended way to handle each task.

Easy and Rare, do it manually. Or script it you want, you know, whatever.
Hard and Rare, document it, and script it if it really annoys you.
Easy and Common, automate it.
Hard and Common, acquire something to help with it, since the automation challenge is likely beyond the cost of just buying (or locating and installing the OSS) tools.
Very Hard and Rare? Hire someone. Really. It’ll save time, and get them to document it.

Collaborative documents such as wikis or Google Docs really do help improve communication a number of ways. For one they’re all access. Second, they can be updated by anything with a browser these days. Third, if doc is out of date or just wrong, you can just fix it. Really, they’re a good idea.

Sysadmins hate documentation, or more accurately like documentation but hate writing the stuff. Happily, Tom has a solution for you:

Checklists.

Lots of checklists.

Checklists are easy to write as they’re just process flow. You can even make them by taking a transcript of a terminal session where you go through the process. If you do them in a collaborative document, as you should, when teammates discover wrongness they can just go in and fix the line. Checklists improve with time as they’re used! Collaborative document writing! You just need to provide the seed!

What checklist do you create first? Things that annoy you. Like creating new users, departing-user change-all-the-passwords procedures, and how to deal with backups. Get enough of them and you can demonstrate that you need a minion.

And a closing thought from Tom:

“Nothing teaches you the value of a good checklist like having to work with a bad one.”