The punchline *

Use checklists to make complicated work easier and more reliable. They can make a huge difference, even for highly skilled professionals. The checklist has to be really good.

About the book

The Checklist Manifesto

How To Get Things Right
by Atul Gawande
Published 2010
Bio: Surgeon and Associate Professor at Harvard Medical School

The book is engaging, super readable, and often terrifying with stories of people at death’s door in theatre. Many improvement type books peak early with a central idea and then pad it out for the rest of the book. The Checklist gets better and better all the way through.

Key things I took from the book

  • In medicine (or other complex areas) people use to die (failure) because we were ignorant – we didn’t know enough about how things work. Now people die because of complexity, we know a lot more, but it’s too much to retain and act on. We sometimes call in ineptitude, but maybe we’re just expecting too much.
  • The biggest benefit of checklists is they can get you a better outcome, with no improvement in skill.
  • Medicine (and many other areas) have been focused on improving components. But you only get small gains this way. The biggest improvements come from improving how the parts of a system fit together.
  • Aviation is the master of checklists – this is why flying is so safe. (Sully, the miracle on the Hudson, can be largely attributed to calmly following a checklist.)
  • The activation phenomenon: when everyone says who they are and what their concerns are at the start of a procedure, they’re much more likely to have the boldness to speak up when they need to later.
  • There are two types of checklist: Read-do, where you read each item off and then do it. Do-confirm, where you do what you know and then stop at determined points to check the list.
  • Expert’s egos get hurt when they’re asked to use a checklist. But they definitely want a checklist used when surgery is performed on them. Takeaway: make it personal.
  • No one sits down and writes an awesome checklist. They have to be tested and improved till you develop something useful.
  • Checklists need flexibility. Things change. One item on the list might be just to make sure the various parties have got together and had a discussion. (When things go wrong you can have a checklist for how to recover from each situation.)
  • Good checklists don’t try and do too much. They need to include things that are sometimes missed. Importance in not necessarily an indicator something should be on a the list – some important things might never be missed. Checklists take care of the dumb stuff that you don’t need to occupy your brain with. They reduce the cognitive load.

Ideas for the software world

It seems obvious that there are many instances where a checklist could be incorporated into software, and could fit nicely into a process.

Checklists in software development could probably prevent a lot of bugs getting through. Could be a simple checklist before making a commit or a more comprehensive list before deploying.

A checklist before we start building a new feature could make a huge difference. Checking we understand the business requirements and we agree on the approach,

* Seth Godin says he just reads a book till he gets the punchline.