Some Random Nerd

View Original

Scott's Law of Automation

I've always liked the idea of being one of those people who have a law named after them- this is my shot. (Apparently "Thompson" is the 12th most common surname in England, so I'm assuming "Scott's Law" would be a little more specific to me- given that "SomeRandomNerd's Law" is unlikely to catch on...

"If you do what you always did then you'll get what you always got" - (ie. I've done nothing and still not got a law named after me); Scott's Law of Automation (which I hope will simply be known as "Scott's Law"1) states that;

Automation should help people do tasks - not take them away.
Any attempt by an organisation to use technology to automate a task that isn't already being carried out by people in that organisation, without properly consulting those people about exactly what the task is and what needs to be done to carry it out, is doomed to failure.

(Feel free to substitute "automation" with "computers" or "software" or "algorithms" - honestly, so long as its "Scott's Law", I'm not precious.)

If you want to automate a task, start out with a person actually doing the task - and then automate it. Otherwise, you won't really know who is complaining about how the task is done, what parts of the task involve challenges that require humans with thoughts and ideas to solve and what parts of the job are boring, repetitive and easily automated. You won't really understand the politics and stakeholders that get involved when the task gets done well, or gets done badly.

In short, it takes people actually doing it to truly understand a task - and if the only people involved are invested in managing a project that takes the people out of the task (ie. thats what they get paid for), then those people are fundamentally incentivised to only understand that task from that limited perspective.

Its tricky to share case studies to support anything about business failure without bad-mouthing your employer (at best...) But if you search for any guide to "automation" in a business context, see how much of the advice - when you strip out the project management tips or technical details - comes back to Scott's Law and the importance of the people doing the task that is being automated. I'll bet that - if its written by someone with successful experience in automation projects - that it will be most, if not all of it.

For example, look at what Capita have to say about What the UK’s largest automation project can teach organisations about efficiency- four points about project management and just one point that is actually about automation;

Automation to support people, not replace them


While boosting KPIs was certainly an important consideration in our automation strategy, so was improving our people’s working lives. The first priority was to prove that this wasn’t about taking jobs. In fact, we’ve actually created more than 150 roles, as we grew our automation practice from eight to 177 people.

Its not unique to Capita.

  1. Not that I have any issue with the pre-existing Scott's Law, but I think mine is more widely applicable outside of Illinois, and will still be valid when/if self-driving cars become more widespread.