Always automate, except when you shouldn't
I love automation.
I started my career as an Excel worm, and I hated it, so I automated my tasks.
That’s how I got into coding. I used Excel macros to make my life easier: I automate the smallest tasks you can imagine.
But I just saw this xkcd:
There is a turning point after which automation is unwise. But we rarely translate this into numbers, so here it is for you.
I like this table, but I disagree with it.
It’s a learning curve
Every tool you use for automation has a learning curve. As a beginner, if you want to automate based on this table, you will never automate anything.
The first few projects will be a “Waste of time”. But is it?
What you (don't) gain in time pays huge dividends in experience.
As you automate more and more, it will take less and less time to achieve things. With experience, you can move the turning point. The more experience you have, the less you spend on automating the task and you save more.
Automation is more like investing. The earlier you start, the better. It will increase your timeframe so you can accumulate savings.
Compare my awesome chart above with this:
Automate the boring stuff.
The first things you should automate are the tasks you hate to do. I don’t care if I spend extra time automating something that causes pain every time I do.
The emotional gains of eliminating something painful are worth the extra time.
You can not just save time but remove overheads.
I trust machines more
With automation, you can also remove the error factor.
Humans tend to make mistakes. If you repeat the tasks over and over you increase the chances for error.
Great automation is error-proof, so with it, you can decrease the chances of mistakes.
It’s usually not just you
Think about other users as well. Your automation will be useful to colleagues, so you can save time for your team.
Or if you automate something at home for a family of 4 then time saving will be 4x.
Timeframe
The table uses a 5 year time span.
So many things will change over 5 years:
The software you use
The environment you work in
Or you don’t do the task anymore.
If you have regular changes, you will need to adjust your automations every time. Automation turns into a painful thing in this case. Time to adjust > time saved. Be careful here!
On the other hand, there are some tasks that we do for longer periods than 5 years.
Conclusion
I like that the xkcd table outlined the math for us, but you should always adjust it to your needs.
Consider these:
The learning curve
Timeframe
How painful the original task is
How many people can save time with it
The mistakes you can dodge
My approach: Automate 99% of the time!
And what is yours? Let me know in the comments.