When (and when not) to automate
Sep 4, 2024
As one of the first ops hires, you probably can multi-task better than anyone at your company, but at the end of the day, you still only have two hands and one brain.
These human limits often drive us to "find more hands," but in most cases, we can't afford or justify another ops headcount, so we have two options:
Recruit teammates to help
Get computers to do it
We will cover the first option later. Today, let's discuss when to use automation.
A quick definition
You don't need to be told what automation is, but let me add one differentiation that may be helpful. (I'm just making up these categories; if you know the proper names, please share them!)
Step Automation: Reduces or removes one step of a task. This could be as simple as improving internal admin by adding a search button or adding a hotkey to go into Do Not Disturb before a meeting. You're reducing an action from 5 seconds to 1.
Process Automation: When an end-to-end process is reduced to a single step. This could be building mini-apps in Retool to turn a complex support workaround into a single-page flow or adding CRM triggers to perform multiple tasks after a meeting.
When (and when not) to automate
I'll be honest, I (Jonni) am a chronic over-automator. I'm a bit allergic to manual work (I'm currently seeing my doctor about it 😉). Youmna falls on the opposite side of the spectrum, and self-admits that she should probably find more opportunities to automate - we all do what works for us!
Although getting systems to automate my work is arguably one of my most maximizing moves, it comes with meaningful downsides if not done at the right time.
Here are a few questions to ask yourself when considering automation:
How much time/energy will it take me to build the automation? If it's over 3x the time spent on the task itself, you may not get the ROI.
How many times have you done the task? The less you've done the task, the less you'll know what edge cases may come up when other team members try using the automation in the wild. A Step Automation is probably ideal to reduce risk.
Is the task a one-time thing or recurring? Sometimes a bug shows up or a one-time project demands manual work. Considering if this automation (or parts of it) can be reused even after this one-time use will help justify the investment in the automation.
How difficult will it be to maintain? Building is one thing, but some automations can be quite fragile. Will you have to consistently invest maintenance work to make it reliable? The best automations allow you to "set it and forget it".
Automation tooling we love
Gone are the days when non-technical operators like us had to rely on developers to build automation. We live in a low-code/no-code gold rush.
Here are a few tools that have been integral for us: