When you should automate stuff or not

One truth I sometimes forget about entrepreneurship: some tasks can downright boring 🥱

And when they repeat, you might think it’s worth automating those.

Whether it's accounting, business processes, or something else, it can take up valuable time that could be spent elsewhere. Automation can be tempting in these situations, but it's important to take a few precautions before jumping in.

1. Make sure that automating a task is worth the effort.

Automating one-off tasks doesn't make sense most of the time. For instance, I once tried to automate the translation of a file with thousands of lines and spent a few hours connecting to the DeepL API, only to be disappointed with the subpar results. I ended up having to go line by line to correct it.

2. Sometimes, automation is worth it even if it doesn't save time.

A few tasks are so tedious that they drain our energy and focus, which can be detrimental to global productivity.

I remember reading a piece of content back in 2015, and it still resonates: don’t automate tasks only based on time spent.

Read it there 👉 https://marcgg.com/blog/2015/09/30/automate-repetitive-tasks/

When I know I’ll automate something, I use a simple, four-step process:

1. Handle the task manually to ensure I know the process inside and out.

2. Check if the automation exists elsewhere (a tool, a Chrome extension, a library, a template, or just an answer on StackOverflow) not to spend too much time on it.

3. Script it: this creates a "semi-automated" process where I still have to launch the script manually each time I need it. Most of the time, I leave it at this stage and build up a collection of scripts that I can use as needed.

4. Automate it fully: define triggers and actions, and let the automation work its magic. If it's related to our product, I'll just add a new feature for users.

Some tools I use for automation include can be “No-code” (Google sheets, n8n, Zapier, Integromat/Make), low-code (Pipedream, Meta API), or fully coded (custom scripts with CRON tasks, microservices).

Lesson: don’t automate for the sake of automating. Make sure it’s worth it first!