Automate it once, or do it forever.

Some tasks can feel like déjà vu. How many times have you set up a project, created and sent invoices, or run deployment scripts? These tiny tasks only take a few minutes, but you do them again, and again, and again.

You end up doing them so much that they turn into mindless chores, and these repetitive chores don’t just waste time, they drain your team's creativity and energy.

If it takes so little time to do, what's the problem?
If a simple task takes 5 minutes, but you do it 3 times a week, you will spend over one and a half working days doing that same task a year. Sounds awful, right? But if you spend an hour or two automating it, you get all that time back, plus you will:
🏆 Free up mental space for more valuable, creative work
🏆 Reduce human error
🏆 Scale your effort without increasing cost
✌️ Have fun: let’s face it, automating boring tasks is often more satisfying than repeatedly doing them

Before you automate anything, though, ask yourself:
❓ Do we understand the full process and its edge cases?
❓ Will automation reduce or eliminate errors?
❓ Is the task repetitive and predictable enough?
❓ What is the return on investment?
❓ Can this process even be automated or does it need manual human oversight?

Automation just to automate can be a bad thing. You need to focus on what you hope to achieve. Here’s some general advice:
🎯 Prioritize based on business value: automation should achieve a measurable goal and bring real value
🎯 Document everything: a year from now, future developers(and you) will need to understand the logic, and perhaps know that they even exist
🎯 Track ROI after launch: how else will you brag about time saved?
🎯 Design for long-term maintenance: requirements will change, your automation should be easy to adapt

A few examples of what we automated:
✅ Project scaffolding: project setup and initial configuration
✅ Invoice generation and time tracking reminders
✅ Deployment pipelines: except that big ol'  Go Live button, we keep that one manual, some moments still deserve ceremony.
✅ Testing

One final thought, while you should not underestimate the cost of doing the same task forever, you should also not overestimate the ROI of automation. Calculate the total cost of doing it manually (Time spent per task × Number of times it’s done = Total Cost), and compare that to the time it will take to develop the automation. If the time it takes to do the automation is larger than the total cost of doing it manually, then it's not worth it.