We build automation and advise against it more often than we recommend it.
Automation pays off for work that is repetitive, has clear rules and is prone to manual errors. Manual control is better at low frequencies, for processes that are still changing, and for work where judgment or customer contact is the core. The calculation decides, not the irritation.
NEXT TO EACH OTHER
The differences at a glance
| ASPECT | Manually hold | Automate |
|---|---|---|
| Upfront costs | None | One-time construction |
| Cost per time | Time, every time | Virtually zero |
| Error prone | Yes, especially when retyping | Consistent, but errors scale |
| Flexibility | High, you adapt immediately | Change requires adjustment |
| With exceptions | Strong | Weak, needs to be taken care of |
| With growth | Takes proportionately more time | Scales without additional costs |
WHEN WHAT
When which
The calculation: time per week times hourly rate times 46 weeks, compared to construction costs. If the payback period exceeds two years, then manual maintenance is usually the right answer.
Also pay attention to stability. You automate a process that will work differently in six months' time twice. Waiting is then not a postponement but the cheaper route.
- Manual at low frequency — a few times a year doesn't earn anything back.
- Manually in case of a changing process — wait until the method is established.
- Manually with many exceptions — rules that always differ are not rules.
- Manually where contact is the value — don't automate a customer relationship.
- Automate your daily work — there is almost always profit there.
- Automate retyping — double entry is error-prone and pointless.
- Automate with growth — scales without additional people.
- Automate with clear rules — if the outcome is predictable.
FREQUENTLY ASKED QUESTIONS
More about the choice
Do you really ever advise against automation?
Regularly, and that is not out of modesty. A project that does not pay for itself results in a dissatisfied customer, and that costs us more than the turnover generates.
What if the work is especially annoying?
Irritation is a bad unit of account. Sometimes the better solution is to eliminate or redesign the process instead of automating it: that costs nothing and yields more.
Can I start small?
Yes, and we usually recommend that. First, automate the one step that goes wrong most often. After a few months you will know whether the rest is worth it.
CONTINUE READING
Of course, continue reading about this subject.
It pays off your case?
Tell us which work will return and how often. We do the calculation and also say if it is negative.