Moving five contacts into a spreadsheet each week may be manageable. Doing it fifty times a day deserves more attention. In both cases, the useful measure is how much work it takes to leave each record correct.
I would not cast automation as a “first employee.” A person can handle exceptions, attend to someone and switch tasks. A workflow has a defined scope and needs someone to respond when it goes beyond it. Comparing them as interchangeable hides work and costs.
Organise before connecting
Take a specific task: copying contacts from a form into the business register. Before automating it, I would decide which fields are needed, how to recognise a duplicate contact and who receives incomplete cases.
Sometimes it is enough to reduce fields or use an export the tool already offers. If the process changes every week because you are still defining the service, it may be better to stabilise it before connecting several applications.
Frequency helps set priorities, but does not establish a universal minimum. A monthly task with a serious error may deserve controls; a daily task lasting a few seconds may not justify implementation.
A small calculation, including exceptions
This example is hypothetical. I assume 100 records a month and three manual minutes per record: 300 minutes a month. With a workflow, I retain half a minute of review for each: 50 minutes. Five cases need four extra minutes to correct them: another 20. I add half an hour a month to check connections and adjust rules.
That leaves 300 − 50 − 20 − 30 = 200 minutes, or three hours and twenty minutes a month. At an assumed value of €25 per hour, that is €83.33 of capacity. If tools cost €30 a month, the balance is €53.33. A hypothetical €600 setup would equal roughly 11.25 months of that balance.
I have not deducted maintenance twice: its minutes are already included. Nor is €53.33 necessarily a payment you stop making; if staff costs remain unchanged, you recover time. To justify cash recovered, I would need to identify expenses that disappear or reasonable additional income.
At ten records a month, the same implementation and maintenance may not make sense. At higher volumes, I would measure again: I would not assume that moving from five to five hundred cases keeps the same exceptions and limits.
Who takes care of failures?
The budget should include support, application changes and a manual way to continue. If you choose to host a tool yourself, there is operational work. The n8n self-hosting documentation covers configuration, security, monitoring and updates. A self-hosted installation does not mean a task without maintenance. For reference, n8n Cloud Starter costs €20/month, billed annually (accessed: 11 October 2026).
I would start with a limited trial that keeps the original record and allows the result to be checked before use. I would measure interruptions too: if you need to investigate the workflow every morning, that time is part of the new process.
You can choose to wait without neglecting the business. Writing down the procedure, reducing duplicates and measuring for a week already helps you choose. Automation makes sense when it addresses a specific burden and you can sustain its upkeep.
Sources
- n8n: self-hosting — Accessed 10 October 2026. Self-hosting involves configuration, security, monitoring and updates; it does not mean operating without cost.
If you have identified a recurring task, you can tell me how long it takes and which exceptions it has before deciding whether to automate it.