Questions and answers
Is automation worth it for a small business?
Multiply three numbers: how many times a month someone runs the process, how many minutes one run takes, and what an hour of that person costs. If the payback comes out under six months it is usually worth it. Over twelve, I advise against it.
A five-minute calculation
Take one repetitive process. Not the company as a whole, one specific task somebody performs over and over.
Work out three numbers: how often it happens per month, how many minutes one run takes, and the fully loaded hourly cost of that person. Multiply them. That is what the process eats every month.
An example: copying orders from email into a spreadsheet, 60 times a month, 4 minutes each, a person costing 50 PLN an hour. That is 240 minutes, four hours, 200 PLN a month. A 1,499 PLN implementation pays back in seven and a half months. That is a borderline case and I would tell you so directly.
What the calculation misses
Mistakes. Retyping data by hand produces errors, and the cost of one wrong order can exceed a month of saved time. It does not fit in a spreadsheet, but it is often the real reason an implementation pays for itself.
Response time. A manual process waits until somebody comes to work. An automated one answers a customer at ten on a Saturday night. With sales enquiries that is sometimes the difference between a job and no job.
Resilience. A process only one person knows disappears when they go on holiday.
When I say do not do it
When the process takes two hours a month. It will never pay back, and it adds one more thing to maintain.
When the process changes every quarter. Automation assumes repetition. If the rules keep moving, the implementation expires faster than it pays off.
When the real problem is organisational rather than technical. Automating a chaotic process gives you faster chaos, not order.
What it costs
Implementation from 1,499 PLN, support from 399 PLN a month. Support is optional but honest: the tools an automation talks to keep changing their interfaces, and somebody has to watch that.
How to work out the payback on automation
- 1Pick one process. Not the company as a whole, one specific task somebody performs over and over.
- 2Count the frequency. How many times a month somebody runs this process.
- 3Measure the time. How many minutes one run takes, including switching between tools.
- 4Work out the hourly cost. The fully loaded cost of an hour of that person, not the headline rate.
- 5Multiply the three numbers. The result is what this process eats every month.
- 6Divide the implementation cost by that result. You get the payback in months. Under six is usually worth it, over twelve I advise against.
One thing to remember
Payback under six months is usually a good decision. Over twelve, the technology will change faster than the process can pay for itself.
Let us check it against your case.
Send me those three numbers from your own process. I will work out the payback and tell you if it comes out badly.