When your own automation pays off instead of yet another subscription

A monthly subscription looks harmless. A thousand crowns disappears into a budget. But subscriptions pile up, and some of them you pay for only because of one feature that does one routine job for you.

Then the question comes up whether it would be better to automate that one job and cancel the subscription. Sometimes yes, sometimes definitely not. Here is how to work it out before you pay for anything.

The basic calculation

Start simple. Divide the one-off price of the automation by what it saves you each month.

Payback in months = one-off price ÷ monthly saving

An illustrative example: an automation for 15 000 Kč replaces a subscription of 1 000 Kč a month. Payback is 15 months. From month sixteen you are ahead, as long as nothing changes.

That "as long as nothing changes" is the whole difficulty. The basic calculation is a good start, not a decision.

What belongs in the calculation even though it is easy to forget

Maintenance. Paying once does not mean nobody ever has to look after the automation again. It connects to other systems, and those change. When the export format or the interface of your accounting software changes, the automation needs adjusting. Add a reserve for adjustments to the price.

Changes in how you work. Every change in the way you work can mean another adjustment. If your process changes several times a year, you will pay for it several times.

Growing numbers of users. Many subscriptions are charged per user. Today you pay for five people, in two years perhaps for twelve. Count on the number you expect, not today's. Your own automation usually has no such fee, although it too can run into the limits of the services it connects to.

The risk that the tool changes. The subscription provider can raise the price, move your feature into a more expensive plan or shut the service down. But it works the other way too: the tool can get better and you get that for free. Do not count only the worse outcome.

Your people's time. This is forgotten on both sides. An automation saves retyping, checking and fixing mistakes. But it also takes time to explain the brief, to try it out and to take it over. If you are not cancelling the subscription and only saving work, count the hours saved per month and multiply them by what an hour of that person's work really costs you.

What happens when it stops working. A subscription service can fail, and so can your own automation. The difference is who notices and who fixes it. With a subscription it is the service provider. With your own automation you have to agree on it in advance.

The same example, more honestly

Here is how the payback changes when you add a reserve for adjustments and a rise in the subscription price. All the figures are illustrative. We chose a reserve of 6 000 Kč and a price rise to 1 500 Kč just for the demonstration; put in your own.

So the same idea can have a payback of 14 or of 21 months. That is why you should work with a range, not a single number. If the automation pays off even in the worse case, the decision is easy.

Then compare the result with one question: how much longer will you keep doing that job the same way? A payback of 21 months and a process that may end within a year does not pay off, however precise the calculation.

When it does not pay off

Your own automation is not the right answer every time. Stay with the subscription if any of these points applies.

• The subscription is cheap. At 200 Kč a month, an automation for 15 000 Kč would take more than six years to pay back (illustrative figures again). That makes no sense.
• The tool is alive and you really use the new features. You would be replacing something that keeps improving without any effort from you. The automation stays as you ordered it.
• The process keeps changing. You would be automating something you will do differently in six months. Settle the process first.
• The work is rare and short. A quarterly report that takes ten minutes is not worth it.
• The subscription does more than it seems. It also gives you backups, access rights, updates and support. Without it, someone else has to look after those.

When it does make sense

• You pay for a big tool because of one feature and do not use the rest.
• The process is settled and repeats often.
• The fee grows with the number of people and the number of people is growing.
• The payback looks acceptable even with a reserve and is shorter than the time you will keep doing the work the same way.

Four questions before you decide

1. How much do you pay each month today, and how much of people's time does the job take?
2. How much would the automation cost, including a reserve for adjustments?
3. How long will your process stay the same?
4. What happens if the automation does not run for a day, and who takes care of it?

If you can answer all four and the calculation still works out, it makes sense to go further. If not, you have already saved money by working it out.

On the Automation page you can recalculate the payback in a simple calculator, and you will also find the cases where we ourselves will talk you out of automating.