Automating Your Second Process: What You Do Differently After the First Success

After the first success, you change three things for your second process. You wait until the first has run for two weeks without you. You keep its form but decide ownership, oversight and schedule anew, and you introduce processes one after another instead of all at once.

How do you know your first process is stable enough?

Your first process is stable enough when it runs for two weeks without you stepping in. The trigger fires on its own, the result lands in the right system, and errors are reported to you so you never have to hunt for them. At imprUvement, I check exactly these points before I sign off a pilot.

Here's an example. Appointment requests that come in by email enter themselves into your calendar, send the confirmation and remind both sides. If you checked three times during those two weeks to see whether it really worked, the process is still running with you as its silent supervisor. That supervision is working time. Start the second process now and you pay for it twice.

Stable means you stop checking because errors get reported to you.

What do you carry over from the first process, and what do you decide anew?

You carry over the form: how you wrote down, tested and signed off the first process. You decide three things anew, because the first time they settled themselves: ownership, oversight and schedule.

With the first process, you were the owner. You looked at every case, and the date held because everyone was curious to see it work. Your second process won't get that attention. It runs in the background, and nobody takes responsibility for what runs in the background.

If the second thing you automate is your payment reminders, ownership may no longer belong to you. It belongs to the person who checks incoming payments today. They need the error alerts, not you.

You keep the template. You fill in new names.

How many processes can your business automate at the same time?

Your business can run many processes at the same time. You should introduce them one after another. The bottleneck isn't the technology. It's your attention during the two weeks in which a new process proves that it holds.

Here's the math behind it. If you introduce three processes in the same week and one reports an error, you search for the cause in three places. If you introduce them two weeks apart, you search in one. After six weeks, each of the three has been signed off on its own.

You're probably thinking: that takes a while. For each single process, it does. For your business it's faster, because no single fault makes three processes suspect at once.

What happens when these three decisions stay open is something I describe in Why a Newly Introduced Automation Slips Back Into the Old Routine After a Few Weeks.

Your next step: today, write three things for your second process on a single sheet. Who owns it, who gets the error alerts, and by when it should have run for two weeks without you. Each of these is a decision you make before you build. You can practise setting decisions once instead of making them again every week in the four-day challenge on impruvement.com.