In this particular context, a process is a series of actions, events or tasks carried out in order to achieve a certain goal. In this sense, your morning routine may be seen as a process. You wake up, yawn a little. You get out of bed before stumbling sleepily to the bathroom etc.
We understand intuitively that these events are linked and that they are in some sense path-dependent (you can’t brush your teeth without getting out of bed). We also understand that many of the sub tasks are the same, but each morning is also unique to varying degrees. You get out of bed on 99.9% of mornings, but there might be one day when you call in sick and barely move. Some mornings you might sleep through an alarm, or experiment with getting up at 6:45 and meditating a little instead of the usual 7:30 rush.
Let’s say that one day, you wake up with an intense desire to more deeply understand your morning routine and in doing so, identify opportunities for improvement. Don’t ask me why – but I did this. I used Google Forms to make a little questionnaire that allowed me to record my activities and the exact timestamps, and one week later, I had my first results.
After a week of recording my morning routine (my morning “process”), I had a pile of data. Unfortunately: In stepping up from a linear timeline of events to a high-level, data-enabled overview, I lost the ability to reason intuitively. No one wants to spend hours poring over spreadsheets of data to find the insights that would streamline their morning, so what do I do!? I used UiPath’s ProcessGold.
Process mining and automation packages like UiPath take the data you collect and use specialized algorithms and visualization dashboards – tools that have evolved to deal with incredibly complex processes – to number crunch and visualize your data to make it intuitively understandable.
One way such tools visualize processes is using a process graph. After plugging in my data – even with my measly weeks-worth of mornings – I had quite the graph! Luckily, ProcessGold’s process explorer makes filtering the one-off links and the rare activities a breeze. In no time I had a high-level overview.
The high-level overview showed me a kind of “ideal” process, which was reassuring. I wake up, get out of bed (at varying speeds – we will get back to this), go to the bathroom and after doing what needs doing, I end up feeding the cat, getting dressed and leaving the house. I rarely have breakfast in the mornings, I usually fast until lunch. From a high-level perspective, my “average” morning looked structurally sound.
I knew from experience, however, that something was not quite right. I navigate to the timings tab and find my vice staring back at me.
The largest chunk of time is under “leaving the house”, understandable considering the last thing I do before I leave is usually working or tidying the house: activities that can stretch for a few hours. More worryingly, the total throughput times of the activities “Get out of bed”, “Yawn”, and “Wake up” were more than an hour! The data told me what I had feared the most: I take eons to get myself out of bed.
The averages were kinder: 12 mins to get out of bed. But it’s the one-offs that cause the most problems. In the analysis tab I found the culprit: It had taken me an hour and 20 to wake up on Friday, and on Wednesday I meekly shuffled between “Yawn”, “Big yawn” and “Get out of bed” for more than an hour.
But it wasn’t as bad as it first seemed. Upon closer inspection, the hour+ sleep in on Friday was due to waking up far too early and snoozing until the alarm. But even with this taken into account, the biggest bottleneck is still getting out of bed.
From the visualizations in the analysis tab, it seemed that Saturday was the slowest wake up on record: with a stunning 43 minutes of sleepy laziness before shifting (I may have been watching Netflix in bed, a little hungover after a brilliant ProcessGold party the night before!). Counterintuitively, the fastest morning was Sunday!
I wouldn’t have expected it, but the day I didn’t set an alarm, I got out of bed like lightning: The reason? My partner had made me coffee, and I couldn’t bear the thought of cold coffee. Asking my partner to jump out of bed every morning to make me coffee is obviously not workable, the solution to my lie-ins may lie here. The subject is process mining and automation: Can I automate my morning cuppa?
Using the high-level strategic oversight afforded to me through ProcessGold’s process explorer, I had the idea of automating a carrot on a stick to help me get out of bed in the mornings. Now this isn’t a digital problem, so UiPath’s software robots won’t help me. But what is a good analogy for a software robot in this context? An automated coffee machine.
It’s a simple story, sure. But this is the magical power of process mining and automation in tandem. Using process mining to get a grip on reality, and using those insights to automate key tasks (where you get the most bang for your buck).
So I went ahead and got myself a socket timer, connected it to my coffee machine and synced it up with my alarm. Not only do I have a meditative evening routine of setting up my morning cup, but I have a steaming cup of joy waiting for me when I roll out of bed. I wonder how I can determine if this automation experiment is a success? Maybe I should collect some data and visualize my processes…