I have been told multiple times over my career, “Don’t send me a notification to approve something on a weekend.” Or do not send an authorization notification on a Sunday, because it is destined to trigger a chime in the next authorizer's email while they are having breakfast with their family. Then when should we when still trying to realize the velocity needed within our organizations?
While I understand the reasoning to not disrupt off-hours, it got me thinking. What is the optimum flow of authorizations when there are single or multiple-tier approvals? How do you achieve velocity while minimizing off-hour disruptions? Couple this with the importance of work/life balance, how do we optimize for velocity while not disrupting our needed time to disconnect?
When thinking about this subject, the Japanese word of heijunka comes to mind. Heijunka is a lean method to reduce the unevenness in production processes and minimize the chance of overburdening a process step. Villanova University describes it as
Heijunka (hi-JUNE-kuh) is a Japanese word for leveling. It is part of the lean methodology of process improvement that helps organizations match unpredictable customer demand patterns and eliminate manufacturing waste by leveling the type and quantity of production output over a fixed period.
Over the past few years, I have been part of many teams rebranding authorization processes for multi-tier approvals. These were written with lean intentions and driving to reduce the meticulous previous versions of “reply all” with Approved written in the email body. The replacements were written to increase the velocity of the process and get the originator the authorization as fast as possible. But at the demise of velocity, also came a perception that I receive all of the approvals on the weekends.
Heijunka (pronounced hey-june-kuh) is a Japanese word that means leveling. In Lean, it refers to the leveling of production, aimed at improving the flow of a process to better match customer demand, reduce waste, and decrease or quit batch processing.
My hypothesis was all over the place before looking at the data. Would originators start on Mondays with a heavy load resulting from the backlog they accumulated from the previous week? Somewhat of a “firefight” strategy on Mondays.
Or would I see the opposite, cleaning of their backlog on Fridays, so that they are setting the stage for success when Monday comes around?
Maybe the approvers would approve everything on a single day as if it was scheduled in their calendars to catch up amongst the already hectic week. Then the tiered approvers; when do they approve? Why do they feel they are getting everything on the weekends? Then what about the administrators, the last step is to transition the authorization to the action. I elected to evaluate one of the processes I was most familiar with, which spanned over two years with thousands of submissions.
Chart 1 - Approval process by step and day-of-the-week
The results showed that the highest day of the week starting the process is on Monday, and interestingly, has a negative through Friday. Along with the data, I saw very little evidence of processes beginning on a weekend. So processes start on Mondays tapering off by Friday with very little requested on the weekends. This trend corresponded with a positive slope, starting the week at almost zero for the approvals. Monday started at almost zero, increasing to the highest peak on Friday, with a slight tapering on Saturday along with practically no process steps completed on Sunday.
Digging into the data a little more, the majority of the Saturday approvals were of the second-tiered approvals. This meant to me that the majority of the first tiered approvals happen later in the week, followed by second tiers were destined to be processed on the weekends, preferably Saturday. The Saturday approvals most likely are the result of “I need to catch up.” This is where the velocity connection occurred to me. These weekend approvals then go to a final administrator's inbox to “process” what the authorization initiated. And what the data showed me was that approving the second tier on Saturday allowed the administrator to complete their process on Monday.
So, from start to finish, the average duration seemed to always be one week. From a timing perspective, approval on Saturday doesn’t do anything to the duration of the process but it does set up nicely to the next step. Additionally, the process slack during the week forces the second tier to conduct their action on Saturday or the duration would be significantly impacted.
What this little case study tells me is to avoid the burden of a process step and to maintain a good velocity of the processes, we need to focus on behaviors of the individual steps. Heijunka's logic would push towards watching the processes as a wave slowly going through multiple sequential steps.
A process cannot be understood by stopping it. Understanding must move with the flow of the process, must join it, and flow with it.
- Frank Herbert
In the example above, if the team wants the process to be conducted faster and remove the weekend burden, the team should strive to have the administrator conduct the final activity on a Friday. Inevitably, this is the “product” that is being produced. Simply putting out the edict of not passing approvals to the next step on weekends would significantly impact the effectiveness of the process in this example. Applying tools to sequence the individual steps with a single-minute exchange-dies (SMED) approach, the team can build the tact times required for each process step. In this example, the second tiered approval has to happen on a Thursday, first-tier on a Wednesday, and every origination starting on a Monday or Tuesday.
This analysis tells me that to accomplish the targeted feats, there is heightened importance on the quality of a process step’s input and the velocity of an individual step to the process. It tells me I need to challenge what steps are truly required and we have some of the foundational data for a process flow map (PFM). It also tells me that our results may encourage more mobile applications instead of processing behind a computer screen. It also tells me that to improve processes like this, it is evident that lean methodology applies to staff departments just as much as production units.
If the goal is to avoid the sense of a weekend burden yet achieve the velocity targeted, we have to look deeper into the process steps. It's not just one person’s action or one process step that creates the burden of performing work on off-hours. Instead, it is the result of the combined sequential actions and can be rectified by process improvements on all leading steps.
Measurement is the first step that leads to control and eventually to improvement. If you can’t measure something, you can’t understand it. If you can’t understand it, you can’t control it. If you can’t control it, you can’t improve it.
- H. James Harrington