23: A Postmortem
We are getting to the pointy end of a SharePoint/Power Automate project here at work. The end goal will be for us to break down silos that our various departments work in and foster a collaborative process. I have been running the implementation with an outside contractor, who is building custom integrations on top of the automations I put together in Power Automate. In preparation for the handover, I have been going through my flow to tighten everything up. This project has been a valuable learning experience in a few ways:
Communication
I’ve come out with the understanding that I need to improve my communication, so that the devs are very clear on what we require. There’s been a few points where I have been the reason for things not turning out how we expected it, because I assumed our contractors knew what I wanted. Spoiler: they didn’t.
Information Flow
This project was a baptism of fire into the world of no-code automations. Working extensively in Power Automate, I’ve improved my ability to conceptualise how data will need to flow as part of the automation and how to structure things so that we get the outcome we need. This will be a valuable skill to apply to any of the side projects I build.
Error Handling
One of the first things the contractors said to me when they saw the initial Power Automate flow was that it lacked any sort of error handling and logging. As a result, I got acquainted with Power Automate “Scope” actions, and then logging info each time the flow runs, so that I can track down errors and debug. This was my first taste of how the pros do it. A big eye-opener for me in how to make a process much more robust.
Prioritisation
In a previous piece, I discussed how I’ve been forced to be more ruthless in how I prioritise. This project is the same I was working on when I wrote that piece.
Scoping
Our company needed to get this project done quickly, so we went into the work with only a rough scope. This was not ideal because we’ve ended up spending longer on this project than we would have liked, and some of this was down to adding things in as we went. A rough scope is good as it provides direction, but in this case we could have delivered this sooner if we were stricter and more explicit on what requirements were.
These are a few of the things I will be adding to my toolbox for use at work and in side projects.
This post was last edited 3 months, 2 weeks ago.