Dont get me wrong im advocate of automation. But lets be reasonable here if you are outside FAANG (resources) you dont have to automate 100% of things . Quick example.
I worked on project that included among other things native mobile apps. It was automated lets say 95% and only thing which remained manual was going to some repository (imagine S3) downloading ios and android binary and then uploading them to transporter/google play. It took 2 minutes at worst. We did it every release so once per week
Automating this would take one day if you know what you are doing. Do the math :).
Definitely. And in addition to the upfront costs, I think there's also a long-term cost to maintaining these in-house automations that should be factored in. What works today might break tomorrow, and someone will need to fix it.
Background in politics, commodity trading, and converted to being a data engineer in 2017. I worked with Django, Flask, Plotly, and Vue.JS, but now Airflow and PySpark for ETL pipelines.
It's always a cost-benefit analysis. Anyone automating without considering that isn't in business :)
I just want to say that to me it's not necessarily always just a time cost. We had a task at my company where another company would manually upload a list to a secured server and someone (non-developer) would take that file and copy and paste the contents into a web app and hit submit. The amount of time it takes to do all of that is around 5 minutes...but the task is all week days for 3 months. We're talking 5 hours of time total over the 3 months but probably an 8 hour day to automate it. On a time-cost basis, it makes sense to use a lower paid employee to do the data entry work than a higher paid developer for more time, but "everyday" also means you have to contingency plan for illnesses and vacations during that 3 month stretch. Now you have a manager balancing resources for this 5 minute task and potentially pulling in one or two other people and training them on it. I voted to automate and monitor, but was rejected because they saw 5 vs 8 hours...
Dont get me wrong im advocate of automation. But lets be reasonable here if you are outside FAANG (resources) you dont have to automate 100% of things . Quick example.
I worked on project that included among other things native mobile apps. It was automated lets say 95% and only thing which remained manual was going to some repository (imagine S3) downloading ios and android binary and then uploading them to transporter/google play. It took 2 minutes at worst. We did it every release so once per week
Automating this would take one day if you know what you are doing. Do the math :).
TL;DR - Automation yes. Hyperautomation mostly no
Definitely. And in addition to the upfront costs, I think there's also a long-term cost to maintaining these in-house automations that should be factored in. What works today might break tomorrow, and someone will need to fix it.
It's always a cost-benefit analysis. Anyone automating without considering that isn't in business :)
I just want to say that to me it's not necessarily always just a time cost. We had a task at my company where another company would manually upload a list to a secured server and someone (non-developer) would take that file and copy and paste the contents into a web app and hit submit. The amount of time it takes to do all of that is around 5 minutes...but the task is all week days for 3 months. We're talking 5 hours of time total over the 3 months but probably an 8 hour day to automate it. On a time-cost basis, it makes sense to use a lower paid employee to do the data entry work than a higher paid developer for more time, but "everyday" also means you have to contingency plan for illnesses and vacations during that 3 month stretch. Now you have a manager balancing resources for this 5 minute task and potentially pulling in one or two other people and training them on it. I voted to automate and monitor, but was rejected because they saw 5 vs 8 hours...
Btw, now they should be called MAANG