Data Thing Part 9: JUST DO IT
SIT REP
- Spent a lot of time doing Microsoft Courses to get my head round what Fabric is and how it should be set up
- There isn't really an official way of using it
- If we go ahead with it my team will need to determine the data standards, processing standards and architecture we want to use. It's difficult to pitch the system to senior mgmt without a clear description of how it will work for us, but that's a lot of work to figure out while its only a maybe
- We already have E5 Microsoft Licences so we won't need to purchase Power BI Pro licences or go in at F64 minimum (YIKES 🤑)
- Storage is cheap and the general consensus on compute is to start with F2, be mindful when scheduling jobs, and use the Capacity Metrics App to monitor usage
- Hopefully (for a small org like us) by the time you outgrow F2 you'll be getting the business value to justify the spend on F4 or F8
- Not sure how many times I can roll over the free trial, or if I like the danger of doing work in the system to test it knowing that it could be wiped out if we don't or can't renew
- As this would be a small addition to our Microsoft bill there's no procurement or new supplier process to follow
DECISION TIME
- I think we should commit to give it a go for a year and see how we get on. In 4 - 6 months I should be able to present real implementations to our directors, and develop a long-term plan
- I don't really want to write a full business case for £250 odd per month
- I book a meeting with our Director to discuss it. Due to annual leave and my impatience, I end up with a date that gives me 1.5 days to prepare a presentation that explains what Fabric is, why we should use it, how it fits with our goals and what I want to do next. On top of all the other work I have after a week off 😁
- I am fed up of faffing around trying to make the perfect case and have everything worked out. We aren't going to really know what we want to do until we can start doing it.
- Meeting concluded - decision made to go ahead with a year trial starting NOW.
This means we can crack on and really use it, then report back. Next up, figure out how we actually buy it (lol) and book a meeting with my team to scope out the architecture design and standards we want to use.
I'll do another post with some of the notes I used on data maturity and strategy 'n all that, in case anyone wants to reuse any of it.
In Conclusion
Find a way to make it happen! 💥
If an official procedure is getting in your way... try... not doing it. you might find its not as mandatory as you thought
And also
We like the idea of being Agile and testing and iterating and all that. But our procedures aren't designed to make it happen. I had to bypass our official processes and rely on half an hour of bodged together "LOOK AT THIS SHINY THING CAN WE HAVE IT PLEASE PLEASE PLEASE" to get approval to give this a try. I'll mention that when I report back in a few months time.
Other posts with tag "data thing":