Hi all, I read the thread below but I think my question is a little more specific.
In my model I have maybe 800 measures so far - not all of which will be exposed to the users. Clearly it’s in everyone’s interests that measures are named clearly and consistently. Hence I wanted to ask the community what you guys do for naming conventions. Let me give you an example. Take the measure “Gross Sales £”. I need to present this as:
Sales (Yesterday)
Sales (Yesterday vs SDLW)
Sales (Week To Date)
Sales (Week To Date vs WTD Previous Week)
Sales (Week To Date vs WTD Previous Year)
Sales (Previous Full Week)
Sales (Previous Full Week vs Prior Week)
Sales (Previous Full Week vs Same Week Last Year)
Additionally all the “vs” measures also need % versions and in some cases % - % as percentage point difference values.
So what I am looking for is a simple, clear and consistent way of naming the measures that I can build in from the beginning and get the users accustomed to. I would like to know what you actually call yours. Do you have an abbreviation convention, for example.
Hi @Jamie. I’d continue to use the short forms you already have, and replace others as appropriate, so, for example,
PY - previous year
PQ - previous quarter
PM - previous month
PW - previous (full) week
P2W - week before previous week
SDLW - same day last week
SWLY - same week last year
WTD - week to date
etc.
I’ve also found it quite useful to output the text of my measures using an external tool like DAX Studio or Tabular Editor, drop the text into a Word document, then work with that a bit to identify patterns that look good to you. You’re right that consistency is the goal, and I’d add (paraphrasing Matthew Roche), “make names as short as possible and as long as necessary”. As well, you can add a glossary page or info button tooltip to your report(s) that reads from a stand-alone Excel file listing all abbreviations and their “expanded” meanings, if that would help the end users.
Following on to @Greg’s excellent post, I wanted to share a great technique I first saw @datazoe use, which was to include a searchable table (with the addition of a slicer on the measure “Name” column) with all of her measure info at the end of her report.
You can play with her live report here:
Also, I would strongly recommend taking a look at calculation groups - if your measures follow repeating patterns (e.g., 20 measures for Sales, same 20 measures for Profits, etc), Calc groups can reduce your total # of measures needed by 75% or more. If you go to YouTube and search for "Enterprise DNA Calculation groups, it will take you to a whole series of great videos done by Enterprise DNA Experrt @MudassirAli.
