When are Beast Modes calculated and when to place in an ETL or leave as Beast Mode?

Sheepadoodling

I have read several varying degrees of opinions on when a Beast Mode is actually calculated. I thought a Beast Mode, attached to a card or assigned to a dataset, is calculated at the objects rendering, regardless of if the Beast Mode is actually being displayed on said card. So if I have 20 total Beast Modes, some only assigned to a card, while others assigned to the card and dataset, and 2 of those 20 or actually being displayed on the card, all 20 are still evaluated when the card is rendered.

Then I read in the Beast Mode facts a rather cryptic explanation….

Do Beast Mode calculations run when not used in a card?

No, Beast Mode calculations only run when they are being used in a card or as a Filter on a card.

That is it for any reference in Domo's help site on Beast Modes. Has this been a feature change or was this always the way it was? The reason why I ask is I'm trying to determine when it makes sense to place a Beast Mode in an ETL as a formula, instead of defining a traditional Beast Mode. Back when using Power BIs, Beast Modes are simply Measures for DAX and are column based style calculations or windows of data. Where I would create formulas defined the dataset creation process for Power BI if this was more row based style calculations.

How does this play into Domo, when do you decide to move a Beast Mode into the formula object of the ETL to create what will look like a normal field to the end user in the dataset the ETL creates? When do you just leave it as a Beast Mode on the card or defined on the dataset for all available cards to consume?

GrantSmith

The rule I have is that if the beast mode doesn't need to be affected by filtering, I'll put it into the ETL instead of a beast mode as it's more efficient and allows the card to load faster.

In the past Domo interpreted all beast modes but some time ago converted it to only process the ones that are being used. I'll typically save beast modes to the dataset so that I have a single beast mode I need to change in the future instead of tracking down all of the duplicates across all of the cards.

GrantSmith

The rule I have is that if the beast mode doesn't need to be affected by filtering, I'll put it into the ETL instead of a beast mode as it's more efficient and allows the card to load faster.

In the past Domo interpreted all beast modes but some time ago converted it to only process the ones that are being used. I'll typically save beast modes to the dataset so that I have a single beast mode I need to change in the future instead of tracking down all of the duplicates across all of the cards.