How to suppress negligible rows/columns when scaling to millions in NR?

Context – We have used a custom dimension (Future 1) and conditional formatting to scale data to millions in Narrative Reporting. Based on the Future 1 selection, data is displayed either in units or in millions.

Note – A member formula was not used earlier since the Future 1 breakup is also applied in some reports for specific line items.

Problem – Rows/columns with small numbers are not being suppressed, even though the values are not visible when scaled to millions, because the data technically exists.

Testing – The compare to zero: decimal precision functionality was tested, but it only works on source data and not on data formatted in NR.