Is it possible you're overthinking this? As an alternative, you might
simply represent each period's growth as a percent of a baseline
figure, maybe last period or period 1. Annual growth is usually
(results as of 31 December this year) divided by (results as of 31
December last year).

Just a guess, but if you went into the boss and asked if they wanted r-
squared represented in a linear regression format, they'd give you a
"bunny in the headlights" look. (For my boss, simple is better.)