Beta, and not r-squared? A beta of one would imply that Google's revenue (or profit?) increase exactly as much as the GDP. An r-squared of one would imply that Google's revenue changes in proportion to GDP changes.
But he probably knows this better than I do, so presumably I'm missing something.
More specifically, a stock that has a beta of 2 follows the market in an overall decline or growth, but does so by a factor of 2; meaning when the market has an overall decline of 3% a stock with a beta of 2 will fall 6%.
In a linear regression model, r-squared represents how correlated the model is with the data, while beta indicates the relationship between an outcome (e.g., revenue) to an input (e.g., GDP). Thus, a r-squared of one in a model where profit is the dependent variable and GDP is the single independent variable suggests that profit is 100% correlated with GDP. A beta of 1 indicates that profit moves in the same direction and to the same degree as GDP.
But he probably knows this better than I do, so presumably I'm missing something.