[ballenf]

Agree, but with a few caveats:

- The cost needs to be fairly easy to measure

- Some degree of competition in the industry exists (otherwise the capital could be used to increase revenue or acquire small potential rivals before they're too large to afford, among other higher return places to spend the capital)

- The industry is mature and sources of larger savings already exhausted

- Prospect for some future much bigger savings small (or at least not in conflict with the current effort)

- Prospect for the industry as a whole very good (in a shrinking market, 10% savings becomes lower priority than selling off assets quickly enough at high enough cost without subsidizing a competitor or startup)

- Industry consolidated enough (or capital costs of the improvement low enough) to finance the upfront cost