Before you measure innivations, you must know what you are measuring! These days banks are moving in the direction of non interest income. One of them is banccassurance. So you must read ab
out it and then try to find out how such income is measured.
You may consider the "CAMELS" (Capital adequacy, Asset quality, Management, Earning and Liquidity and Sensitivity) ratios for analyzing financial performance of banking institutions.
Two answers : a global answer and an answer by business.
First, the global answer could be to see the dynamics of the following ratios : operating expenses/net banking product or operating margin/net banking product.
But today, a bank has different businesses. It would be better to measure the same ratios in the différent businesses of the bank and the weight of each business in the bank without forgotten the relationships among these businesses.
RORAC (see above), but also the other way around RAROC, or with both numerator and denominator adjusted for risk (RARORAC). The tricky part is: how to link these ratios to company value?
I find interesting the suggestions of Kijjambu and Gueugnon. Much like the bottom-up (synthetic) beta calculation in corporate finance, I too think a weighted index of stream-based profitability ratios would be more interesting. Moreover, once you develop such an index, comparison of performance versus traditional indicators could unravel new insight. Which ratios you want to use to build the index is up to you. Since the approach is organic itself, I'd say traditional ratios like ROA/ROE/RAROA/RAROE should be okay. You may also incorporate latest advances in Z-scores or its alternatives.
On a side note, Prof. Schueler's idea is intriguing, and tricky (as he himself pointed out). I'd like to experiment with it before recommending. If you have enough leeway to experiment (i.e. your advisor doesn't object), I'd say go ahead with RARORAC.