Dear Bank IT Manager,
My most sincere apologies for being so quiet all this time. But I have been rather busy these last weeks… Did I mention in my latest post that P&L complexity should not be under-estimated? :-)
Anyway, let me try to elaborate a little more about the complexity of P&L tools. Here you have a second thought, more to come soon (hopefully):
Consider exception as your main rule.
A common fallacy is that the new strategic P&L platform will not need to deal with too many exceptions. Probably in parallel with the new strategic P&L platform, there is another initiative in the business to standardize processes, systems, calculations,
reporting, etc. Don’t get me wrong, I do believe that this strategy is serious and the management will put a lot of effort on it. But this reminds me of an old song that I have re-discovered thanks to Michael Bublé: “Tell me, quando, quando, quando”, i.e.
Tell me, when, when, when…
Especially when we are speaking about big banks, these sort of standardization programmes can be extremely challenging. One obvious case is the different regulatory rules in each country, which will impact in our platform functionality whether we like it
or not. It will also be difficult to standardize rules among different departments specializing in completely different investment products, either in nature (e.g. equities vs. debt) or in complexity (e.g. vanilla vs. exotic derivatives). There are other showstoppers
to be considered: change-reluctance in certain departments, complex impact analysis of any change, etc. We also need to be very conscious that certain changes, once agreed, might need a significant time to be applied for technical reasons.
So let the standardization programme team do their work and applaud gratefully for any small success. But in the meantime be selfish and consider exception your main rule. If you want to build an acceptable P&L platform it could be sufficient to design a
standard exception handling. But if you want to build a good P&L platform you need to make exception as a core feature of your functionality. In other words, do not consider exceptions as a collateral problem to deal with; instead consider them as an
intrinsic feature of the P&L world. There will be different rules across departments, products, locations, etc. and they will vary over time. So make your exception treatment really maintainable because you will probably need to revisit them rather often.
Service-oriented designs can be your allies here.
Drilling down a little bit I would suggest that you design your application to deal with:
- Multiple upstream systems: Different logic, formats, timing and even data granularity. Do not over-believe in upstream de-commission plans. It is incredible how difficult it is to de-commission any application.
- Different calculations: Different formulae for the same P&L concept depending on the product, location, department, etc.
- Multiple workflows: Some simple workflows in certain departments might need to co-exist with more complex workflows: adjustment types, approval iterations, data investigations at different levels, etc.
- Flexible reporting: Different levels of granularity, reporting hierarchies, aggregation rules, etc.
Hope it helps,
Miquel Febrer, Director, GFT Iberia.