Dear Bank IT Manager,

You might wonder why I have disappeared for a month and a half. Well, it has been a hectic period with a lot of projects commencing at the same time. When are banks going to move away from the projects that strictly follow the calendar year? Why do all engagements need to start on a 1st January and finish on a 31st of December? Well, it is not the intention of this post to complain about this.

So let me continue with the Six Ws method. In my latest post I wrote some thoughts related to the “Where”, stressing the importance of analysing where stakeholders are and processes happen. Despite the new Cloud era, location still matters e.g. to establish a proper operational model and methodology or to consider specific regulatory requirements. In any case, now we should speak about…

When:

The first “When” question should be at proposal time, i.e. when the P&L tool needs to provide each piece of functionality. It seems an obvious comment, but a few failed projects prove that is not that obvious. Surfing on the over-simplification wave, some people might tend to think that a P&L platform is a compact functionality and therefore it has to be delivered at once. We should never confuse end state with implementation plan: Which pieces of functionality are more urgent from the business perspective? Which functionalities are strictly required to cover a certain regulatory or audit requirement?

Our requirement analysis should pay  special attention to when users require the data. In complex organizations this question can become really difficult to answer. Our analysts might be tempted to make a generic statement: “Data has to be ready before controllers go their offices in the morning”. That should never be acceptable. Which users? Probably Asian users will need the data before Europeans, and those will need the data before their American colleagues. But even if current users are in a single location, are you sure that there is no plan to outsource this function to another continent? My experience is rather tough here: We had a tool that worked properly with users in London. One day the controlling function was moved to Far Asia. In a few months we had to re-engineer our batch to take 8 hours less. I promise you, that it was not straight-forward but we learned a lot.

Additionally we should make an effort to refine other dimensions of data: When will users need information about each product type, when will they need to report each P&L component (Trading, Funding, Dividends, etc), at which granularity, … For example; some people talk about book streaming as they want to have the book’s P&L as soon as it is ready. That seems sensible, but are we sure that the user’s workflow does not require aggregations before completing their workflow? So what is the real SLA?

In general it is a good investment to refine, refine and refine the “When” at an early stage. It will avoid generic complaints about the performance when the platform is in production. Therefore the team will be able to focus on the detailed SLA that are a real requirement.

Hope it helps,

Miquel Febrer, Director, GFT Iberia.