Strategy begins with IA
All successful analytics strategies must first focus on putting fundamental data foundations in place. Without the right preparatory work, valuable insights may remain inaccessible. This is where implementation analysis (IA) comes in.
IA can be defined as a specification of all work that must be done to successfully implement a new technology service. In the world of data analytics, IA enables banks to quickly identify gaps in their data and deploy effective – and rapid – remediation strategies. The ROI of the service is huge, providing valuable learnings that massively outweigh the small IA fee.
When delivered by an external agency with an objective point of view, IA can massively accelerate the process of turning a bank’s data into actionable intelligence, as it is able to cut across internal politics and provide a realistic assessment of what the institution needs to do to begin its data journey.
Once the necessary data health challenges have been fixed, the bank can then begin on its analytics journey. This begins with normalisation. Data sets must be harmonised and standardised into one consistent format and cover as wide a set of relevant transaction and market data as possible. Furthermore, each data entry should be as comprehensive as possible, with all relevant fields captured for every entry.
Again, the initial IA can provide a valuable roadmap for this stage of the journey – after all, achieving such a unified, cleansed and enriched data set is often far from straightforward. Within market-facing firms, trades are being executed across myriad electronic trading venues including bilateral liquidity streams and by traditional over-the-counter protocols (i.e. voice).
Within the FICC markets, each trading network adheres to its own messaging language for passing and recording trades and there will often be wide variation in the fields captured for a given trade. To add to the complexity, data which firms bring in from external sources will have been processed in a way which is unique to that data provider and cannot simply be added to this new unified data set.