While we may not focus on it, or even know it, our lives are surrounded by key performance indicators (KPIs). Amazon, Uber, and eBay all carry ratings for the goods and services they provide and even the briefest interaction with a help desk is followed by a request for feedback. Do those consumer ratings really influence our behaviour? Well yes, in part, but no doubt you, like me, have ignored some of them, doubting their authenticity and accuracy.
Commercial and regulatory organisations such as Experian, Moody’s and central banks improved on consumer and member ratings by adopting a more rigorous approach that captures more granular data, combines information from more sources and applies non-trivial aggregation algorithms. We then see sophisticated KPIs in the form of credit scores, sovereign and corporate debt ratings, liquidity coverage ratios and so on.
These KPIs are certainly more objective and carry more weight than subjective consumer ratings for several reasons: the source data is sourced in a more sophisticated way eliminating bias, the data is more voluminous and therefore more precise and has a certifiable lineage. Lastly, the aggregation algorithms are often in the public domain allowing them to be scrutinized, challenged and improved.
However, objective ratings can still be rather ineffective when it comes to influencing behaviour. Who cares whether the rating is 980 versus 981 and does it matter much if the value fluctuates from day to day or adopts some sort of cyclic or systematic variation?
The problem is that these ratings do not relate to anything measurable in the real world that we can equate to.
Incentives and Penalties
So governments and regulatory bodies go further when it comes to forcing behaviour to change, in particular when thinking about corporate behaviour in regulated industries. What often results is a carrot and stick approach: financial incentives and penalties based on KPIs. These monetized KPIs manifest themselves in many ways; examples include bonus payments, performance penalties, fines, balance sheet provisions, risk charges, sales commissions, cost of sale and so on.
For example:
- The Basel Committee on Banking Supervision mandated that banks comply with regulations and those that do not face higher risk charges.
- Violating the U.S. Health Insurance Portability and Accountability Act, which includes privacy laws, can cost up to a quarter of a million dollars in fines.
- In the U.K. a train operating company can be penalised thousands of pounds when a service arrives more than 5 minutes late.
Although monetized KPIs have a direct impact on the financial performance of any enterprise - private, public, government or corporate etc, monetized KPIs can also facilitate the funding and success of projects, because the return on investment can be measured and justified. Your project could be cancelled without the right dataset and data tools to measure impact.
The Need for Detail and Nuance
The monetization process is usually complex in order to take account of the vagaries of human behaviour and the sometimes illogical evolution of processes and legal entities around the world. Company mergers and acquisitions add to the mix. The resulting processes are often non-linear, statistical or non-additive (for example a KPI must be within a target range 95% of the time). One consequence of this is that monetized KPIs often demonstrate erratic behaviour whereby a small change can result in a disproportionate outcome. Another consequence is that it can become difficult to attribute a change in KPI to a particular underlying cause when the data are reviewed retrospectively.
So monetized KPIs bring complexity to the enterprise:
- First, the KPIs need to be calculated accurately and monitored daily so that trends both positive and negative can be identified early.
- Second, the daily or weekly changes in the KPIs need to be analysed so that the underlying cause of changes can be mitigated.
- Finally, every transaction and decision in the enterprise should be tested against the KPI to demonstrate that the proposed action will produce a positive outcome.
For this reason, all of these tasks need to be performed in a system that can operate at the speed of thought, responding to user interaction in a split second, because the enterprise cannot slow down or afford to wait 24 hours for the result of a newly programmed batch report.
ActiveViam’s Atoti ticks the box on all three of the requirements needed to perform KPI monetization: flexible, accurate aggregation across multiple dimensions, retrospective explainers and predictive What-If analysis are the characteristics that are needed for an enterprise BI platform and ActiveViam’s Atoti platform is unique in being able to satisfy all of them making it a natural choice for this use case.
