Why data definition is important in Finance
Companies usually have two main sources of data to understand their performance:
- Financial Information (FI) which includes elements such as invoices, credit notes and accruals which are processed and posted into a general ledger. This leads to a statement of Profit & Loss and Balance Sheet for internal and external reporting.
- Management Information (MI) which is based on data extracted from trading systems at a more granular level than the FI can be. It is then transformed and loaded into a database for use.
There is a trend for MI and FI to converge as often variances between the two can be a bone of contention in a company; data teams and finance functions continuously question each other’s numbers, with finance questioning the MI’s validity and data questioning the FI’s completeness and definition.
This convergence can be achieved through building a new version of the FI from the core building blocks of the MI.
Clearly, this will either eliminate variance between the two or make any variance fully explainable, as the differing transformations on the same data can be compared. However, while this sounds good it could lead to grave issues regarding integrity of reporting:
- MI & FI are supposed to be independent sources of information that can be used to validate each other.
- FI is useful for validating MI scaling, timing and translation.
- MI is useful for checking the completeness of the FI.
These last two points are potential benefits that require the clear definition of both MI and FI elements:
- It is typical to have a clear definition of where the data comes from in the MI, as the ETL (Extract, Transform, Load) processes, which take data from the trading systems and transforms them for use in databases, contain this definition. However, this expertise can be lost in organisations or poorly disseminated leading to confusion about what the MI contains or doesn’t.
- The corresponding data definition in the FI has often been seen as more difficult to obtain; Financial Control teams with intense workloads are more concerned with categorising income and expense correctly, as opposed to making sure each element is understood and the definition recorded. Without this definition in the FI, the benefits of mutual validation between MI and FI can’t be achieved and the integrity of both is in question.
If the FI data definition is completed and maintained, the MI can be shaped to fit individual elements (invoices/credit notes etc) of the FI, according to this definition. This process of shaping the MI to fit FI elements will highlight any scale, time and translation issues in the MI.
Once this shaping is completed so that each element of the FI has a corresponding part of the MI, gaps in the FI can be discovered - there may be additional MI that has not been used to match an FI element. This additional data is perfectly suited to help make accruals to cover gaps in the FI. It may also be the case that the FI has double counts or overlap of certain income or expenditure due to complexity and the MI will be able to show this. Cash may be debited without a corresponding element of FI to explain what the company has paid for and to whom; the corresponding “loose change” in the MI should help explain this.
Clearly, there are big benefits to having well defined independent MI and FI systems. Attempting to force the two to match, without considering the data definition, risks compromising the integrity of both systems. The FI ceases to explain the cash position and liquidity of the company because it’s tied to the MI, and therefore financial reporting cannot be done in good faith. As the MI attempts to match this untrustworthy FI, it too will not be trusted for decision-making, rendering it useless.