What data is required for compliance?

Compliance requires both internal and external data. In the latter case, data is usually sourced from content providers, aggregators and so on, raising the question of whether dataset X is better than dataset Y. After a previous discussion of qualitative aspects, we are taking a closer look at quantitative aspects in this expert article. When making these kinds of judgments, we often fall victim to logical fallacies.

Let’s imagine that those responsible for compliance are primarily using dataset X, drawing on dataset Y for additional queries. Dataset X will usually be imported into an application for standard checks, while Y can be retrieved online if necessary. If the compliance representatives find what they’re looking for in X, they don’t see any reason to search in Y, and won't notice if what they found in X is in fact missing from Y. However, if they don't find anything in X, they will search in Y as well and may end up finding results. To conclude from this that Y is better because more results can be found there is a common logical fallacy. The Venn diagram below shows a first example in which X and Y are equally good and a second example where X is better, even if more results can sometimes be found in Y. The above conclusion is therefore logically incorrect. If the compliance representatives instead used dataset Y for the standard checks and X for online inquiries if nothing was found in Y, they would come to the opposite conclusion.

We therefore need to turn to psychology. It has been proven that our capacity for memory is not uniform. Compliance specialists are more likely to remember the first and last search result than all of the results in between. According to Murdock (1962), these tendencies are known as “primacy” and “recency” in psychology. It’s therefore not surprising that a logical process of inquiry comes to different conclusions than an individual’s subjective evaluation.