DE
“Do you know what is common between economists and pathologists?” This was the question our Behavioural Finance professor asked us in his farewell speech during our graduation ceremony many years ago. Without awaiting any answers, he said: “Economists can always explain everything. What happened, why did it happen and what were the exact causes and reasons? They will show you very sophisticated and convincing graphs, will slice and dice data, will provide you with a mount of evidence confirming their explanations … the same as a pathologist will do … but unfortunately, when the patient is already dead.”
The Evolution and Pitfalls of Data Analytics
In the last decades, we have learned to use and take advantage of advanced data analytics, ML, and AI in areas such as Compliance, Risk, and Performance Management. We are modeling the behavior of our business partners, counterparts, and clients based on their actions in the past trying to predict how they will interact and behave in the future. The use of probabilistic approaches to identify risks, threats, and opportunities has become the new normal in the GRC space.
However, if you look around and take a closer look at how this is done, you will realize that the main, and in many cases the only, source of data is the past. The past, which is mainly represented by incidents and losses, performance figures from past financial statements, executed financial transactions, questionnaires, and assessments, in rare cases combined with purely static data describing the relevant circumstances in detail.
Reinforced by our operational blindness (there is a nice and very expressive German word for this phenomenon called “Betriebsblindheit”) we would spend enormous efforts to seek, find, and advocate newer, more sophisticated quantification techniques, and algorithms … invented and used by NASA, Google & Co … and which promise to improve analytical performance such as accuracy and stability of forecasts and quantifications by a few marginal percentage points. In recent years, I have often been sucked in and absorbed by such kind of egocentric exercises and have witnessed these kinds of ego-driven discussions far too often.
A Wake-Up Call: The Client’s Feedback
I remember quite well the feedback we received from a customer after one of those exercises a few years ago. I seem to remember it was about modeling and predicting the risk and probability of customer attrition. Quite a disappointing and discouraging statement, that in my humble opinion, sums it up and explains a lot. The client said: “That wouldn’t be anything I wouldn’t see, explain and conclude based on common sense”.
Although we were driven, motivated, and spurred on by the innovative spirit and work we were doing, with the firm conviction that we represented an avant-garde in the Risk Management industry, at that moment I remembered the words of our professor … and really felt like a pathologist, or at least not like someone who is going to “save someone’s life” with his findings and insights.
Shifting Focus: The Forward-Looking Approach
This case has affirmed my previously suppressed suspicion that we have been focusing on the wrong place. Not because analyzing past incidents and losses, financial statements, transactions, or questionnaires and assessments are not important or not useful, on the contrary, they are a very important source of data … nor because it is not important to apply the latest and most powerful quantification techniques and algorithms, of course, it is, … but because our analysis has been almost entirely Backward Looking, one-dimensional, and focused exclusively on the past.
The questions popped up in my head: How can we change this? … How can we make our quantification, modeling, and analysis more Forward-Looking?
My team and I were convinced that in order to make our analysis and modeling (of whatever type) more Forward Looking we had to: 1) break out of the one-dimensionality and cover more dimensions of our counterparts, 2) start to capture, measure and analyze the “Present” of our counterpart’s actions and interactions instead of only the “Past”, and 3) apply a holistic, 360° concept, which we called the “Understand Your Counterpart – UYC approach.
So, besides all the historical data available, we started to analyze how our counterparts behave in real-time, what the media writes about them, what their digital, cyber, social, business integrity and environmental fingerprints look like, how they talk to us, what kind of words and expressions they use, how their voices sound, how they behave in all digital channels we give them access to, how they use their devices during these interactions, what they read and what they focus on when they visit our platforms, what worries them, what they are interested in, what and how they write to us when they chat or email with us, how they are connected to their environment, what their spillover risk is, etc.
By proactively analyzing all this data of the “Present”, of the “Now”, in a Forward-Looking manner, we are not only dealing with much broader, more diverse and insightful data dimensions, but we have also managed to significantly reduce the number of False Positives (compared to the traditional one-dimensional analysis) and generate a variety of immediate and highly reliable Early Warning signals for a wide range of Compliance, Risk, and Performance Management topics and issues.
The Future: An Exciting Journey Ahead
Where are we heading? … Where are the boundaries?
Well, I don’t know exactly, but I am sure of one thing: since the world we live in and human behavior are so diverse, dynamic, and multifaceted, this must and will be an exciting journey!
At the end of his speech, our professor said: “Try to be smart economists and not pathologists!”
