Watch your step! Sales managers and managing directors in B2B confuse correlation and causality.
Data-based decisions in sales are not always ad-hoc better than intuition. The reason for this is the frequent confusion between the terms causality and correlation.
How nice it would be if managing directors or sales executives regularly knew why something happened. Why individual customers churn; why one product does not sell well or sells more than others; why in the end a promising sales lead does not become a customer, regardless of how good our salespeople are.
Questions and logical enquires you would like to get answered quickly and easily.
Humans are the result of causality
A person who thinks and acts rationally feels comfortable when he knows the reasons for something. Often causes are searched for in a convulsive way, only to get an explanation for a particular fact. The mind, trained to seek understanding, is out to find a cause-effect variable for everything, to establish causality.
Example price policy: If customers ask less or no more for a certain product, then this is of course exclusively due to a too high price. We wish B2B sales would be that simple.
It is not. Life is not as simple as we would often like it to be. It doesn’t always follow an understandable pattern neither can we so quickly establish cause-and-effect principles. Life is complicated and sometimes entirely random.
Something like this is unsettling, especially in harsh business environments managed by KPI and sales controllers. That’s why sales directors and managers are even more looking for manageable explanations.
Because those who know the reason for a specific customer behaviour can (re-)align their sales strategy accordingly. He can, for example, lower the price of a product, do better than the competition or produce it cheaper to gain a cost advantage.
With a manageable number of products that a company offers, this may still work. But the increasing saturation of the markets means that customers do not have the choice between a few, but between more and more suppliers. To differentiate themselves from the competition, these providers are therefore differentiating their range of services more and more frequently — the result: growing complexity.
As complexity increases, the number of possible correlations increases.
A forecast from the Federal Statistical Office, based on the volume of digital data generated annually worldwide, confirms an increasing complexity in all areas of business. While companies and consumers produced a mere 33 zettabytes (one zettabyte is 1021 bytes, i.e. sextile ion bytes, i.e. a 1 with 21 zeros) of data in 2018, the forecast for 2025 is 175 (!) zettabytes.
The megatrends digitalization and industry 4.0 mean that more and more data will be available to companies. It, therefore, makes sense to make business decisions based on data. After all, executives cannot ignore the facts hidden in the data. And data does not lie. Or maybe it does?
The answer is: it depends. It depends on how you read data – and what decision-makers and managers ultimately derive from it. The term business analytics shows, in principle, why managers are gathering data. Analytics covers methods from descriptive to predictive data analytics.
Analytics can be backwards-looking, i.e. data analyses relating to the past, but also forward-looking, regarding forecasts and predictions.
If we refine these two categories further, we can talk of Descriptive Analytics for business analysis and pattern recognition, Predictive Analytics for data-based forecasting and Prescriptive Analytics for the predictive, action-oriented recommendations.
In short, there are many areas where data analysis can help your business. There are, however, quite a few decision-makers and managers in sales that believe the amount of data is the only pre-requisite for better decisions.
The problem here is that data-based decisions are not always automatically the better decisions! The reason for this is a confusion of the terms causality and correlation.
Correlation and causation definition – what do they mean?
The Dictionary of Statistics says about the concept of correlation:
“A correlation measures the strength of a statistical relationship between two variables. In a positive correlation, the more variable A, the more variable B, or vice versa. A negative correlation means: the more variable A, the less variable B or vice versa. Correlations are always undirected, i.e. they contain no information about which variable causes another one – both variables are equal.”
Concerning the concept of causality, it says:
“If there is a relationship between cause and effect between two characteristics, we speak of causality. Correlations can give an indication of causal relationships. A person who smokes a lot (characteristic X) has a higher risk of developing lung cancer (characteristic Y). Important: A correlation alone is not proof of a causal relationship.”
Quite simple, right?
Why correlations between variables are important
However, many sales executives and executives still often confuse correlation and causality or negate that the comparison of data is a correlation rather than causality.
We remember: With the thought of a cause-effect relationship, i.e. with a supposedly rational explanation, many people fall asleep better than with the idea that it is merely an indication of a possible causality. The uncertainty that something could be connected without one knowing this cannot and will not be afforded in competition.
But it is precisely correlations between data and defined KPIs that give sales essential indications as to where it might make sense to adapt or change the sales strategy.
For example, predictive analytics can show which customers could buy more of a product in the future. Or recommendations for action can be derived with which sales could achieve better results overall.
Recommendations for action can be derived with which sales could achieve better results overall.
Correlations can, therefore, help to prioritize in which strategies and sales actions the sales team should work, to achieve specific results.
CALCULATE NOW THE ROI OF QYMATIX PREDICTIVE SALES SOFTWARE
Causality and Correlation Using KPIs in Sales as Examples
Key Performance Indicators (KPI) play an essential role for successful sales controlling and for corporate management.
Let’s assume that the KPI turnover for product A is declining compared to that of product B. Product A is more expensive than product B. Now it can be concluded that the higher price justifies the decline in sales, i.e. is the reason for it. However, this conclusion does not necessarily have to be true, because both variables – higher price and sales development – initially mean nothing but a negative correlation: the more variable A (price), the less variable B (sales).
But managers in sales want quick solutions and are therefore dependent on equally quick explanations. That is too short, however. The falling turnover for product A can perhaps be explained by the higher price compared to product B. The reason for this is that the cost of product A is higher than the price of product B. It can be – but does not have to.
The price is just one variable among many others. Another can be, for example, the combination of the individual ingredients that the product contains. Most customers may not like this combination. Or: The product name of A “pulls” less at the customers compared to the product name B, which again would be a matter of marketing.
Salespeople should still take the calls.
Data mining and analytics is helpful to support forecasts or deriving recommendations for action. By comparing different data or KPIs from sales, specific correlations can be visualized. They reflect the actual state of a relationship between two variables (KPIs).
What conclusions the sales department draws from this and what decisions it makes at the end based on the database is still up to it. An analysis and evaluation of correlating data can, therefore, only ever support decisions.