One of the most common issues people face when utilizing data is falling into the trap of repeating the same information in each presentation.
This is not a negative as most managers want to (or need to) see a month over month comparison in deliverables. However, doing this also means that an analyst could be missing the opportunity to share some actionable recommendations with the relevant people. This also means that the analyst is not making the most out of their data.
Maximising the return of your data can be a challenging endeavour especially if the analyst does not have an understanding of the data on a granular level. IE: the more the analyst understands what they are presenting, the easier they will find data maximisation. This is because data maximisation is heavily dependent on the analyst and the three Cs: Correlations, Causation & Conclusions
Correlations
A Correlation is, in effect, a connection between two or more things. For example, an analyst at a flower company might identify that sales peak in February of every year. This correlates to the long-standing and accepted relationship between flowers and Valentine's day. Therefore, it can be concluded that sales spiked in February due to Valentine's day.
A more robust example would be one that was popularized in the book Freakanomics. The author was able to show that the decision of the United States Supreme Court in the case of Roe v Wade in 1973 (where abortions were legalized), lead to a 15% reduction in crime over the succeeding 3 decades. The correlation was that legalizing abortion meant that the most vulnerable members of society could safely and legally have abortions, reducing the number of at-risk youth in the immediate 15 years. Therefore, with fewer people that fit the mould of people who would normally perpetrate crimes, the crime rate, understandably fell.
The same concept/thought process can be applied to an office setting. For example, an analyst could identify that when average spending in the entertainment sector decreases by more than RM40, the education sector sees a rise in average spending of RM40 or more too. If the analyst can show this is not a fluke or a one-off, it can be concluded that a genuine correlation between the two sectors exists.
Causation
Causation is the relationship between two temporally simultaneous events when the first event (the cause) brings about the other (the effect). Causation is the part of working with data that allows us to separate the fluke occurrences from genuine correlations. For example, the analyst from the example above could state that the rise in spending in the education sector causes the dip in spending in the entertainment sector. However, the analyst would need to prove this causation. A simple way to do this would be to identify the people who spend money in both sectors and to determine what percentage of those people fit with the pattern. If the analyst can show that a majority of people fit this pattern, it can be claimed that causation has been established.
It is important to note that causation is not easy to prove because lines between cause and effect are not always straightforward. Therefore, an analyst would need to dig deep and be willing to abandon a correlation quickly once it is clear that a pattern cannot be established. As guidance, an analyst should be able to prove a correlation after examining 5 instances and returning 3 positive matches.
Conclusions
Conclusions are the bookend to the entire pattern and where the analyst sees the return on all of the work that was put in to determine the correlation and prove the causation. Conclusions require the analyst to look at confirmed patterns and apply a mixture of logic and creative thinking to come up with an explanation of the correlation. For example, since it has been proven that people either spend in the education or entertainment sectors, we can conclude that most people will prioritize the education-oriented needs of their kids over their own desires for entertainment.
It is at this stage that an analyst can provide commentary on the links that were proved and deliver actionable recommendations. Therefore, a bank might want to create a promotion where a cinema offers free admission to a child if accompanied by two adults and the bank’s credit card is used for the purchase of the movie tickets. This means that the analyst is no longer limited to showing the month on month change in spending by sector and can offer insight that will enable their business to grow.
This is a snippet of one of the courses offered by Dicorm Consulting. If you are interested to learn more about our courses or if you would like to discuss how we can help you make the most of your data, contact us for a free, no-obligation consultation.
Comments