How to Correlation, Causation and CRM

Bloggerping / February 15, 2021
How to Correlation, Causation and CRM

If you read a lot, like me, you can notice almost everyday is a new study that refutes some earlier research. Some cause cancer – so it’s good for you. you know the drill. What is going on here? Do we not easily know what our research is saying? Can no one interpret the data correctly?

This would have no meaning for CRM, except that with the advancement of big data and analytics, the front office – that is, the relationship between vendors and customers – is similar to many other efforts that rely on data analysis. Here is my take on all this.

About correlation

Very often the research we get in the popular press and in business interactions represents the findings of correlation studies. Simply put, correlation refers to how strongly two things or events are related to each other, and it takes some sophistication to understand.

We can think of correlation as a probability, but we need to understand what it means. A coin toss has a 50/50 chance of having a head or tail, so the 50 percent probability is absolutely neutral.

If something was 40 percent likely to happen, it would be negatively correlated. In other words, nothing will be more likely. However, the 40 percent probability of something happening is not zero, which is why we still rain on days when it is likely to be less than 50 percent.

Therefore, a probability greater than 50 percent is what we are usually looking at, and the higher the number, the better the correlation. A 90 percent probability is interesting, but 60 or 70 percent – not so much, for reasons that are so far clear.

Nevertheless, 90 percent correlation is not a definite thing. Using the weather analogy, we sometimes see sunny days when there is a 90 percent chance of rain.

In business, we tend to use correlation a lot, but it frustrates many because correlation alone will not tell us about another important part of the story, the work-cause.

About the reason

The reason behind correlation is. It is this data, added to the correlation data, that will provide the necessary information for decision making. Therefore, for example, a sales person evaluating prospects may look for a high correlation between a prospect’s requirement profile and the seller’s solution.

This is a good start, but it is missing something very important. It does not say anything about the prospect’s motivation, which can only be met through more traditional means, such as making sales calls.

What? Is not correlation enough? Consider this: At the level of correlation, a prospect in need of a solution looks just as one who just buys something from his opponent.

The reason, in this case, is another term for the purchase signal. If you look at the buying signals and not just the correlation, the purchased customer will look very different in this one dimension, which is still visible.

In sales and marketing analysis, we are mostly focused on correlation, and this means that we are far from foolish in making our predictions. I’m not trying to go into anyone’s case, but the fact that we are so rooted in correlation just tells us where we are in the lifecycle of analytics applies to CRM. There is more work to be done.

Another way of looking at the situation is through the lens of qualitative versus quantitative data. So far, I have been focused on quantitative analysis – such as getting those 90 percent signals. When we are dealing with quantitative findings, very often we are looking at correlation data.

Consider a candy bar

Finding the reason requires more sophistication, but it is often the qualitative conclusions that tip the balance. Interestingly, you can develop quantitative conclusions on qualitative conclusions, but it takes a little more work. You need to ask different questions, and you may need to answer to get a quantitative result.

Work-cause detection begins with asking open-ended questions. In my book, Solution for the Customer, I use the example of creating a new candy bar. The quantitative approach may ask about preferences, such as whether you prefer coconut, milk chocolate or dark, like peanuts, almonds, pistachios, nougat – the possibilities are almost limitless.

At the end of your research, you may have a very detailed understanding of how much your target audience likes the various components of a candy bar, but you won’t be close to creating something that will sell.

The qualitative approach is less sexy in many minds, as it means that you won’t get enough information to do the work – but consider it. In designing a candy bar, it will benefit you greatly if you also ask open-ended questions about what people like about them, or their favorite memories in a candy bar, or how they fit into a person’s day. Huh.

Those questions are also almost limitless, and the answers will surprise you and possibly tell you a lot of needs in a crowded market.

Leave a Reply

Your email address will not be published. Required fields are marked *