Analyzing data requires a good deal of caution, because there’s always a risk of inferring two unrelated trends are linked. That’s known as confusing correlation with causation, and it’s a problem highlighted in hilarious fashion by author and Harvard Law School student, Tyler Vigen, in his “Spurious Correlations” blog.
The blog shows several graphs, each containing two correlating trends that are in no way related … for example, the divorce rate in Maine and per capita consumption of margarine. One does not cause the other, even if the graph shows a correlation of 99.26 percent. Yet, people often see causation where it doesn’t exist and make other mistakes when analyzing data.
When you look at your sales team data, there are some ways you can avoid making some of the most common erroneous conclusions.
People sometimes make the mistake of trying to extrapolate findings from a statistical set that’s too small. If you look at 10 customer accounts and find that five of them bought products on a Tuesday, that doesn’t mean 50 percent of all your product sales happen on Tuesdays.
A sample size of 10 will have a 31.6 percent margin of error. As the sample size increases, the margin of error decreases.
Sample bias is when an analyst eliminates relevant data in a way that skews the outcome. Usually, that happens unintentionally. For example, decades ago, pollsters would call people at home to ask which presidential candidate they preferred. Today, though, homes may not even have a landline – about 60 percent of adults under age 45 only use cell phones. So calling only landlines, or only cell phones, could have a big impact on the results of a survey.
If you’ve upgraded your sales software recently, make sure you import old data. Otherwise, you might inadvertently omit a lot of statistical information that could help you identify long-range trends.
Accounting for variables
A random sampling has to account for variables that influence behavior. One of those is geography. If your company sells products regionally, nationally and internationally, those markets need to be analyzed individually.
Some variables are easy to identify. Shopping habits around holidays and during a recession, if not accounted for, may cause you to draw incorrect conclusions about sales data. Your results will be most reliable if you can focus reports in a way that eliminates as many variables as possible.
The biggest way to avoid mistakes in sales analysis is having the right software. Sales force automation is a necessary tool for successful sales teams today because it provides the insight you need to make data-driven decisions.
For more information about maximizing your profits through data analysis, read our white paper, “Understanding Your SDR Needs: Identifying Your Input and Output Ratio,” by filling out the form below:
If you’re interested in learning more about how our sales force automation solution can help you, request a demo today.