Correlation Analysis and Your Marketing Efforts

Your vision is to know if there is a correlation in sales and marketing.

Ideally, you would have clarity into whether your marketing efforts are tied directly to your revenue. You could try correlation analysis.

You’ve been putting a lot more time and energy into marketing efforts. You may have also tried some new avenues, such as podcasts, trade shows, and conferences in an attempt to expand your customer base. Sales have been on the rise, but do you have to spend on these types of efforts to expect higher sales? The dilemma is, how do you tie these things together in a presentable way to consider this information in determining a budget and spend?

One solution is to look at a type of statistical analysis called correlation analysis.

The purpose of correlation analysis is to look at how strong the relationship is between two variables. It is important to note that this isn’t going to tell you that one variable causes another.

Positive Correlation: If one of the variables, let’s use sales, and the other variable, let’s use marketing, increase simultaneously with each other, you have a positive correlation. The slope of the line on your graph would be upward.

Negative Correlation: If sales go down as marketing goes up, this is a negative correlation. The slope of the line on your graph would be downward.

No Correlation: If there is no relation between the variables, you will have 0.

Taking the positive and negative concepts a step further, we can look at the Pearson correlation coefficient. This is how you put a number to the value between variables. The range is -1 to 1. Results that are negative indicate a negative correlation and values above 0 indicated positive. When your results are closer to 1, that means the linear relationship is stronger than it is if it were closer to zero. So for example, 0 to .3 is weak, closer to .5 is moderate, and .7 to 1 is strong.

Using a practical example of weather and vending machine sales at an outdoor mall, on a hot day, vending sales are up. This shows us that as temperature rises, vending revenue is up, and it’s likely going to be a strong correlation, of .7 to 1 range. What would the example look like if the correlation coefficient were closer to zero? Restaurant sales at the mall may be an example. People are probably just as likely to eat at the restaurant, regardless of the temperature.

So there must be a formula for correlation analysis, right?

Yes, and it’s a long one. Since the purpose of this article is about how you efficiently apply correlation analysis to your business, I am not going to put the details here, but you can read more in the reference section. If you want to look more into the formula, I recommend you first get up to speed with standard deviation. I highly recommend you use Excel or Google Sheets formulas for correlation. This is a link to a sample spreadsheet.

Step 1 in finding correlation:

List out or run a report of the two variables you are looking to compare. In the example, we are looking at a year of sales revenue compared to marketing spend by each month.

Step 2: Enter the correlation formula into any cell and choose the two columns you are comparing.

=CORREL(B2:B13,C2:C13) → Enter and note you see .71.

Step 3: Check for outliers.

Add a scatterplot graph, by highlighting columns B and C, and clicking on ‘Insert’ → ‘Chart’ → ‘Scatter Chart’. Then go the the chart editor on the right, click on ‘Series’ → ‘Trendline’ → and type needs to be ‘linear’. Scroll down a bit to check the box for ‘Show R^2’.

Notice that if you take the square root for R^2, you get the correlation value. This is a good way to visualize your data and an alternative for getting your result.

Look at your graph to see if any of the plots are extreme outliers. In our case, there are none. However, to test this, try removing the $111,000 spend and look at how your correlation value increases to .74. Now put it back, and move to the next step. If you had an extreme outlier in your sample, you should consider removing it and documenting so that you are looking at the most regular business spend.

Step 4: Review this result.

Our sample correlation coefficient is .71. As a quick spot check, it’s confirmed the number falls where we would expect, between -1 and 1. Since this result is closer to 1 than it is zero, there is a positive correlation between sales revenue and marketing spend. This doesn’t mean that higher marketing spend is going to cause an increase in sales. What we can say is that these two variables move together. So, the more spent on marketing, the higher sales, and higher sales mean more spent on marketing. If our result would’ve been 1, that would mean that there is a perfect correlation between sales and marketing spend and that 100% of the time, they both move in the same direction. If our results would’ve been negative, then as more was spent on marketing, sales would’ve been dipping down, and as sales go up, marketing spend would be dipping downward.

In addition to knowing that these variables move together, you could also assume now that if you budget $300,000 in sales next month, your marketing spend will have to adjust up accordingly as well.

Try saving the sample spreadsheet and adjusting the numbers to see how this changes.

Try running a correlation analysis the next time you are wondering if two variables in your business activities are related and comment how it went!

Written by Nicole Hullihen, August 8th, 2021

Learn more about how I can help.

References recommended to learn more about correlation analysis

https://www.djsresearch.co.uk/glossary/item/correlation-analysis-market-research

https://www.statisticshowto.com/probability-and-statistics/correlation-analysis/

https://www.investopedia.com/ask/answers/032515/what-does-it-mean-if-correlation-coefficient-positive-negative-or-zero.asp

https://www.investopedia.com/ask/answers/041015/what-does-negative-correlation-coefficient-mean.asp

https://www.investopedia.com/terms/p/positive-correlation.asp