Marketing Analysis Math – Don’t Freak Out

Make decisions based on data.

In its simplest form, business is math. If the unit economics work in a businessʼs favor, then the business will succeed, otherwise, the business will fail eventually. Even if youʼre math-phobic, there are a few essential metrics that you should know to better serve your business. I know, there might be others who are able to use analytic tools to do the calculations for you, but you should still understand the underlying meaning of the metrics provided. Hereʼs a quick summary of four equations that you should know: 

  1. Customer Acquisition Cost: How long does it take your business to breakeven once a new customer is acquired? To do this calculation, you first need to know your customer acquisition cost (CAC). This calculation should take into account the expenses related to all customer facing roles involved in winning a new customer. For example, if you have an in-house sales and marketing team, the salary of everyone on the sales and marketing team, plus the cost of software, advertising, and employee benefits should be factored into the equation. Once you collect all of these numbers you can calculate CAC: (Total Marketing Costs + Total Sales Costs) / Number of Customers Acquired. CAC is a simple, but powerful calculation. Marketers can calculate CAC per marketing channel to understand if particular channels acquire customers more cost effectively. Alternatively, senior managers can better understand how changing the composition or compensation of a team can alter CAC, and therefore the profitability of the business.
  2. Independent t-test: The independent t-test is perhaps a bit more complicated. It’s used to determine if two sample means are statistically significant, or if the difference in means is simply the result of random error. The equation provides a score, called the “t value.” Once you have calculated t, you must compare it to a “critical t value.” If t is larger than critical t, then you can conclude that the observed difference in means is statistically significant. Itʼs worth noting that you select the critical t value based on the level of confidence youʼd like to have and on the size of the sample, which is called degrees of freedom. A common confidence level is 95 percent, meaning that there will only be a 5 percent chance that the variation between means is due to random error if t is larger than the critical t. Independent t-tests have a number of useful real-world applications. For example, letʼs say you want to test two web pages to see which one converts better. In this case, the two page variants are your independent variables, and the conversion rates of the pages are your dependent variables (the means you wish to compare). You can use an independent t-test to determine if the difference in conversion rates is statistically meaningful. 
  3. Net Promoter Score: On the whole, how do your customers feel about your business? Thanks to the net promoter score (NPS) youʼll be able to answer that critical question. The net promoter score is derived from a one question survey presented to customers: “on a scale of 1-10, 10 being very likely, how likely are you to recommend this business to your friends or colleagues?” As detailed in a Harvard Business Review case study, Enterprise Rent-A-Car administered the NPS survey to their customers. Thanks to the insights generated, they found success against their competitors. Hereʼs how you can calculate an NPS score: Detractors (0 – 6 scores) – Promoters (9 – 10) / Total Number of Respondents x 100. Track your organizationʼs NPS score over time to track how customers perceive your organization. You can even take a page out of the Enterprise playbook and start compensating customer-facing representatives based on the NPS scores associated with the customers they interact with. 
  4. Pearson Product-Moment Correlation: To what degree are two variables correlated? You can answer this question by calculating r, also known as Pearson Product-Moment Correlation. An r score is a score from -1 to 1. The closer the score is to either -1 or 1, the more correlated the two variables are. For example, the relationship between IQ and GPA has an r score of above .5, meaning the two measures are more than moderately correlated. Whereas the relationship between watching violent television and committing a violent act has an r score of .2, meaning the correlation is weak. If r is a negative number, it simply implies that the correlation is negative, for example, income and number of years spent in prison is probably negatively correlated. The fewer years you spend in prison, the higher your income. Calculating r can be done with the help of Excel or similar spreadsheet tools, or it can be done by hand. In either case, calculating r can help to break through office conjecture by providing colleagues or managers with a statistical explanation of relevant relationships.


Now that you know these formulas, surely youʼll admit that math can be useful. The next time youʼre faced with an important business decision, donʼt just rely on your gut. Try to put some of these formulas into play to understand how various decisions could impact your bottom line.

For more thoughts about marketing metrics, read “Start Measuring Downstream Brand Metrics”.

Copyright ©John Trenary 2022

Leave a Reply

Blog at