Which statistics are you using to track online marketing KPIs?

One of the most common statistics used is the average, also known as the Mean. But did you know that the average could be giving you the wrong impression about your marketing campaigns’ results? Have you tried the Median?

Today, I discuss when to use and not use mean or median when analyzing your online marketing strategy.

** Mean, Median, and More for Online Marketing KPIs**

**Mean**

Also known as the arithmetic mean, or the average, the Mean is calculated by adding up all the values you have and dividing that total sum by the number of values. Let’s say 10 punters waged on your online sportsbook, or you’re calculating the average revenue per user (a significant KPI), or you sell sports accessories, and 10 people made orders on your store. And these punters waged the following amount of money – $20, $35, $50, $100, $150, $20, $70, $80, $55, and $85.

The Mean will include getting the sum of all the money, which gives you $665, and when you divide that by 10, you will get $66.5. When budgeting for your next marketing campaign, this value will help you know how much you need to spend to maintain a sustainable ROI.

The only issue with this approach is when there are outliers. These are values that are either too high or too low. In the values I just gave, let’s say one guy came and spent $1000. When you add that to the $665, you get $1665. And when you divide that by 11, you get $151.

This means that the average expenditure per person is $151, which isn’t true since very few customers spent more than $100. In such a case, the Mean would be giving you the wrong picture of your business. And that’s where Median saves the day.

**Median**

On the other hand, the Median is calculated by ordering the numbers in ascending order and picking the value that appears in the middle. In the list of values I just gave, we have a list with 10 numbers and a list with 11 numbers. Let’s start with the list with 11 numbers (odd number).

The formula to get the median = (n+1)/2 = 11+1 = 12/2 = 6

Our median will be the number in the 6th position

Let’s order the values

$20, $20, $35, $50, $55, **$70**, $80, $85, $100, $150, $1000

Median = $70

Now for the even values (10 numbers)

Formula = (n+1)/2 = 10+1 = 11 = 11/2 = 5.5

But we can’t have a position 5.5, so we’ll have to interpolate between the 5th and 6th position.

Let’s order the values

$20, $20, $35, $50, **$55, $70**, $80, $85, $100, $150

Let’s get the mean of these two values = 55+70 = 125/2 = 62.5

Median = 62.5

As you can see, adding the extra-large value ($1000) didn’t skew the data that much, and now we have a better perspective of the data we have. Francesco Mascadri and Christopher Penn, both marketing consultants, agree that the Median can help budget for your campaigns or analyze data in a better way compared to the Mean.

So, when analyzing some aspects of your online marketing campaign, and you have to calculate the Mean, don’t forget the Median.

**Mode**

Mode is another measure of central tendency that can help analyze your data. It’s simply the value that occurs most. In our values above, $20 occurs twice, so it is our Mode. But for a sports betting company with millions of customers, calculating the Mode will be more complex than that. Like the Median, the Mode is not affected significantly by the outliers. The only issue with the Mode is that it’s not always practical since it doesn’t represent the whole dataset.

**Standard Deviation**

This is another statistic that you can use alongside the Mean. Standard deviation measures the degree of variability from the mean/average. For instance, in the values we just used, the Mean was $66.5 before adding the $1000 value. As mentioned earlier, the Mean can misrepresent data when there are significant outliers.

But the standard deviation can paint a clear picture by saying that the average expenditure per user in a day is $70, with variability (standard deviation) of about $50. This means that some days people spend more, and other days they spend less. A high standard deviation necessitates an investigation to understand what’s happening.

For instance, if you run a sports news website, and on some pages, some people find your content engaging and spend hours there, while others only spend a few seconds and bounce, you need to investigate. Maybe your content targets only one demographic, but you still need to address more demographics with the content.

The same case applies when calculating the conversion rate of several marketing channels. You can calculate an average of the conversion rates. However, some may have a very high conversion rate, while others may have a very low conversion rate. Standard deviation helps you understand the variability between the various channels.

**Correlation Analysis**

This is the study of the relationship between two variables. Correlation is measured on a scale between +1 and -1; values close to +1 show that the variables have a positive correlation. Meaning that if you increase the value of one variable, the other variable will also increase. But if the values close to -1, it shows a negative correlation. And if you improve the value of one variable, the other variable is likely to reduce.

You can use correlation analysis to analyze the;

· Effectiveness of a particular feature or design on your website.

· Ad format or configuration,

· Content length (web or social media)

· Click-through rate with or without images in your campaigns, and as many variables as you can find related to your business.

Correlation analysis is often conducted using surveys, where marketers strive to find the variables with the highest positive correlation value. The two methods of conducting correlation analysis are Spearman’s Rank and Pearson’s Coefficient.

Remember, if the values are close to 0, then the variables have little to no relationship.

**Wrapping Up**

In this age of digital marketing, data is everything. KPIs are examples of different groups of data you can use to measure the performance of your business. But data would be of no use if there was no way to analyze it. That’s where statistics/parameters such as Mean, Median, correlation analysis, Mode, and standard deviation can help. They get rid of the guesswork out of marketing and improve your ROI in the long run. Mean is the most common and helps give a general picture of the data. But standard deviation, correlation, and median help paint a clearer picture.

**Partner With US**

Scoresandstats is a top sports news website that offers the latest news in various sports, including soccer, football, baseball, and basketball. They are also an advertising agency ranking for several keywords through SEO and with a large following on social media. As long as you’re in the sports and betting niche, you can rest assured that you’ll be advertising to the right audience.