As a small business, you are probably leveraging your social media to market your services. Social Media Marketing is an excellent tool for promotion. Especially when your budget is close to $0 dollars. I know because I am in the same boat as you. Today I want to help you understand social media data and look at its variability.
What is social media data?
When I am talking about social media data, I mean things like the number of likes or replies on a message you sent on Twitter or another app.You may have a way of doing this through a tool but I always collect this info by hand.
So typically, I will use a spreadsheet to track the text of each message, the number of likes, the number of replies, the date, the length of the message, and if there was any image or video. All these give me important clues to what worked or didn’t work during the past month. Yep, I do this as my end of the month ritual.
The thing is, social media data has a lot of variability to it. We need to understand that and statistics gives us a way of dealing with it. We are going to take a look at the median, outliers, box and whisker plots, and the standard deviation.
I previously discussed the median in my marketing insights article. This is one way out of three for describing an average. It has advantages over the mean and the mode. The median won’t just tell us the average it will also tell us about outliers in our social media data! That way we can look at a particular threshold and pick out our highest performing and lowest performing messages.
This is how you calculate the median.
Arrange all your numbers from lowes to highest. Find the number in the middle of your data set. That is the median.
If you have an even number of digits, add the middle numbers and divide by two.
Take this random set of numbers.
In order we have: 2, 4, 5, 7, 12, 14, 23, 25, 26, 28, 30, 36, 37, 41, 48, 48, 51, 52, 64, 65, 88, 93, 94, 97, 99.
As you can see the digits can repeat and that is okay.
The median of this set is 37.
Calculating the Outliers
Now that we know our median we can calculate our outliers.
The first step is to calculate the lower quartile. This is the halfway point of our digits below the median. There are 12 of those so we average the middle numbers.
14 + 23 = 37
The second step is to calculate the upper quartile. This time we find the halfway point of our digits above the median. Again this is an even number so we average the middle numbers.
64 + 65 = 129
The third step is to find the interquartile range. We subtract the lower quartile from the upper quartile.
64.5 – 18.5 = 46
Our last step is to calculate the lower and upper boundaries.
Lower Boundary: Lower Quartile – (Interquartile Range * 1.5)
18.5 – (46 * 1.5) = -50.5
Upper Boundary: Upper Quartile + (Interquartile Range * 1.5)
64.5 + (46 * 1.5) = 133.5
We have no outliers because our data set is within these boundaries.
Here is an example of a data set that did have outliers in it.
The outliers of the data set are in green.
Any spreadsheet program including Google Sheets will have the functions to compute the median, quartiles, interquartile range, and boundaries for you. I still think it is good to know how to do it by hand. If your sample is small enough you can try it for yourself!
Box and Whisker Plot
The best way to visualize the variability of your sample is with a box and whisker plot. They are very cute.
As you can see, this is very informative at first glance. We can see there is more variability above that middle line which represents the median. The left end of the rectangle is the value of the lower quartile. The value of the right end of the rectangle is the upper quartile. We can also see from the length of the right whisker that the digits beyond the upper quartile are more extreme.
The Standard Deviation
The last analysis of our social media data to look at is the standard deviation. The standard deviation tells us how far each of our numbers in our sample is from the mean. It can tell us if something is far away either above or below our mean.
The mean is what we usually refer to as the average. So add up all the numbers and divide by how many there are.
We can now find the standard deviation. I used an online calculator so that there wouldn’t be any errors. This one is more complicated than finding the outliers from the median.
As we can see most values fall between one standard deviation away from the mean. But we can also see what falls outside the range and this shows how our data varies too.
What are our goals with Social Media Data?
I think that we want a consistent outcome which means reducing our variability from month to month. We want a smoothly running machine. So we have an expectation of what those values should be.
Obviously, we want to see higher values for our likes and replies. That is something our outliers can tell us. Pay attention to the social media messages that outperform others by such a large margin. There will be clues into why they were so successful.
I plan on offering more ways to use statistics in the future to understand our content marketing efforts and our social media marketing. So please subscribe!
I can provide you with measurement and analytical insights into your blog in plain language.
- Setting up Google Analytics for your Blog
- Hooking you into Google Console
- Translate data into words so you know how your blog is performing
If you are interested email firstname.lastname@example.org
Featured Image SCREEN POST