As a Product Manager, you are expected to support your ideas and decisions with data. Working on a brief more often than not will require you to spend hours interviewing users and carrying out seemingly endless desk research.
Oftentimes you will find yourself lucky to have found the right ready numbers and statistics to back your ideas up (or better yet, you have a data scientist on your team who can cook them up for you). Sooner or later though, you will have to roll up your sleeves and dig through the raw data in search of diamonds yourself.
And let’s say you’re the lucky one, your company provided you with a copy of all the data collected on the subject over the years. Usage data, log of all the user interactions, or else. A number of records so large you don’t even know how to read it out loud.
How to make sense of all this noise?
You might, dear Product fellow, start by applying some of that Descriptive Statistics sauce.
So, what exactly are Descriptive Statistics?
To put it simply, Descriptive Statistics is just a way of talking about data (numbers) that is easier to wrap your head around.
Just like when you’re in a bookstore picking a book, you look at the cover, the title, and the pictures to get an idea of what that book is about. Similarly, Descriptive Statistics provide basic facts about the dataset that help us understand its characteristics.
We can group these statistics into four groups (but you can bet there are more to discover as we go down this rabbit hole):
1. Measures of Frequency
- Count
- Percentage
- Frequency
2. Measures of Central Tendency
- Mean
- Median
- Mode
These help you get some basic orientation, a way of locating the central point the data is distributed around. Usually, a great starting point when looking at a dataset.
3. Measures of Dispersion or Variation
- Range – Minimum and Maximum values
- Variance
- Standard Deviation
Say you just found out that users interact with your product 100 times a week on average. That might be a very good, maybe even great result. But don’t pour the bubbles yet.
Do the users in fact interact with the product that often, or maybe there are a few maniacal outliers who have been skewing the stats by clicking that thing day and night (a bot maybe or a bug in the code) while the majority of users try the product only once and leave to never come back?
Understanding the spread of your data adds invaluable context to the numbers like the mean or median. It will help you identify the golden nuggets of insights or narrow down the focus of your research right where it hurts.
4. Measures of Position
- Percentile Ranks
- Quartile Ranks
We will use these to describe how the values fall in relation to one another. If you ever received congratulations for scoring in the 99th percentile of all test or performance results, I’ll help you break down what it means. If you haven’t, keep on fighting soldier. You’ll make it one day.
“I think I’m starting to get the picture”
Fantastic.
Chances are you already were familiar with at least some of the terms we introduced above. Other, you might have not known by their name but deduced by looking at the graphs and data visualizations.
Time to get into the nitty gritty and learn where these numbers come from and how could we apply them to make you a better at all things Product and get you that job or promotion you were trying to manifest for the past few months.
More on that in the next post which I’ll link here as soon as it’s finished 🙂
That’s great thoughts man!