5 Tips for Analysing Data

When you think about analysing the data in your business it can feel quite overwhelming. Having large amounts of numbers staring back at you can send many people running for the hills. If that sounds like you, then read on.


In this post, I want to provide a few key points that will help you break down the barriers of data analysis and help you unlock some of the secrets that are hidden within the numbers.


So, let's take a look.



1. Use context


When looking at data the first thing that I would encourage anyone to do is give your numbers context. This context is vital in understanding performance.


You may look at your data and see that you have sold fifty units in a week. But is that good or bad? Is that the same or worse than last week? Or is it above or below target?


Adding context to your numbers helps you understand more about the number than just its value, it will tell you about its relativity to other values. Using a traffic light system will provide a visual indicator as to whether the values are good or bad compared to their comparison.




2. Consume Your Data One Bite at a Time


The first thing that you need to do with your data is to break it down into chunks. These bite-sized chunks help you look at your data at a high level without drowning in the detail.


If you break your data into departments, categories, segments or geographies, it will help you look at groups within the data rather than looking at everything together. This will allow you to see the particular sections of your business where you need to look at in more detail.


Imagine that you have one hundred products that you sell. Looking at all one hundred of them in one go may tell you the largest and smallest, but you won't be able to easily spot whether similar items are all selling in the same way. When these one hundred items are divided up into segments, you can see which segment is performing best, and then within each one see which items are performing best.


Having multiple types of segments can also be useful. Such as characterising by both colour and size may tell you different things and show how different cuts of the data are performing relative to each other.




3. Create a hierarchy


Now that you have segmented your data into chunks, you should try to arrange them into hierarchies. This will help you follow through the data in a structured way. A hierarchy is essentially a set of attributes that can be layered on top of each other to group the data into levels, with each level getting more granular as you go down through the levels.

For example, you could go from World to Continent to Sub Continent to Country to Region to City. There is only 1 World, which divides into 7 Continents, 15 Sub Continents, 195 countries, countless regions and even more cities.


You would then start your analysis with the largest grouping and work your way down to the lowest level. So identifying the Continent of interest, then the subcontinent, country and so on, until I get to a city.


By following the data in this way you don't have to digest too much in one go, you can use the levels of the hierarchy to tell your data story and direct your focus to the next interesting data point. Other examples of hierarchies could include manufacturers and brands or company departments and functions.




4. Trend It Out


Using trends will show the trajectory of a value over time. This will help show whether the overall trajectory is up or down, whether a particular period in the data is driving the overall picture or whether a certain segment is performing differently over time compared to the others.


Trends can be set as long or as short as the data will allow and can tell different stories at different points. Looking at a set of over 10 years will provide a much greater level of context than looking at the last week, but yet both are very relevant in different situations.


You can also add time periods together to help take some of the peaks and troughs out of the data. Creating a time hierarchy, that is based from Year, to Quarter, to Month, to Week to Day will allow you to follow the data through a year without needing to look at all 365 days in one go.




5. Action!


One of the biggest sins of data analysis is the lack of action from its findings. I have seen first hand how the use of data can change the trajectory of a business through changing decisions based on the findings. Making sure that the information isn't wasted is a very important element, especially once you have managed to filter through all those data points to find the golden nugget at the end.


It should be noted that the data will always be able to tell you 'the what', but it can't always tell you 'the why'. Understanding the true drivers to performance and how you change it may be down to external factors that aren't carried within the data, but once you know what is happening you can set about finding out why it happened.


Once you have managed to get to why something is happening then you can see how you can either fix it, in the case of a decline or do more of it, in the case of growth. The important thing is though, to do something with it.



6 views

Recent Posts

See All