We normally associate analytics and data science with mathematics and logic. What is often forgotten is that the data we analyze represents real world situations, people and businesses. This means that before we get to the stage of making complex calculations based on the data, we need to first ensure we’re accurately interpreting everything, mapping what will be analyzed and talking with everyone involved in the data project.
It’s also important to understand that the data that is processed and analyzed must be able to generate actionable insight. Although data is the new oil, it only truly offers value when it delivers meaning and knowledge. As a result, alongside the all-important work of organization and analysis, communication is a key focus for data professionals, ensuring data generates insights, enables decision-making and supports businesses in building a data-driven culture.
Here are some of the key techniques that are used to ensure that data is communicated clearly.
Measuring what matters
The first stage in ensuring data provides valuable insight is determining which aspects of business performance are being measured and how this information will be quantified.
If, for example, a business knows its website received 100 visits this month, does it also know whether this represents a high or low number?
Once it knows that, in an average month, the website receives 10,000 visits, it becomes clear that the 100 visits shows a significant drop in traffic. If, however, the site typically receives 10 visits, the 100 visits represent a growth of +900%.
Always think about having a comparison baseline. Gaining this additional contextual information also allows you to set realistic goals and compare how close you are to achieving these.
In short, it’s important to be selective about what is being measured and to remember that an absolute number is unlikely to give clear answers without relevant context.
Keeping language clear
Once you’ve made sensible decisions on what is being measured, it’s important to communicate this in language that everyone can understand, including people working in different areas of the business. That means standardizing the terms used in dashboards and reports so that each indicator is represented with a unique and clear name.
Terms like “revenue_1” and “revenue_2” won’t be entirely clear, while “revenue_online” and “revenue_instore” will ensure people across the business understand exactly what data is being gathered.
Also, avoid using abbreviations when naming variables if there’s a chance of confusion. Choose “conversion_rate” over “CR”.
Clarify with graphics
Any visual representation of data needs to be carefully considered, with graphics chosen according to what exactly is being communicated.
This can help with the creation of clear and concise graphics and avoid visuals being difficult to interpret.
For example, for a visualization where you need to include many dimensions, a pie chart might not be the best option, as placing many different “slices” becomes a loaded visual and difficult to interpret.
It’s also important to check that captions are visible and that users will be able to read all values.
Adding color to the insights
Many companies prefer to use their own branded colors in dashboards and reports. Others prefer to put positive numbers in green and negative numbers in red. The important thing is that there is a plan informing how colors are being used and that this makes interpreting data simpler.
For example, if I want to highlight a dimension within a chart, I can put a stronger color on that dimension and a weaker color on the others.
You should also avoid using too many colors as this can simply cause confusion.
Dashboards and reports are powerful tools to generate knowledge, so it’s important to develop a clear plan, which considers the story you want your dashboard to tell. It’s also possible to create an initial version of your dashboard and ask users for feedback after a few days of use so that you can make adjustments if necessary. Analytics is about continuous evolution and you can always improve your analytics and reporting.