At ClicData we try to keep things simple.
Like many others, we do frequently use terms such as metrics, measures, dimensions, and facts, but we find that there can be a vast difference in their meaning from one person to the next.
We hold a vision to make the field of data information analysis and data visualization accessible to everyone—so we use a set of definitions we’ve devised that are based on the experience we’ve had in the field and the work of experts that we respect.
Representation of the data
The objective of these terms is to provide businesses and individuals with valuable information gleaned from data. Today, information can be delivered to us humans either visually or audibly. The rest of our human senses—smell, touch, and taste—remain unexplored avenues for software applications to communicate business insights to human beings.
For the most part, the best sense by which we human beings can obtain a wide range of differentiation is our visual sense: our eyes. Through our eyes, our human brains are capable of detecting hundreds or thousands of subtle variations through visual cues. Audio input wouldn’t lend itself as well to detecting the differences between columns or pitches or frequencies.
Stephen Few illustrates how the brain sees and potentially understands many values at once in his book, Now You See It: Simple Visualization Techniques for Quantitative Analysis. He describes the many ways that the brain can quickly differentiate between variables such as width, color, length, shape, orientation, and location.
Today, data visualization is the best means to communicate data and information.
Surprisingly, the types of visualization methods in use today are not much different than those that were available 50 years ago or more. Apart from those in use by a few visionaries in this area—including, Stephen Few, D3, HighCharts, Google Charts, and even the news media as they attempt to simplify (and sometimes over-simplify to the point of being wrong) data to their viewers and readers—the list of data visualization mechanisms has not changed much for years.
They still include tables, charts, gauges, and other similar indicators. Even a single phrase, word, or number can report a metric.
Here are some examples of data visualization components:
These data visualization components (what we in ClicData call widgets) sometimes do not provide enough information. In order to spell out the full picture, it may become necessary to place several of them together, to link them together to provide more detail (drill-down), or to put them in context of other related information.
Some data visualization components are highly condensed, providing aggregated views of the data. Others are very detailed and require more space than a page or a screen can provide. The level of aggregation or detail depends on the information that the user needs.
If a manager wants to know if a customer has paid a particular invoice or wants to get an overview of the customer’s activity over a period of time, they can get that information simply and effectively from a report—a table that contains rows of data that spans one or more pages.
But if the report consists of many pages, our brains have difficulties consuming and assessing the information. Trying to remember facts from page 3 when you are looking at page 18 requires more effort than if everything was in the same physical space. Still, when used for reference purposes only, reports can be highly effective tools.
More often than not, and especially now that data is so abundant, we are better served when we can aggregate or condense the data to give us a deeper insight into what the information is telling us. Line, bar, and column charts, and pivot tables are one of the first tools we reach for. If we want to identify correlations between two or more numbers, then x/y charts, scatter charts, and bubble charts will be quite useful. Many people make the mistake of using pie charts because of its visual appeal rather than its capacity to deliver accurate visualization.
The difference between measures and derived measures
To produce these types of visualizations, we need numbers—numbers that measure anything from test scores to the number of sales in units to something (or someone’s) performance.
Sometimes we can access the numbers simply and directly; other times, we need to aggregate them or average them. These measures—such as sales revenue, number of visits, number of units sold, school grades, weight, height, age, blood sugar content—can sometimes be used to create other numbers. For example, mass and height can be used in a formula to determine body mass index. This is called a derived measure, but it is still a measure.
Some numbers are not measured at all; they are created as an objective, a target, or a measure to be used alongside another measure to create a derived measure. So a target measure of 20 units, for example, combined with the actual measure of 18 units can be combined to create a derived measure of 90%, which is the percentage indicating how close we are to achieving the objective.
This approach is useful because it “normalizes” all the data into one metric, and that allows comparison. It is much easier for the eyes and brain to capture the right information if the data is normalized. It is the equivalent of not having to flip between pages of a report, except it avoids having to remember temporary values, wasting previous memory, and processing brain cells.
Some derived measures are easy to calculate. Others are much more complicated and might require access to data from different data sources.
In the scenario depicted in the tables above, the sales targets for each territory were created by a sales manager. She used Excel to estimate and adjust the target measures so that they meet the overall national objective. The actual measure, the number of units sold, are not tracked in the same way. Most likely each time a sale is made, a record of that sale is registered somewhere like an order entry system or an online web e-commerce application. The challenge becomes bringing that data together in such a way that combines the two measures and aggregates them at the right level.
Ask yourself, what do you need to know?
Once you understand the data and metrics available to you as well as what you can potentially derive from them, the biggest challenge you will ever have is determining what you need to know: in other words, how to transform your data into usable information.
There are many paths to do so. For example:
- You can study the data and attempt to visualize by employing a variety of data visualization components. This is called exploratory (bottom-up) data analysis.
- You can determine what is important for you or your business to visualize on an ad-hoc or frequent basis and identify the data needed to support those indicators. Some people define key performance indicators (top-down) and work backward towards the data.
Whichever direction you go, bottom-up or top-down, we recommend that, no matter what, you don’t start with technology, you start with your head.
Technology can help but not if it distracts you from the data or the information you are trying to derive from your data.
Try this exercise.
Imagine that you are the head of a large corporation, and you just arrived back at work after a long vacation. You would like to know how things are going enterprise-wide. But nobody is allowed to talk to you. Each department head can only hand you a single sheet of paper, and on it, they can put whatever they want to fully communicate to you about the overall status of that department.
What would those pages look like? What information would they need to contain to make you immediately aware of where to focus next or what area to dive into for more details next?
If you are an owner of a small business or running a household, the same exercise is applicable. Imagine that you walk in the door and you want to look at a whiteboard that gives you all the essential pieces of information you need to know about your responsibilities.
What would the whiteboard contain? Something like this? Would it have more detail? Less? It is important to focus on what is essential to run your business or household.
Think of dashboards as these single pages of information, an aggregation of metrics and indicators that allow you to act on them.
- It is the thing you do in your head before you make your grocery shopping list and decide what to buy.
- It’s the calculation you do in Excel to review your cash flow and sales numbers in order to determine if you can hire that next employee.
- It is the preparation you do in PowerPoint before presenting it to your bosses.
Now make it happen automatically and on a frequent basis so that you can focus on the actions to take based on the data—and not have to spend time on data preparation, aggregation, presentation, and publication. In short, a business manager needs to have key metrics available at all times to make meaningful decisions.
The metrics and performance indicators you use to reap the information you seek will depend on your business, your focus, and your responsibilities. What is essential to a Finance Director may not be important to the Logistics Director, for example.
In a nutshell
In this article, we outlined the importance of identifying your metrics, your derived metrics, and your performance indicators.
We revealed that you can conceptualize dashboards “bottom-up” or “top-down,” but whatever you do, we advise that you don’t start with the software. Even if an app is easy to use and includes so-called “intelligent” features, they don’t understand the needs of your unique business. So start with your head.
And we provided a simple exercise to help you understand the best approach to designing dashboards for your company.
Ready to turn your Data into Information?
Connect, normalize, combine your business data and create interactive and real-time dashboards to make informed decisions with ClicData.