Dazzling, dynamic data dashboards…or duds?

Data dashboards are probably one of the most common analytics tools used by companies and interestingly, they are not unlike your vehicle’s dashboard.  Think about that for a minute.  In your vehicle, you have a lot of important pieces of information directly in front of you that help you make decisions.

Your Mechanic website

Your speedometer tells you how fast you are going.  This helps you adjust your speed depending on speed limits or if you see a photo radar vehicle ahead.  You also have a fuel gauge.  This is useful so you can prevent getting stuck somewhere because you ran out of gas unexpectedly.  Then there’s my favorite, the check engine light.  Ok, I’m sure many people have ignored this indicator – Penny on the Big Bang Theory sure has – but it is an important leading indicator that your car might just stop at a most inconvenient time if you don’t check that engine soon.

Now think about all that information for a minute.  Vehicle dashboards are designed to guide you and provide you with useful information so you can make informed decisions about your driving and your vehicle.  Data dashboards are also meant to provide guidance to businesses by helping them operate more smoothly and avoid making costly or fatal decisions.

Unfortunately, many companies get lost in the data or technology behind dashboards and often miss the point of the dashboard entirely – to help guide the business in making the right decisions at the right time and with the least amount of confusion and complexity.  Dashboards are supposed to provide structure to the massive amounts of data most companies collect and focus on the key performance indicators (KPIs).  It seems so simple but so many companies get it wrong. 

In my experience, here are 6 ways companies have created complexity and confusion where it’s not necessary.

TOO MUCH DATA

More is better right?  One might think that the more information you have, the more informed you are.  This is not always true.  Too much data can result in analysis paralysis and can lead to bad decision making.  Putting the wrong data in a dashboard – even if it’s accurate – can distract and make people lose focus. 

The data in dashboards needs to be focused and targeted towards answering specific questions. Imagine the image below is how your vehicle dashboard looked. You’d definitely need more training in driver’s ed to understand this but the bigger question is would you need all this information just to drive around town? How would you keep focused on the road and getting to your destination safely?

Pinterest – Bill Donoghue

So, how can this be mitigated?

• Start with what questions you need answered.  Don’t pull a bunch of data and then try to figure out what questions you want to answer (or ask for that matter).  If you don’t know your actual KPIs in advance, you will never know when to stop putting more information in the dashboard.   You will also never know how you are doing and where to make improvements that have the biggest impact. 
• Create thematic dashboards if you need to track multiple aspects of your business and ensure you review them in different meetings. For senior leaders, you might create a dashboard with a key metric from multiple areas – a business health check dashboard.  However, for department leaders a separate dashboard with more detailed metrics might be needed.

FOCUSING ON TECHNOLOGY

There are so many business intelligence (BI) tools out there, it is difficult to determine what is really needed. There are websites that tell you the top 15 tools, or the 24 best tools or what are BI tools and then when they have your attention, promote the tool(s) they sell. Every company with a BI tool has widgets, gadgets, portals, and solutions that promise to make your dashboards dazzle.

Make them interactive.  Real-time data.  Data visualization.  In the cloud.  Easily accessible. Collaborative. Easy to share.  Connect multiple data sources. Data is in one place.  Predictive analytics. 

It all sounds perfect, doesn’t it?

Some companies will jump from tool to tool because they think being an early adopter gives them an edge in the crazy new world of data science.  What actually happens though?  They end up going counter to the purpose of the technology in the first place.  Why have multiple tools and technologies if each one is supposed to streamline the work, reduce effort, and put everything in one place? 

The focus also strays from using data to make decisions and towards what the tool can do or putting in time and effort to constantly implement new tools.  Before you go to a new tool, assess what you want that tool to do. Understand what you need from a tool first. Then you can figure out what tools would be appropriate for the job. As mentioned, there are lots of sites that can help you narrow your search. If you put in the effort to pick the tool that works with your business, why would you keep changing just because it’s newer? 

REAL-TIME DATA

If you can’t respond in real time, why do you need real-time data?  In your car, you can adjust your speed, fill up your tank or release your emergency brake when you see the dashboard information.  If you can’t react in real-time to the information in your data dashboards, why do you need to see the information in real-time?  If your tools allow for this AND your data is easily available as a live feed, great.  Feel free to take advantage. Otherwise, put your resources to better use, such as doing deep dives, machine learning, or predictive analytics on your data to find actionable insights.

VISUALIZATIONS

Almost every tool can do visualizations.  I’ve seen some really pretty dashboards that contain pretty useless information. I’ve also seen some very simple dashboards that provide just the right information. Unless you have a real need to make something simple to visualize because the information is inherently complex and contains a lot of varied data, almost any tool can provide the basic visualizations that most companies need.  Maps, bar graphs, line graphs, even pie charts – almost all tools have them.  Use visualizations that people will understand and make the data easy to interpret.  Take it easy on colour too. Keep it simple, remove distractions, and keep the focus on what the data is saying. 

NOT HAVING TARGETS OR GOALS

Data without targets are just interesting facts.  If I told you I grew a pumpkin that weighed 10 pounds, you might say “Wow! That’s one big pumpkin!” But if you knew that the average pumpkin weighs about 12 pounds, you might not be that impressed. A target gives you context and a point of comparison for you to measure improvement.

One of the main reasons we have a speedometer in our vehicles and don’t drive based on our own comfort level is because the law sets speed limits as a safety mechanism and to prevent deaths.  We need to set benchmarks, goals, or limits on dashboards as well. Otherwise, how will we know if we are doing well and should continue our actions or are doing badly and need to improve and come up with new ideas?

INACCURATE OR BAD DATA

Don’t bother with a dashboard until you fix underlying data issues.  It’s worse to make data decisions on bad data.  If you can’t calculate something properly, find a different metric or spend time fixing your data.  Now, this doesn’t mean your data has to be perfect but know your margin of error and don’t use data to make decisions if you don’t trust how it’s collected and analyzed.

Harvard Business Review did a study in 2016 that estimated bad data costs the US $3 trillion a year. Another study by Gartner estimated that companies were losing, on average, about $15 million per year in 2017. Data quality is a real issue and a real challenge. It is important to fix the root cause rather than publish bad data just to say your company uses data dashboards. As the adage goes: garbage in, garbage out.

AND SO….

When companies do any of these 6 things:

1. Put too much data in a dashboard
2. Focus on technology instead of what they need from the data
3. Trying to use real-time data when real-time reactions are not possible
4. Over complicating visualizations
5. Not having targets
6. Using inaccurate or bad data

it will often lead to lack of adoption of a dashboard, dissemination of inaccurate data, or outright abandonment of the tool or dashboard.  If a dashboard is not useful, too complicated, inaccurate, or people don’t know how to take action using the information, this will lead to bad decision making. People can also become complacent in their roles or, worse still, will think their “gut feelings” are a better way to make decisions.