Jump start your analytics program with visualization and communication


With so much information available about high-performance analytics, business intelligence and visual analytics, it can be difficult to know exactly where to begin, especially if you don’t have a team of statisticians standing by.

I'm frequently asked by customers who hope to take advantage of analytics how to get started.  How do you advocate for analytics without dragging your organization behind?  How do you show the value to your highly effective counterparts who like the way things have always been done?  If your company lives on a highly operational mindset, creating that culture of analytics can be a challenge.

Thankfully, analytics covers a huge range of opportunities, and it doesn't take an army to get started! Below are some techniques to start creating an analytical program using the resources you already have available.  

Understand the business need

When approaching any new analytics effort, it is critical to ensure the objective and focus is centered on an actual business problem or need. Performing analytics for analytics' sake can be fun, but it ultimately won’t be as effective in creating analytically-minded organization. So, start with talking to the business. What are their current pain points? What is their current focus? Who can talk with you about tangible needs?

I recommend discussing opportunities with a person from the business as it gives a dedicated avenue for sharing your results and can help you build your network for discovering future opportunities. Talking to the business doesn't mean you need a sit down with a senior executive. Take advantage of the contacts you have. When talking with the business, be aware that as you are starting out you may not have the data, systems or statistical expertise to solve every need you encounter. Therefore, listening and avoiding setting unobtainable expectations will be important.

Revitalize your reporting

Frequently, companies can find themselves overrun with hundreds of operational reports. If your business puts a lot of emphasis on reporting, you can capitalize on this opportunity!

Without completely re-architecting what you have, consider adding some new reports or visualizations that focus on providing decision making intelligence or even adjusting a few reports already in use. Keep it simple and targeted. So, what are some tips for designing reports to answer the “why” questions? Consider making your reports more interactive. Below are some specific ideas for how to put more power in the users’ hands.

  • Allow users to start at a high level and then subdivide the data into various categories or time periods. Not sure what this means? Think pivot table.
  • Provide the ability for users to select a subset of a report then jump to the detailed records behind those aggregated data points.
  • Display data geographically on a map. Depending on the context, mapping your data to its geographic region can be very informative.
  • Interlink your reports so that selecting data in one report highlights those same data elements on another report. This could highlight possible correlations in the data.

Remember, the goal is to give your users the ability to further investigate the numbers and details behind a report so they learn why things are the way they are. This can be a great first step to creating a culture of data-driven intelligence.

Start with the basics

Not all needs require complex statistics and a PhD! Just using simple descriptive statistics (such as mean, median, mode and standard deviations) while spending time profiling and reviewing the data can be very insightful. Be curious and ask the “what if” or “could this be” questions. The key here is use what you know. Below are some ideas to get you started:

  1. Profile and prepare data - It's important to first understand your data. Before applying any analytics, spend some time just looking at the data (i.e. profiling). This can be using a formal profiling tool or reviewing the data manually. Sometimes it is very helpful to meet with a subject matter expert to understand its idiosyncrasies. What data and metrics exist? Are they spread out across multiple sources that need to be joined together? Is the data to support your analysis sparely populated? Are the data metrics normally distributed? Are any transformations or standardization required? As you review and prepare the data, it will give you ideas for where to focus. Typically in an analytical effort, 80 percent of your time is involved in data preparation and cleansing. Therefore, don't overlook this important first step!
  2. Investigate graphically - Once you understand the data you are analyzing, consider reviewing the data graphically. Sometimes it can be easier to identify areas for further analysis or standardization through visualizing the data. This doesn't require a fancy graph – just a simple scatter plot, box and whisker plot, pareado chart, or geographical plot can be very insightful. Depending on the effort, sometimes you can meet the entire business need merely by graphically representing the data in a new way.
  3. Perform analytics - As mentioned above, you don't have to get fancy with your analytics to provide valuable information to business leaders. I am often amazed at how many business questions can be solved using simple statistics and data visualization. The statistics used in this step will be highly dependent on the question you're trying to answer. Consider starting by using the metrics you have available to review the averages and standard deviations. If you are interested in comparing different groups, consider using t-tests, which allow you to ascertain if the different groups averages for a specific metric are statistically different or not. For example, if everyone thinks customers in one state typically buy more than customers in another state, you can compare the average revenue from each state using a t-test to determine if there truly is a difference statistically. Not all problems can be solved with basic statistics; but as you are just getting started, stick with what you know.  
  4. Organize results - When determining how to best articulate your findings, remember the business is focused on results and not the methods used to get there. So, start with the bottom line. What is the answer to the question they are asking? When presenting your results, keep it simple. If you need to showcase the data, use charts and visualizations where possible. If the business question couldn't be answered (and this will happen from time to time), can you find some other value to provide that will aid the spirit of their initiatives and objectives? I recommend not coming empty-handed where possible. Next, follow the results with all pertinent assumptions (for example, if the data you were analyzing was sparsely populated, this would be important to note). The goal here is to avoid being overly technical while still raising awareness.
  5. Offer a recommendation -  Finally, conclude your results with any follow-ups or recommendations. These could be in the form of ideas for further analysis, recommendations for additional data collection or strategic suggestions to enable the tracking long term for this type of analysis. It is recommended to tailor your recommendations around the audience; if this individual has no authority to change the way data is collected, then suggesting different data collection practices may not be relevant or helpful.

Publicize your victories and build on past success

After your first success in helping the business use analytics, find ways to publicize it. This will help demonstrate the power of analytical thinking, plus it helps use data to support business objectives and hopefully creates more opportunities for future work. If you have a company newsletter, find out if your story could qualify. If you have an internal company website, see if you can get a spotlight feature or article added to the front page. If those options are not available, use what you have even if this is just word of mouth. Take advantage of the opportunities that come through the publications while continuing to seek out new ones as well.

As the successes continue to grow, you will build justification for bringing on additional staff with statistical backgrounds or analytical modeling experience or getting formalized training in statistical modeling. Building a culture of harnessing analytics to drive your business doesn't happen overnight. Start small and stay committed. With each new victory you will reinforce the power of analytics, and over time a new culture can emerge. Remember, the goal is to empower business leaders with data-driven intelligence. As you demonstrate this capability, your business will grow and so will your opportunities to continue using analytics.


About Author

Jennifer Nenadic

Manager, Enterprise Analytics Services

With a background in data management and analytics, Jennifer has aided SAS’ leaders and external customers in strengthening their business by strategically transforming their current systems into intelligence-generating solutions using advanced analytics techniques and data management best practices. Jennifer has worked with clients across various industries to find creative designs and solutions to meet their evolving business needs. Jennifer holds Bachelor’s degrees in Computer Science and Textile Engineering as well as a Master’s degree in Advanced Analytics from North Carolina State University. Follow me on Twitter! My username is @JenniferNenadic

1 Comment

Back to Top