One of the hottest trends today in the business intelligence and analytics spaces is “self-service.” The word self-service is thrown around lightly in many situations and often carries different expectations for different people and organizations.
Before we go into the details of self-service analytics it is important to have a quick view on what self-service is in general. It is important to understand, and remember, that successful self-service empowers users to take on the responsibility of carrying out most of the activities as part of their decision making cycle in an independent manner. This includes accessing data, generating insights and sharing information with little dependency. However self-service does not mean that users are on their own. In every self-service driven organization, IT plays an important role of making sure systems are available, that they contain sufficient and high quality data and for ensuring that users have appropriate capabilities.
From WHAT to WHY
Traditional business intelligence (BI) approaches largely utilize existing data from the data warehouses and focus on enabling the user to answer the WHAT part of the business query. Examples of WHAT style business queries could be:
- What have we sold this week?
- What is our current sales revenue?
Though there are traditional “self-service” BI approaches delivered in the market they mainly cater towards making it easy to generate typical BI outputs which people like to use to look at their “What” data such as reports and dashboards. The issue with this is that once displayed the data is rather static.
Not satisfied with this limited static view of information, and driven by the need for more information and deeper insights, business users have driven the BI market to adapt and deliver capabilities that go beyond standard reporting and dashboarding to deliver discovery approaches. These discovery approaches enable users to carry out more investigation and interrogation of the data to find the answer to WHY part of business query. In some ways this can be seen as trying to understand the root cause of the WHAT.
Examples of WHY style business queries could be:
- Why are my sales down in France?
- Why are we not hitting our sales target?
In many of the tools seen in the market focusing on the data discovery aspect of business intelligence the ability to get answers to the WHY of any business query is still limited. Many of these self-service discovery tools still have pre-defined path of analysis that do not allow the end user to follow unique paths of thinking that may not have been previous envisaged.
Data visualization versus analytical visualization
When meeting with various customers and their community of business users I have often come across the fact that users are not clear about the difference between data visualization and analytical visualization. There are many ways one can view data and interact with data, and there are numerous highly interactive and innovative ways of visualizing information, but most of these methods are still showing users their same historic data now wrapped in a beautiful and shiny wrapper around it.
Visual analytics or analytical visualization, though, is based on the principals of data visualization (ability to communicate information clearly and efficiently to users via selected information graphics), it takes users to the next level of exploring their data where visualization is coupled with analytical techniques for obtaining deep insights towards planning and decision making.
A simplest example is using a line chart graphic as a visualization. Using data visualization techniques, a user can do a highly visual trend analysis of a key metric over a period of time and do various interactions of filtering data, comparing trends across periods or even run some what-if analysis with pre-defined calculation paths. In this visualization and discovery process there is Analysis but not Analytics.
When it comes to analytical visualization, it would take the historic trends as inputs to sophisticated analytical algorithms (such as forecasting) and provide future predictions for the key metric. Application of analytics to this data can further extended the scope of analysis from predictive to prescriptive, yet powered by extremely easy to use visual and interactive methods to run scenarios which are much more advanced than pre-defined what-if calculations.
Another example I can share is from the analysis of unstructured data. With an enormous amount of information now being captured in unstructured data channels, such as emails, twitter feeds, social media discussions, online user comments, etc., not analyzing this source of information can leave organizations trailing in the wake of their competition.
Many visualization tools can process unstructured data, generate new metrics and represent information say in the form of a word cloud. However, it's only with the application of analytics that users can not only view the highly visual word cloud, but they can also do more sophisticated analysis such as understanding the consumer sentiments hidden in the textural information and thus make more informed decisions.
Fear of analytics
Even though business users want to make more informed decisions by using analytics, they struggle due to their fear of analytics. This is mainly due to two factors:
- Most end users lack the skills (which could be built with appropriate training) for handling analytical tasks.
- Most of the tools on the market lack approachable analytical capabilities, meaning they either do not exist in the tool or they are so complex that users require a lot of hand-holding before they can feel comfortable using them.
The end result is that most business users still ask their analytical experts to generate possible scenario outcomes, forecast results and to convert the answers into visual results that are easy to interpret. This whole approach is contrary to the basis of self-service.
Analytics for the masses
It is a known fact that using analytics improves businesses. The art of getting analytics into your business to improve it is to make analytics available to the masses. Self-service analytics is the key to getting more people using analytcs in your organization.
Success comes only when analytics can be utilized closer to the business, which requires an environment that makes analytics approachable and easy-to-use in a self-service manner. Organizations that have managed to deploy approachable analytics into the hands of their users have been rewarded with outcomes that contribute directly to reducing the organizations bottom line or growing its top line.
Self-service analytics has also brought additional side benefits such as freeing up existing analytical experts to focus on bigger projects and more strategic tasks rather than the analytics needed to answer day to day user questions. Such an approach has helped to bridge the gap between the silo’d organizational structures of business users, limited to reports and dashboards, and the analytics experts supporting efforts to move organizations to be analytically driven.
It is important to remember that just giving access to self-service analytics on its own is not sufficient. It has to seamlessly integrate with the self-service BI capabilities. This helps and encourages users take one step at a time towards analytics and get used to it within their familiar BI world as well as enabling the sharing of analytically driven insights.
SAS has helped organizations empower their users with self-service analytics by combining the power of BI and analytics in a single solution, >SAS Visual Analytics, which is a great starting point for helping their users take their first step towards analytics. Analytically mature organizations can leverage SAS Visual Statistics, a simple to use play-pen environment for analytically driven data discovery and to build and refine predictive models extremely fast.