Predictive Analytics and Self-Service BI
The Increasing need for Predictive Analytics
The world of Business Intelligence continues to change rapidly. Lengthy, costly, IT led projects are being replaced by more cost effective and user friendly options. Business Intelligence is no longer just for organisations that can boast the finances, the necessary infrastructure and the skilled resources.
Nowadays, driven by an increasing demand from small and medium sized organisations and the trend of “customisation”, BI tools are becoming more lightweight and powerful. This shift provides an opportunity not only for small organisations who, historically, may not have had the capabilities to drive their business forward by using data effectively, but also for large organisations seeking to reduce operational costs whilst maintaining a strong analytical capability.
SMEs in particular have become a huge market for Business Intelligence products, and as such, these businesses have begun to influence the direction and trends within the BI space. A key trend to emerge and become popular over the past 5 years is Self-Service BI, an approach to analytics that is designed to reduce the IT department’s involvement in analytical projects and increase the ability of users, regardless of their level of expertise, to visualise their data with tools that can be easily adopted and do not require knowledge of the technical protocols.
According to Bill Schmarzo, “Business Intelligence is the world of descriptive analytics: retrospective analysis that provides a rear-view mirror view on the business—reporting on what happened and what is currently happening.
Predictive analytics, in contrast, is forward-looking analysis: providing future-looking insights on the business – predicting what is likely to happen and why it’s likely to happen.”
When people have a piece of data, the first thing they want to do is explore that piece of data, to understand the information that it might hold. This is why descriptive analytics first became popular: it enabled people to understand what the data was reflecting. As time went on, people were no longer satisfied with merely visualising historical data, they wanted to understand future trends in order to make better decisions. Thus, organisations became increasingly keen to move on to the next stage in data reduction: Predictive Analysis.
Self-Service BI tools entered a second revolution, the aim of which was to enable users from any level, even those without background knowledge of statistics, to conduct analysis. This would enable the CEO of a small company for example, to visualise their sales by region and to build models that could predict the sales of a product next season, all done with little background knowledge of IT or statistics. The Marketing Manager of a company might also be able to visualise sales after the launch of a marketing campaign and build a model that would help them find their target customers for the next campaign.
Vendors in the Market
SAS, the market leader in advanced and predictive business analysis, provides a sophisticated tool – SAS Visual Analytics. Combined with the SAS Visual Statistics add-on, it provides the user with an easy-to-use interface to explore and visualise key data metrics, as well as an ability to create and compare different statistical models within the same user-friendly environment.
Many other BI vendors, who traditionally focused on reporting and visualising data, are now moving from traditional BI tools to BI tools that include prediction functions. Such vendors include Tableau, Qlik, Information Builders and MicroStrategy. These tools all allow R integration. They provide point and click GUI to generate R code, allowing the user to build predictive models in a code-free environment.
The Current State of Affairs
SAS Visual Analytics along with SAS Visual Statistics are the most sophisticated tools capable of combining the Self-Services BI and the Predictive Analytics currently available on the market. The main statistical model available includes Decision tree, Regression Modelling, Logistic analysis and text mining. However, the model is automatically generated, and little customisation of the model can be carried out. A higher level of customisation will require access to other tools within the SAS suite.
Most of the other BI vendors are entirely dependent on R. This can have both pros and cons. For example, limited customisation is allowed on the interface of the BI dashboard, and building your own regression model in the interactive dashboard is not supported – this is because most BI software will select the most influencing factors automatically for you while limiting your availability to add and drop specific underlying factors. However, a statistician who knows how to write R code could always amend the original R code to customise the model.
The predictive analytics functionality for Self-Services Business Intelligence tools continue to improve in terms of the wealth of algorithms, level of automation, scalability, model portability, web enablement, capability of access to large data sets and ease-of-use.
According to Gartner, leading industry analysts, the number one criterion necessary in delivering self-service data discovery and analytics is ease-of-use. They discovered that up to 41 percent of organisations cite this as one of the top three sought-after features when they considering an advanced analytics platform.
As the self-services BI market continues to grow and attract new users from diverse backgrounds, ease-of-use becomes indispensable. Predictive Analytics in Self-Services BI will require less coding. However, even though Self-Service BI technology is improving as a whole, both in terms of analytics capability, level of customisation and ease-of-use, there is always a trade off between level of customisation and implementation complexity. Vendors will go separate ways to differentiate themselves, depending on their technical capability and how “analytics-centred” they chose to position themselves.
Demarq is an independent, software agnostic consultancy. Our Analytics Enablement Service provides support across all aspects of the BI development lifecycle: from the development of data visualisations to the implementation of a full strategic BI reporting environment for operationalised analytics.
If you would like to understand more on how you can leverage the true value from your data then please contact us at email@example.com