Paul Doyle, Director BI Services, GNet - A Nihilent Company
The challenges in implementing Business Intelligence (BI) and analytics strategy are varied and multi-faceted, and the selection of BI tools to support the strategy and business needs can be daunting. BI teams are continuously under pressure from business users to deliver flexible, fast, self-service and cost effective BI solutions. In addition to managing time to market and cost, data architecture, support, and maintainability also need to be considered, all while accommodating changes to business requirements. Critical success factors for managing expectations of an organization’s analytics platform include, aligning with business strategy, understanding user needs, balancing time to market and data architecture, and
Where to start
As organizations look to increase revenue, become more operationally efficient or build competitive advantage, it is essential for them to measure the past, discover current trends, understand cause and effect, determine appropriate changes and measure impacts of change in order to drive improvements. To that end, it is important to begin by assessing the current state of the organization and to understand the questions the business would like to answer. For instance if business leaders were able to answer how changes in temperature or changes in precipitation affect sales, what would they do differently? Without being able to answer questions such as how changing a supply vendor or part on a bill of materials would impact their product quality, what decisions are they unable to make? Once these types of questions are uncovered and an understanding of why it is vital for the organization to be able to make sense of the vast amounts of data at hand is achieved, analysis of how the current analytics and reporting tools are, or are not, meeting their objectives can be undertaken.
Time to market
Two significant challenges exist with delivering on an organization’s analytics needs, time and everz evolving business requirements. In a typical scenario, analysts spend an enormous amount of time gathering, massaging, standardizing, grouping, segmenting, and loading data from a myriad of systems in order to gain insights. In traditional data warehousing implementations, this challenge is tackled by BI and IT teams through data modeling and data extract, transform, and load (ETL) processing. These types of solutions can alleviate the time it takes for an analyst to collect data but may miss the mark in terms of timeliness of providing answers when the lifespan of the requirement is not considered. Analysts require tools that provide the flexibility to gather data and produce reports in formats they deem best suited for consumption, provide the ability to share reports with the rest of their team or organization, and provide the flexibility to work at the pace of business.
Know the user base
In addition to aligning with business strategy and delivering at the speed of business it is important to factor in the types of individuals who will be working with data and their role within the organization. BI solutions and tools are seldom one-size fits all, frequently needing to support standardized reports, actionable dashboards, self-service reporting, ad-hoc analysis, and data discovery. For instance, sales professionals may require eye-pleasing visuals on a mobile device,which quickly show where action is needed and which include a guided experience to get to root cause;an inventory analyst may require a self-directed experience with the ability to create content and derive new metrics for the organization; a restaurant manager may need IT certified printed reports. It is important to understand the different usage patterns of the BI solution and to provide tools to meet those needs.
Given the explosion of data volumes - the rate at which data is created, and the importance of data security - it is more important than ever to create robust data architecture to sustain an organization’s foundation for analytics. The data platform needs to provide functionality to support structured and unstructured data, and securely bring together on-premise data with external data or data stored in the cloud. While relying on traditional data orchestration methods of processing data in batch with ETL alone is not sufficient going forward, ETL tools will continue to be part of the equation.Evaluating the ability of ETL tools to orchestrate structured and unstructured data both on premise and in the cloud is essential.
Taking it to the next level
Organizations are increasingly looking at advanced analytics to gain a competitive advantage and differentiate themselves in the marketplace. With this, the analytics platform needs to support more than just traditional aspects of BI like reporting, querying, pivoting, and slicing and dicing. It needs to enable going beyond being informative to being predictive and prescriptive, including such features as built-in machine learning algorithms and support for creating proprietary algorithms and visualizations through technologies such as R and Spark. In addition, the analytics platform should support the use of predictive and prescriptive analytics by line of business users, not just data scientists, and provide the capability to incorporate advanced analytics models into line of business applications.
BI and analytics solutions can provide significant advantages to an organization for revenue growth, operational efficiencies, cost savings, and competitive advantage. Understanding the analytics needs and objectives of the business and providing the right tools and approach to meet these needs are key to a successful BI journey.