All too often, business leaders are told they can’t have the analytics they want and which they really do need. “Not yet” they are told. “Only after we ….” [insert your favorite time consuming, expensive prerequisites] Is the all too common refrain. As a result, there are many situations where organizations postpone, water down, or even abandon initiatives to improve analytic capabilities.
Here are some real-world examples:
A large manufacturing company acquires several smaller players. The senior executives want to have an integrated view of performance metrics across the newly expanded enterprise. They are told that before that can happen, the newly acquired businesses must be moved to the acquiring company’s ERP system.A natural resource company reorganizes itself moving from an integrated structure to separate business units. Senior leaders want to be able to analyze data about new businesses. They are told that the prerequisites include upgrade and modification of the ERP system and an overhaul of company data structures.An energy company starts down the path to add new and improved analytic capabilities. After some initial success, and new insights start to emerge, executives push back pressing to wait until standardization of processes on new systems.An insurance company with hundreds of individual brokerages wants to have a near-real time view of its overall risk profile. The CFO is told that because the individual brokerages are not on common systems, it will be prohibitively expensive to develop this capability.
Sound familiar? What were once legitimate barriers to progress on data analytics are now merely complexities to be overcome. Business leaders should no longer be told “not yet” or that they will have to wait years to get the information they seek. “Analytics first” should be the default position in today’s cloud & app-driven world.
For years, ERP vendors and implementation partners had advised their clients that business processes needed to be standardized on a foundational platform like an ERP as a prerequisite to establishing data analytic or business intelligence capabilities. Or, in cases where multiple business process systems persisted, or data sources were many and varied, a complex data warehouse project would be a prerequisite. This advice was primarily based on the state of technologic capabilities and implementation methods. It was widely accepted as best practice for a long time.
An unintended consequence of this historical approach was that many organizations finished their ERP projects without ever getting to the improvements in data analytics. The transactional systems for finance, supply chain, HR, and customer processes were completed. The data work that accompanied these projects focused on completeness and accuracy in the context of a transactional and accounting environment. Even then, however, ERP implementations did not fully meet the needs of finance and accounting. Additional systems and tools were necessary to support planning, budgeting, forecasting, reconciliation, consolidation, and reporting. The transactional world overshadowed the needs of business analysis.
Further, the completeness and accuracy requirements of historical data was part of the reason for delays in delivering analytic capabilities. Consider the data cleansing and preparation work for ERP implementations, which is time-consuming and resource-intensive. Data quality in the context of performance management and analytics should be a function of its use and value to the business, which is different than its role in transaction processing and even historical reporting.
Information that is timely, meaningful, and actionable in the context of business analytics generally does not require 100 percent data completeness and accuracy to be useful. The pursuit of data completeness and accuracy as a prerequisite to build and use data analytics capabilities is often counterproductive. Data quality does not need to be perfect to provide valuable insights. In fact, the best way to drive data quality improvement is to analyze your data.
But it’s not just about differences between the data needs of process systems and analytics. The other reason you don’t need to wait is technology innovation. Technology has changed and we can do things today that we could only dream about just a short time ago. It is possible to do things more efficiently, in less time, and with better results. Why? Improvements in technology to support a transformational agenda and new lower cost deployment options are a key part of the answer. We live in a cloud-based and app-driven world. We have new and better multi-tenant applications. We have proven in-memory computing platforms. We have new architectural approaches for data management. New web-based intuitive user interfaces are driving deployment and user-acceptance to very high levels. New technologies and approaches are driving down the costs and time for implementation. There is good reason to expect “better, faster, and cheaper”.
One of the traditional challenges in delivering analytic applications is integration to transactional systems. Analytical systems use a dimensional-based modeling environment, while the transactional systems use a relational-based environment. Recent significant advances in technology and data-centric architectures enable enterprises to consolidate and share data in an analytics-first world. Data hubs share and deliver data in real-time and in a multi-dimensional way. New tools and approaches allow the storing and retrieval of large volumes of data without the need for pre-determined data structures.
We now have the capability to more easily integrate all forms of internal and external data. With speed and ease of retrieval, new data management tools have made possible the ability to perform analysis at a very granular level at affordable prices. We gain the ability to do frequent analyses and drill back to the transactional details while eliminating the need for a separate data warehouse to support basic analytics.
Is there any good reason not to analyze your data? No. Do not accept “not yet” or “only after.” Remember, it is only by using your data that you can wrangle out insights, determine how best to improve data quality, and become an analytics-driven organization.