Bimotics views on Cloud Data Warehouse, Big Data, Business Intelligence, Analytics, Google Cloud Platform and Google BigQuery.
Monday, March 7, 2016
Is enterprise ready for Analytics Marts?
Observing consultants and IT organizations implement large business intelligence solutions, I find often that the first project either fails entirely or never gets past the initial phases. Lessons learned from these setbacks are often rooted in the client/ business not knowing the data they really want or political battles over which information and metrics are most important. Technical architects have avoided such battles by instead providing data marts, so that managers can help themselves to any data and build metrics in a self-service manner. By giving the client or business everything, the problem is solved.
But does this approach really help the business in the end? No. Managers remain misaligned serving their best guess of what will get them praise instead of addressing the true business need. Internal to the organization, data proliferation occurs where meanings get blurred and maintenance is so difficult even labels lose their original purpose of a sufficient description. Data silos of big data proportions are saved per division which is wasteful and duplicative.
Generally, I am describing an enterprise problem. Small and medium-sized businesses (SMB) suffer less from these political problems as they cannot afford much system customization and the resources needed to maintain it. Instead the SMBs tend to stick with the standard and best practice fields and data points. Because of this, Bimotics can provide our customers a solution we call the “gallery of analytics”. This gallery hosts all the business analytics available given the operational and financial application data marts for which the customer has data. Generally, these pre-built analytics reflect best practice operational and sales processes that are fundamental in all business looking to grow. The image below is an example of what the analytics gallery looks like.
The value to business owners is that they do not need to know what metrics they want before they bring on analytics. Instead they pick and choose the available analytics that makes sense to answer a particular business problem. They also can prioritize these analytics based on the business strategy they laid out. If the business changes direction, then the business owner can change out the analytics to reflect this new vision. Not having to go back to the drawing board saves precious time. This gallery approach to analytics puts the definition and the prioritization of metrics and at the end as well as provides breadth and flexibility to a manager.
Can this same principle be applied to solve an enterprise problem? Although very complex to build, can an “analytics mart” based on only the standard fields and best practice processes of major enterprise applications such as SAP, Oracle, Microsoft and SalesForce be built using the similar principles as our “gallery of analytics”? This “analytics mart” would cut across the different platform so that advanced metrics are available. For example, metrics, like support center effectiveness, are shown as a blend of financials in SAP with support data from Siebel. This proposed approach solves one of the greatest barriers that keep enterprises from successfully implementing business intelligence, by defining what to measure up-front, avoiding the interdepartmental politics in its allegiance to only standard and best practice processes.
Why haven't companies implemented "analytics marts" already? The answer is twofold. First building an analytic mart across enterprise systems houses many technical complexities especially when looking at all the software versions and system customizations that exist per enterprise. This is not to say that a solution is technically impossible however. Second, budgets split by division and departments need to be continuously spent in full which enable data silo behavior over cross department collaboration and process analysis. In other words, large organizations have budgetary policies that encourage managers to make blind purchasing decisions. How much of a fundamental shift would need to occur in an organization to embrace the sharing of data and metrics sharing?
The key to customer adoption of the analytics mart rests on a consolidated drive to improve your business and the willingness of managers collaborate holistically. Will enterprise managers have the courage to allow themselves to be measured against fundamental business process standards, in addition to evaluating how well their assigned divisions support the overall corporate strategy?