Analytics that just work

Analytics, business intelligence, and their predecessors have been full of promise since the evolution of the personal computer. And for the most part, that’s where they’ve remained: full of promise.

By now, most organizations have a decade’s worth of historical data, much of it secured in sophisticated data warehouses. But few have transformed that data from bits into insight. In storage, the data is effectively useless. With the exception of fairly narrow financial reporting, the gems and nuggets promised to business users have remained elusive.

In our experience, there are three components of a successful Business Intelligence (BI) solution:

  • Tight integration with databases and IT systems,
  • carefully and cleverly developed analytic algorithms, and
  • quick and intuitive business interfaces.

If any of these are missing, the BI implementation will fail.

The BI solution needs to be reliable. If the data is incomplete, out of date, or simply wrong, business users will not trust it, and they will not use it. Once a system has gained a reputation of being untrustworthy, its usefulness is sacrificed. Tight, thorough integration with IT systems and databases mitigates this risk. A company must invest the time to ensure that their data is accurate, well-understood, and accessible to analytical tools. All too often, this crucial step is rushed or bypassed entirely.

The heart of any effective BI solution is robust analytics. The metrics and measures must be tailored to the specific business questions that your organization wants answered – not what the average organization needs. This requires a significant investment of time, and precludes a quick out-of-the-box integration. Your business needs must be fully understood before your measures and metrics are developed and translated into the mathematical language of analytics. Only then will the solution answer your specific questions.

Even if your tools are accurate, trustworthy, and answer your business questions, they still must pass the performance test. In an ideal world, this means no learning curve and no latency. A manager should be able to open the software and instantly drill down to the answer. Intuitive design and optimized performance lead to a better user experience and a more useful BI solution.