When you played baseball as a youngster, and stepped into the batter’s box, the last thing you wanted to be was an “easy out”. Ironically, today many healthcare organizations are looking for the “easy out” to rapidly develop the business intelligence reporting needed to address regulatory reporting demands, population health 
management and chronic condition management, to name just a few.
The pressure to quickly stand-up an enterprise data warehouse, put data governance in place, start loading and cleaning data is intense just to get to the point of creating dashboards and offering mobile BI. Overloaded Healthcare IT teams are dealing with demands to compress traditional time-frames of 18-24 months to get the BI foundation in place down to as little as 4-5 months, start to finish.
The Future of Big Data
With some guidance, you can craft a data platform that is right for your organization’s needs and gets the most return from your data capital.
This situation begs the old saying of “do you want it fast or do you want it right?” You can bet the answer today is both. Generally, healthcare organizations develop a BI strategy that examines the current state BI architecture, envision a future state BI architecture, document the gaps and create a time phased roadmap to build out the infrastructure, software and development required to meet the business needs. Just describing the process tells us that it will be complex and time consuming, right?
It is time to examine the traditional waterfall development approach to building enterprise data warehouses necessary for addressing the current demands for business intelligence in healthcare. While we could dialog on the use of Agile versus waterfall for development, I want to propose a more radical approach: a business intelligence assembly line process. We need to reorganize our thinking towards automating the steps to deliver raw materials quickly, i.e. identifying a data source, gauging its quality and delivering it for assembly, before adding it to the BI visualization process. The data normalization and data quality can be addressed with another manufacturing idea: continuous improvement over time. Instead of traditional thinking of having to move all data into a common data model, this approach allows for data sources (raw materials) to be substituted, if needed, in the assembly line process for business intelligence reporting over time.
The people, process and technologies need to be aligned to the assembly line process for delivering high-quality products (dashboards and reports) faster and faster from re-usable components in the process. Engineering a new product would start with a design goal and identify how it would be built as opposed to only delivering what the underlying data warehouse can deliver. This assembly line approach provides the flexibility to introduce new concepts like big data or streaming real-time data as manufacturing techniques to speed up assembly line delivery without the big bottleneck of forcing everything through a common data model or wait for assimilation into the enterprise data warehouse.
Trusted data sources, like trusted raw material vendors, will be the secret to rapid BI manufacturing success. Do you like the idea of a business intelligence assembly line versus standing up a whole factory? Let’s poke holes in the idea to see if it is a better alternative!
