Lampa, who is also a managing partner at data management consultancy Archipelago Information Strategies, presented strategies for the shifting analytics landscape Wednesday in a presentation entitled “Support Your Local HiPPO: Analytics for Better Business Outcomes.”
While it may be of little surprise that Lampa suggests IT and business work together, he goes a step further and puts “business optimization analytics” at the core of the modern enterprise analytics program. With business optimization as the beginning “target,” analytic team members can be more nimble with changes to data models and sources, as well as the expectations floating around them. Surrounding this business optimization core is a closed loop starting with business process discovery, and on to statistical analysis and data warehousing and BI, and then collaborative BI and workflow routing and alerts that direct back into process discovery.
This loop, built on business data requests and adjustments, creates the basis of what Lampa coined as “agile on steroids.”
“Analytics projects are highly nuanced and follow the way humans think and decide, and that doesn’t follow a prescribed, structured approach,” he said. “We start with one premise and then we’ll find a tangent we need to chase down because it’s proving to have a lot of reward. The rule book is rewritten every day.”
And the most massive rewrites and edits to that “rule book” tend to come from the size and variety of the data itself. Businesses are dealing with higher volumes of data from more and more external data sources, each with their own quality questions, ownership issues and unstructured needs. Lampa said the direction for analytics teams on these outside data streams can be found in initial business process definitions. On the more technical side, there should be a tight connection between analytic systems and ERP, including regular exchanges of alerts and changes, in addition to monitoring the data quality of external data feeder sources. Analytics pros must also find their own answers about their desired amount of data quality and investigation, especially with pools of data from sources such as public social networks.
To reach more concrete and revelatory analytic discoveries with new and existing data, Lampa said part of the preparation is becoming open to uncertainty.
“It’s a chaos that has to be embraced and institutionalized,” he said.
Ideally, the feedback loop and preset understanding of the changing nature of data should bring more comfort on analytics for the “HiPPO” – i.e. the boss who will be making the decisions based on this data. To get to this point from a team-building perspective, Lampa said having a Chief Analytics Officer or an analytics evangelist anointed from your IT or data team works to link the business basis with the down-and-dirty data. Expanding across the enterprise, Lampa advised that this analytic planning be linked to BI centers of excellence and the stated role of “collaboration architects.” who invoke analytics activities with those on the process side of the house.
Analytics programs should be couched in terms that allow for adjustments to models and even a reset, if needed, before reports are cranked out. Not an excuse for a program that drags on, Lampa said this approach frames the time and effort it takes to find and refine the most accurate data that is increasingly undergoing a “paradigm shift.”