In my last post, Rethinking Business Intelligence, I explored how organizations must learn to ask the right questions rather than simply ask for data as a starting point for analytics. There are a number of forces impacting the organizations that deliver analytics services to the business, and how these organizations can reimagine their relationship with analytics to address the critical need for speed in delivering information and insight. These forces include:
- The plummeting cost of data storage, which has fed massive growth in the volume of data in federated, unstructured systems, and is producing data of exceptional diversity, but unknown value
- Cloud technologies have become ubiquitous and acceptable even in industries that have favored security over access
- For organizations with the right skills, open source applications have the power to find patterns and trends in seemingly random bits of data
- The proliferation of sensors has accelerated the capture of real-world data—information generated in a physical environment like bio-metrics and food consumption—and brought more of the physical world into a digital ecosystem, which is transforming how services are delivered and experienced by users
As traditional relationships between clients and service providers have blurred, the race is on to see who is fastest at accessing and exploiting new information sources for commercial gain or to solve business issues (see Figures 1 and 2 to see how the traditional and new approaches compare).
The window to realize economic value from analytics is getting smaller with acute and polarized business challenges, and shorter business cycles. It took AirBnB—a company that helps people list, find, and rent lodgings—less than 12 months to go from concept to live operations in 2012. It now has 1.5 million listings in 190 countries. Success lies in finding accurate answers to key business questions at speed.
As the potential to exploit data has increased, so has interest across functions outside the traditional analytics hotspots of marketing and risk. Strategy, sales, distribution, collections, HR, customer service, and IT are all looking to business intelligence to improve performance.
Progress has, however, been hampered by a continuing lack of skills and experience across all levels and functions, and especially in information sciences, in statistics, modelling, and interpretation. Old data-accessibility issues—from infrastructure constraints to data privacy—remain problems. Delays in major technology programs have resulted in fragmented, incomplete solutions and the need for workarounds. Likewise, regulatory controls and interventions are increasing in many industries, and shareholder pressure for growth and profitability further complicate the process of gaining insight.
We believe there are three key building blocks to creating an analytics capability that can withstand the forces outlined above and deliver the outcomes shareholders desire. Here we will explore the first one.
Reimagine the approach to sourcing and using data
Data governance is hard to implement because it focuses on enforcing data quality in source systems, which is hard to do quickly. Combining technology, process knowledge, and analytics creates intelligent operations that allow organizations to extract and govern only the data needed to drive the insights business users can trust, without modifying source-system data. This approach embeds the best principles of data governance to provide inbuilt data lineage (information on the data's origins and life cycle) so business units can rapidly and interactively prioritize the inclusion of new elements into a model or data set—for example, adding weather data to a model predicting delivery times for major infrastructure projects.
Fig 1. A traditional approach seeks to integrate and cleanse all or significant data before it is used for business intelligence
Fig 2. With an intelligent operations approach, only the information needed for a specific business purpose is used, without requiring changes to the source data
Information supply chains are built with embedded governance that establishes data lineage, a data dictionary, and metadata management protocols.
To reimagine how data is sourced and used, begin by defining the data required for an intended purpose and the associated governance requirements. It is then possible to build as many information supply chains as required. Finally, build in organic data governance with a consistent operational model for documenting, prioritizing, resolving, and communicating data issues.
Critically, this approach runs in harmony with existing data sourcing and management initiatives as logic and code can be transferred into the operational system when the resource is available to do so.
If properly understood and leveraged, the proliferation of data, cheap storage, and sensors can be harnessed through advanced analytics methods to help businesses reimagine how to embed deep insights into increasingly high-speed operations.