The University of Central Florida Business Incubation Program hosts CIO Community of Practice meetings, which provides a Big-Enterprise view of leading technology focus areas. CIOs and key IT representatives from leading Central Florida companies including Darden Restaurants, Harris Corporation, HD Supply, Siemens PG, The Boeing Company, and Walt Disney World Company met this week to discuss Big Data and analytics.
The discussion centered on how to transform fragmented data architecture into an environment where business decision-makers have consistent, reliable information access and can make better informed business decisions. Participants described their organization’s use cases, challenges, tools, roadmaps, and strategies. If you are an enterprise data management or enterprise data architecture practitioner, you are probably not surprised to hear fundamental blocking-and-tackling disciplines dominated the conversation. Participants described how master data management, data governance, data lineage, and data staging fundaments are pre-requisites before to turning data into information, scaling business intelligence delivery, and achieving the insights required to gain competitive advantage (see Figure 1).
Figure 1: Data Cycle
Participants described how their firms could achieve competitive advantage by:
- Obtaining a 360 degree behavioral view of customer and organization’s contact
- Operating as a single company across multiple divisions and product groups
- Understanding how product demand mix drives revenue and profitability
- Correlating current activities and predicting future events
- Performing real-time A/B testing of sales, marketing, and logistics strategies
- Completing data integration projects within 90-100 days
When describing the macro-challenges preventing widespread intelligence delivery, distorting human insight, and inhibiting competitive advantage, each attending company representative expressed a similar organizational context:
- Uneven data management maturity across the organization
- Emerging master data management practices
- Minimal identification of single source of truth
- Little agreement on core data entity representation
- Enterprise Information sharing platform not in place
- Fragmented data silos and data repositories
- Ad hoc, project-level data integration
- Limited data virtualization and data services
- Proliferation of Excel spreadsheets
The result, 90% of business decision makers believe they do not have the information required to make informed business decisions. Vlad Rak, Vice President Global Technology Strategy at The Walt Disney Company, described how to craft an architecture transformation program within the unique organizational context. The data architecture transformation program roadmap includes:
- Define analytic capabilities (i.e. KPIs, 5 year strategy)
- Build an information management capability model
- Assess organizational maturity
- Define a business case
- Create reference architecture
- Establish business-driven data governance
- Collect and bundle projects into an architecture transformation program
When defining analytic capabilities, consider posing a foundational question; what actions can the data drive? For example, can the business use sensor information and drive new business offerings. A business-driven data governance theme couples information management and analytics with business performance.
A business case helps energize interest in data, and turns data consumers into data stewards. When defining a business case, link information management to revenue growth and/or cost efficiency. Within Big-Enterprise organizations, defining the convergence between micro-projects and macro-programs is critically important. I often describe how trust is required to successfully share web services and increase Service Oriented Architecture adoption. Community of practice participants mentioned trust is required to successfully share data. When an organization does not have a consistent data governance strategy, verified data lineage, and a single data integration platform, up to 80% of project time is spent verifying data integrity and massaging data. The team spends only 20% of project time analyzing data. The ratio is unfortunate, because the analysis effort drives intelligence delivery and human insight.
Professor Morgan Wang described some insights gleaned from analyzing health care insurance data and measuring how the American Affordable HealthCare Act would impact insurance premiums and available health care dollars in the State of Florida. Professor Wang mentioned carefully adding more data attributes and analysis techniques into the analysis model could yield more accurate insight. He also mentioned competitive value lies in creative analysis rather than raw record count. Professor Wang mentioned Big Data (i.e. petabytes and terabytes) is not always required to gain competitive advantage or transform the world.
When discussing SAP HANA, participant discussion covered how filtering, reducing, and caching reduces datasets and right-sizes processing, network bandwidth, and storage. While many main traditional analytics vendors were mentioned (i.e. Business Objects, Cognos, SAS, Informatica, Information Builders, MicroStrategies Oracle, SAP, and Terradata), very little time was spent talking about specific analytical tools, and participants outlined how blending people, process, with technology tools is a critical success factor.
Surprisingly, the discussion mostly revolved around analytics, and participants expressed very little pragmatic, production experience with Big Data tooling. Specific Big Data tools offered by innovative vendors (e.g. Cloudera, HortonWorks 10Gen, DataStax, Splunk, Pervasive Software, and WSO2) were not even mentioned during the half-day session.
After reviewing my data (notes from the conference), the following insights stand out:
- Master data management and agreement on core data entities is critical to bridging silos and gaining synergy across divisions and departments.
- New business models are possible when aggregating sensor data, yet new data reference architecture components are required
- Fund transformation programs by aggregating multiple data projects and reviewing historical baseline spend
- Overcoming traditional Big-Enterprise adoption hurdles and maturity barriers requires a ‘business-driven’ approach. Up-front time is required to establish business impact and buy-in.
- To optimize success and adoption, map self-service Business Intelligence tools to end-user skills
- Predictive analytics requires a continual discovery period to identify analysis patterns
I look forward to attending future UCF CIO Community of Practice meetings, and would like to extend my appreciation to the coordinator, Robert Rich.