Douglas Briggs

Managing a Heterogeneous BI Landscape

Douglas R. Briggs Director, Business Intelligence Washington University in St. Louis

Who am I?

•  Education: –  BA, Williams College (go Ephs!) –  MS, University of Illinois at Urbana-Champaign

•  Business: –  Director of Business Intelligence at Washington University –  14+ years at Anheuser-Busch Inbev, managing BI and Data

Warehouse support, development, and architecture –  Managing Editor of Data Hub Australia (

http://datatalent.com.au/datahub-australia)

–  Advise companies throughout the Midwest on BI & DW matters

The place I love to work •  Washington University is:

–  Not The University of Washington –  Not George Washington University

•  Washington University in St. Louis is (US News & World Report 2014 rankings): –  #14 National University –  #22 Best (Graduate) Business School –  #18 Best Law Schools –  #6 Best Medical Schools: Research –  #5 Best Graduate School: Genetics/Genomics –  #3 Best Graduate School: Physical Therapy –  #2 Best Graduate School: Occupational Therapy –  #1 Best Graduate School: Social Work

•  Our mission: –  Washington University’s mission is to discover and disseminate knowledge, and protect the

freedom of inquiry through research, teaching, and learning.

… but enough about me, let’s talk about data …

Background and Ancient History

•  BI used to be done “inside the house” •  As cloud became more viable, more vendors supporting it,

more stuff went into the cloud, including BI! •  Some companies rushed into the cloud, resulting in –  Failure to deliver ROI –  Failure to satisfy user expectations –  Failure to “play nicely” with IT’s other systems

•  Cloud got a tarnished reputation, BI seen as risky for moving “out of the house”

•  Now, confusion!

Current State

•  Companies charting a course that accomplishes three goals: –  Plan: create a roadmap to a destination that is reachable given

the company’s culture, resources, and landscape –  Deliver: provide safe and robust service where it’s needed –  Earn: Maximize ROI

•  How do we do this? We need to know: where is BI now, and where is it going?

A Map of BI Tools & Uses

Mining & Exploration

Big Data

Dashboards Ad-hoc

Reporting &

Analysis OLAP &

OLTP Excel-based Reporting &

Analysis

Presentation Quality/"Pixel-

Perfect" Reporting

Narrower Wider

More General

Questions

More Focused

Questions

Deployment Footprint

The Natural Partition

•  Enterprise-wide reporting solutions: –  High-quality presentation –  Standardized, repeatable processes

•  Department-wide reporting –  Intermediate-scope business questions –  Multiple delivery channels –  Parameterization, customized look & feel

•  Specialized data tools –  Data mining, powerful analytics, data discovery –  Used by enclaves of power users who know the data intimately

Implications

•  Vastly different tools for different audiences!

•  Single tools can be unsuitable for use across the segments

•  You can refocus your landscape on best-in-class solutions for each segment

•  Some vendors have multiple products in their suite that can work in each segment, but don’t discount niche players focused on specialized needs

Your Challenge

•  Remember your goals: –  Create the viable roadmap that gets you from here to there –  Deliver the goods –  Make it pay for itself

•  You need the appropriate segmentation between components that provide the best facility and agility –  Facility = effective feature-set given your institutional needs –  Agility = robust, extensible solution that meets present and future

needs

Goals for delivery: safe, robust, accessible

•  Safety means all of these things (and more): –  Regulatory compliance –  Data & information protection (including models & algorithms) –  Business process resumption & disaster recovery

•  Data governance can help facilitate this •  Credibility is the first casualty if it’s inconsistent or

unscalable •  Delivery at the user-interaction endpoint is tailored

Managing the Transformation - Technology

•  Rip-out & Replace: –  Focus change and time-box the impact –  Costly, and highest risk for “institutional inertia” to cause delays &

overruns

•  Phased: –  Minimize risk and pivot quickly –  Landscape is constantly in flux –  Requires thoughtful landscape architecture & effective

communication to foster credibility!

Managing the Transformation - Organizational

•  Roles: –  Development & support teams –  May require changes to staff, retraining, or reorganization of

leadership

•  Processes: –  Governance, architecture, & business alignment –  Grow it over time, especially if you’re phasing in BI components –  Vendors & integration partners are excellent resources for best-

practices

Make it Earn its Keep

•  Find your partners and advocates

•  Select targeted opportunities: –  Short ROI horizons –  Cheap investments with outsized returns

•  Handcuff the organization to the results –  Nothing motivates like skin in the game –  Vendors too!

Strategies

•  Decouple segment-based solutions from each other

•  Leverage SaaS paradigms selectively

•  Allows departments to build effective point-solutions

•  Your landscape will (probably) become heterogeneous organically – that’s okay.

Change is coming …

The vendor offering landscape is (once again) gravitating toward consolidated solutions

•  OLAP à OLTP •  Increasingly sophisticated, presentation-style graphics to

merge traditional dashboards and pixel-perfect reporting •  Hardware acceleration and advanced storage & retrieval is

allowing traditional data discovery and exploration tools to encroach into the Big Data space

Your next steps

•  Understand alignment between your current landscape and the segmentation map –  Talk to your BI architects, they will show you the fit/gap –  Close gaps with effective tool choices & process changes

•  Develop the strategic plan, but … •  Make sure you seed the path (especially in the beginning)

with the projects that meet your goals: –  Part of a feasible plan –  Deliver the goods –  Pay for themselves

Take-aways

•  Natural user & use-case partition à landscape segmentation

•  Best-in-class is a more effective strategy than it’s been for the past several years

•  Expect some solutions to migrate to SaaS organically

•  Expect tool consolidation to muddy the waters

Questions?

(I have some answers)

Please keep in touch!

•  My interests and specialties: –  BI competency centers –  Data governance –  BI change management

•  Contact: –  [email protected] –  www.linkedin.com/douglasrbriggs

Thank you and good luck!

