BI on a Budget
Yes! You can deliver cost-effective business intelligence.
Clarise Doval Santos is CTO
of InterActive Systems & Consulting and a senior project manager
for Kaiser Permanente.
Her career encompasses
all aspects of
designing,
implementing,
and managing
enterprise solutions—
including
the development
and engineering
of decision
support systems, business
intelligence (BI) suites,
Online Analytical Processing
(OLAP) tools, portals and
dashboards, datamarts, and
data warehouses. Here,
Doval Santos shares her
10-point strategy for
campuswide BI that won’t
break the budget.
EDITOR’S NOTE: Doval Santos
will co-present “The Economics of
BI: How to Drive Cost-Effective
Implementation Strategies” at
Campus Technology 2007 in
Washington, DC, July 30-Aug. 2.
Want to be considered for Campus Technology's Top 10? Send your countdown and a brief background/bio summary to [email protected]
- 10
-
Ask yourself: What is BI, anyway?
- Is it analysis? Reporting? Data mining? Predictions?
- The answer is, it doesn’t really matter how the rest of the world defines BI—it’s how you and your users define it that counts.
- 9
-
Think: Architecture, architecture, architecture.
- In real estate, architecture means location; for BI, it’s all about aligning data
flow to business processes and enterprise data models.
- Match up information architecture, business architecture, and program architecture.
- 8
-
Use rapid prototyping as you try to understand users’ needs.
- Build your game plan while helping users understand what is possible.
- Use tools like Marvelit Dash and Palo (an Excel OLAP plug-in) to deliver results in as little as a week.
- 7
-
Don’t bite off more than you can chew.
- Break your overall data warehousing/BI program into a series of small
projects that will provide quick wins.
- Stay aware of your process: Discover, define, design, develop, deploy,
determine the next steps, and test along the way. Repeat.
- 6
-
Exercise rigorous project management.
- Many think that modern application development methods are antithetical
to "traditional" project management. Not true! Each needs the other to
achieve business advantages rapidly and without "scope creep."
- 5
-
Involve your users every step of the way, and in a timely manner.
- If projects take too long, lose relevancy, don’t deliver on current needs,
don’t fit well with business and IT procedures and standards, and/or are too
hard to learn, they will not lead to useful applications.
- Remember: The least cost-effective project is the one that never gets used.
- 4
-
Get buy-in from all organizational leaders who have reporting
and analysis needs.
- Sustainable BI projects don’t exist in a vacuum; they require input from all
stakeholders.
- Your project must be part of an overall enterprise data strategy.
- 3
-
Data stewardship matters!
- The quality of source data is very important. It all goes back to GIGO
(garbage in, garbage out).
- Assign a data steward to work both with the users and the source system
experts, to verify data integrity.
- 2
-
Consider open source software as seriously as you would
proprietary products.
- Open source tools and BI suites like Pentaho (www.pentaho.com), Jasper-Soft, and SpagoBI have
matured into true competitors of the proprietary vendors.
- 1
-
Keep your BI strategy up front, constantly.
- Have an overarching plan with both strategic and tactical tracks.
- As each project within the tactical track is completed, iterate the strategy
with lessons learned.
- You’ll see how your priority of cost-effectiveness is served by your strategy.