Does On-Demand Business Intelligence Make Sense? April 2, 2009 David M. Raab Raab Associates Inc. www.raabssociatesinc.com
May 13, 2015
Does On-Demand Business Intelligence Make Sense?
April 2, 2009David M. RaabRaab Associates Inc.www.raabssociatesinc.com
On-Demand ComputingDefinition of On-Demand
• Remote resources• Self-provisioned• User-operated
Comparison
In-house On-Demand Outsource
local resources remote resources remote resources
IT-provisioned self-provisioned vendor-provisioned
user-operated user-operated vendor-operated
Advantages of On-Demand
• Low cost / no fixed investment
• Easy to expand (or reduce) capacity
• Easy to use / no tech skills
• Quick deployment / easy to change
Business Intelligence
“Information used for decisions” • Standard reports • Dashboards• Ad hoc exploration• Operational / predictive /
distributed BI
BI Components• Data warehouse (source,
integrate, structure)• Presentation (BI system)…most of the work is in the
warehouse
An executive dashboard.
A traditional BI architecture diagram.
Does On-Demand BI Make Sense?
On-demand is good.Business Intelligence is good.
…so, they must be good together, right?
(Remember peanut butter and chocolate.)
(But what about broccoli and ice cream?)
Business Intelligence Challenges
• Need tech skills to set up
• Requires IT support
• Costly to deploy
• Time-consuming to implement
• May be short-term or one-time project
• May require large computing resources
BI Challenges vs. On-Demand Advantages
BI Challenges On-Demand AdvantagesNeed tech skills to set up Easy to use / no tech skills
Requires IT support Easy to use / no tech skills
Costly to deploy Low cost / no fixed investment
Time-consuming to implement Quick deployment / easy to change
May be short-term or one-time project
Easy to expand (or reduce) capacity
May require large computing resources
Easy to expand (or reduce) capacity
Does On-Demand BI Make Sense?
On-demand is good.Business Intelligence is good.On-demand strengths match BI challenges.
…but what about objections to on-demand?
Objections to On-Demand
• Security
• Process integration
• Reliability
• Vendor lock-in
• Inadequate tools
Responses to On-Demand Objections
• Security: often better than in-house
• Process integration: APIs more available
• Reliability: most BI projects are not mission critical
• Vendor lock-in: negotiable
• Inadequate tools: pick the right vendor
Does On-Demand BI Make Sense?
On-demand is good.
Business Intelligence is good.
On-demand strengths match BI challenges.
Objections to on-demand can be met.
…but the real issue with BI is complexity.
Complexity in BI
• Complex data preparation
• Complex data models
• Complex requirements
• Complex analytics
Does On-Demand BI Make Sense?
On-demand is good.Business Intelligence is good.On-demand strengths match BI challenges.Objections to on-demand can be met.
Can it reduce complexity?
Strategies to Reduce Complexity1. Make standard technology
more efficient• Simplify the tasks• Automate the tasks• Do fewer tasks
2. Use non-standard technology• Columnar and in-memory
databases
Strategies vs. Types of Complexity
Data Preparation
Data Models
Requirements Analytics
Simplify x x x
Automate x x x x
Do less x x x
Database x x
Does On-Demand BI Make Sense?
On-demand is good.
Business Intelligence is good.
On-demand strengths match BI challenges.
Objections to on-demand can be met.
It can reduce complexity.
Is there any problem on-demand BI can’t solve?
Problems On-Demand BI Can’t Solve
• Data isn’t available
• Too costly to move data
• Project has no clear objective
…do these in-house or outsource.
When Does On-Demand BI Make Sense?
In-houseOn-
Demand Outsource
General on-demand advantages
• In-house tech skills lacking x x
• In-house tech staff unavailable x x
• In-house cost too high x x
• In-house time too long x x
• Short term / one-time project x $
• Large hardware requirements x x
General on-demand issues
• Security risk x x x
• Process integration x x* x
• Reliability x x* x
• Vendor lock-in x x x
• Inadequate tools x x* x*depends on tool
When Does On-Demand BI Make Sense?
In-houseOn-
Demand Outsource
On-Demand BI issues
• Complex data preparation x x* x
• Complex data models x* x
• Complex requirements x* x
• Complex analytics x* x*depends on tool
Issues on-demand BI can’t solve
• Data not available x ? x
• Too costly to move data x x
• Project lacks clear objective ? ? ?
Picking the Right Vendor
• Technology– Provisioning– Functionality– Scalability– Complexity– API / integration– Maturity
• Terms of service– Data access– Service levels– Intellectual property
• Professional services– Strategy – Project management– Data integration– Analytics
• Pricing– Upfront and ongoing cost– Impact of volume– Contract period
• Financial stability
Final Thoughts
1. On-demand BI offers substantial benefits
– Cost, time to market, flexibility
2. Vendors differ substantially
– Cost, time to market, flexibility
3. It can’t solve all your problems
4. Choose projects and vendors carefully for success