Blog: William McKnight

« Data warehousing dead? | Main | Scott Humphrey Pacific Northwest BI Summit »

Providing ROI with Business Intelligence

Below are some learnings and tips on BI/DW ROI from Neil Freake, Senior Manager of BI/DW at Scotiabank, a client who I had the privliege of presenting with yesterday at the Shared Insights show. The presentation was titled "Providing ROI with Business Intelligence." This was an additional slide so it's here, as promised, for the attendees... and all of you.

- Financial Models need to factor in attrition rates
- Calculating “lift” is often not a viable argument (wholly dependent on assumptions)
- Benefit is based on Profit NOT Revenue (ergo…understand margin calculations)
- Financial Models must be more accountable when defining costs (i.e., SOX)
- First Run Cost Estimates are always off +/- 25%
- Payback Calculation – two camps
- Cost Camp: begin to calculate as soon as you incur costs through cumulative life of benefits
- Benefit Camp: Calculate payback after project has been implemented (Not recommended or realistic)
- Soft Dollar benefits are easier to argue / more difficult to prove
- Hard Dollar benefits are rare, more difficult to argue and almost impossible to prove.
- Most organizations use a 10% interest rate when calculating NPV
- How well you define your assumptions will dictate how well your model stands up to financial scrutiny
- Every financial model has weaknesses (highly subjective)
- Would have to be in a lab to prove benefits
- Double dipping benefits – most common error in business cases (who has the right to claim a benefit?)
- ROI is purely an internal (and sometimes external) marketing tool
- The higher the dollar amount the more scrutiny your project will face

Share: