Pages

Monday, January 17, 2011

ketika anak kuliahan mencoba memahami pertanyaan penting tentang Business Intelligence

Taken from here.

1. What happened and what is the cause?  
Find technology solutions that provide ways to identify the root cause of issues. This may include filtering or drilling into the drivers behind a specific outcome.

2. What led up to this?
By viewing statistical trend lines, you can determine whether a specific issue happened over time or abruptly. Only with trending information will you be able to accurately assess a problem.

3. If I do nothing what is likely to happen? 
Use trend lines with forecast projections to show future outcomes.

4. Is this a systemic problem or an abnormality? 
Use control charts to show how much change there is from one result to the next – for example, in quality or even sales cycle time – to see if you have a consistent process.  A lot of variance indicates a systemic issue.

5. What happened the same time last year, last quarter, last month? 
Look at the differences in the results achieved the same time last year, last quarter, last month, etc., to see if you are tracking similarly or if there is a difference indicating some change in seasonality or the process itself.

6. How does this compare to a benchmark or best practice? 
Review the difference between results being achieved in one part of your business to another, perhaps a best practice within your organization. You can also use external benchmarks to determine areas requiring improvement.

7. What are the contributors to my issue?
Here, Pareto charts can show you the percentage contribution of each result (great for applying the 80/20 rule.) Or, histograms can show the major “buckets” such as Accounts Receivable aging.

8. What would happen if I made this change? 
The beauty of truly interactive dashboards means that you can now do real-time scenario analysis that shows the result of one or more changes.

Beautiful. I think the business intelligence will more fascinate me in the near future..

No comments:

Post a Comment