Simplicity and Transparency
- 1. Draw No Boxes Simplicity and Transparency How to do Effective Data Warehousing and Business Intelligence David Holcomb, Ph.D.
- 2. Draw No Boxes Goals of this Presentation • Challenge the way you think – Words versus Actions – Principles to drive action • Offer some simple thoughts on DW & BI • Have some FUN!
- 3. Draw No Boxes “Simplicity” Defined Most people say “Simplicity” is . . . Making things easier! For this discussion, let’s use this definition of “Simplicity” . . . Don’t make it harder than it has to be!
- 4. Draw No Boxes “Transparency” Defined • Fewer common definitions of “Transparency” exist. From a data warehousing/business intelligence perspective, the one I use is: • The ability to take any number and break it down to the elementary or organic detail.
- 5. Draw No Boxes Simple DW/BI Model
- 6. Draw No Boxes The Model – Simply Stated
- 7. Draw No Boxes Principles • Get Data In – Data is an Asset – optimize your assets • Get Data Out – Information is Power . . . If available where and when needed, comprehensive, and transparent • Underlying Processes – Be an easy place to do business!
- 8. Draw No Boxes Data is _____ • Data is Dumb! – It doesn’t know where it came from – It doesn’t know what it allowed to do – It doesn’t know what it can be used for • Data is the artifact of a process – It doesn’t spontaneously generate – It doesn’t spontaneously combust – I know sometimes it does not feel that way! • So, Data is an Asset - Optimize those Assets
- 9. Draw No Boxes Data is an Asset • If we treat Data as an “Asset” – what does that mean? • Well, with other “Assets” we . . . – Acquire the asset • Determine you need it then go get it – Prepare the asset • Get it ready for action – Deploy the asset • Put it to work – Manage the asset • Maintain, improve, optimize, maybe even “sunset” the asset
- 10. Draw No Boxes Chair, Table, and Mabel Acquire the asset 1. Determine type and quantity 2. Buy chairs 3. Have them delivered 1. Determine type and size 2. Buy table 3. Have it delivered 1. Determine the skills desired 2. Interview 3. Hire Prepare the asset 1. Take it out of box 2. Clean it 1. Take it out of box 2. Clean it 1. Get office, PC Train on systems Deploy the asset Put in office or conference room Put in office or conference room Put in front of customers Manage the asset 1. Clean periodically 2. Fix or toss 1. Clean periodically 2. Fix or dispose 1. Review performance 2. Offer more training 3. Give helpful hint 4. If not working well, dismiss All of the assets are treated basically the same way. How about data?
- 11. Draw No Boxes Data Treated as an Asset Acquire the asset 1. Get it all • Data has Content and Context • It all has value, get it all Prepare the asset 1. May need some additional structures 2. Some ETL may be necessary; additional dimensional, aggregation, or temporal models may be needed Deploy the asset 1. Point tools at it 2. Give people direct access Manage the asset 1. Purge the data that is not of value 2. Add disk space to hold it all as it grows 3. Add more structures Interestingly if we use these principles, we actually know the requirements. Why do we keep asking?
- 12. Draw No Boxes Dilbert Break Does this sound familiar? Careful of Taking Orders!
- 13. Draw No Boxes Dilbert Break #2 A little knowledge – watch out! Requirements?
- 14. Draw No Boxes Why is this Important? • Every time, I talk to people regarding the main problem with their data warehouse, I always get the same answer: – It does not have the information I need • Why is this: – The DW doesn’t have all of the information – The people don’t know the information is there • Let’s assume you were the owner and these were your assets, what would you do?
- 15. Draw No Boxes Acting Like an Owner • You would . . . – Get all of the assets – Make sure: • It’s comprehensive with some sizzle (Product) • Everyone knew about it (Promotion) • It was accessible (Place) • It was at a good price (Price) Ok, so let’s do some Data Warehousing! Time for Action!
- 16. Draw No Boxes Action – Data Warehouse Native Formats Dimensional Model Aggregated / Temporal ModelsAcquisitionAbstraction Enterprise Data Warehouse Model CRM/OLTP Budgeting Compensation Web Logs Get the data in its native formats Audit Capability Full Transparency Benefits
- 17. Draw No Boxes Get Data Out Do you have problems with Dueling Numbers? The Secret: “He who is transparent first, WINS!”
- 18. Draw No Boxes Getting Transparency • Leverage the infrastructure from GDI • Utilize drill techniques (based on the business) – Organizational Design – Organizational Behavior • Understanding how organizations design themselves and how they behave, will enhance your transparency and adoption
- 19. Draw No Boxes Organizational Design • Organizations design themselves to get work done – Work Definition: Someone will do something at some time. • They design themselves basically 3 ways – Hierarchically – Functionally – Geographically • Other designs: – Teams: basically a microcosm of the macro organization – Matrix: Please note, everyone reports to someone and often the reporting with follow the hierarchy • So if we know this information, should not all reporting follow these designs?
- 20. Draw No Boxes Organizational Behavior • Rule #1: “The discussion will skip a level” – When the CEO looks a report, the discussion will not be about the results of their directs, but how their skip-levels affect the numbers • Rule #2: “If you can’t be first, don’t be last” – Competition between similar organizations is implied by reporting • Rule #3: “My kingdom, my tundra view” – Leaders want to be pointed to where they should look – You’re job is not to put out fires, your job is to point out the smoke
- 21. Draw No Boxes Putting it Together A Report (could be a dashboard) Levels drill to next level - Action plans - Questions - Real People, Real Numbers When you get to the bottom level, the user can reference the transactional system and verify numbers.
- 22. Draw No Boxes Underlying Processes • Metadata & Data Lineage – Meanings, data values, etc. for the fields and the path back to the data origins – The Get the Data In strategy gives you full lineage and maximizes the opportunity to have meta data and full transparency • Governance – Explicit Authority, Access Control, Standards & Policies. – Good governance should “Make the right thing the easy thing” • Interaction Models – The processes for maintaining the relationships between the data warehouse organization and two groups (end users and the data providers) – Make it easy for end users and data providers to work with the data warehouse – “Be an easy place to do business” • Organization – Often, no one owns the data warehouse. This lack of ownership leads to a lack of strategy and a disjointed warehouse environment. I propose that someone must own the data warehouse AND must realize they are working on behalf of the entire organization.
- 23. Draw No Boxes A Quick Re-cap