Yesterday I was talking to one of the wicked smart software engineers I know—Dave Gustafson (who also plays a killer trombone, BTW). Dave was talking about techniques for smart, pragmatic refactoring of code, and I couldn’t help but see the parallels to information governance and data quality.
One method Dave was mentioning was code smell. Yes, code smell. “What’s that?”, I said. Dave explained that it’s like refrigerator smell…you know when you open the refrigerator and you smell something, but you don’t know quite what it is or where it is? But you sure know that something is awry.
Oooooo. Isn’t this how many data quality and information governance initiative start? Something has gone wrong. Maybe you weren’t able to ship product…maybe you have too many Days Sales Outstanding in Accounts Receivable… You have information smell. Now you have to figure out what exactly is wrong and how to fix it. If you’re really good, you’re also thinking about how to avoid the problem in the future.
Just this week, I talked to a customer who was setting up an information governance program. “Why?”, I said. He explained that their charter, driven by IT management, was to take data to the next level. “Why?”, I said. He explained further, saying that they needed to be more proactive in reporting than so reactive…additional data requests were taking far too long to satisfy. “But WHY did you decide this was something you needed to do?”, I asked again. This question was easy to answer for him. They initiated the project because they weren’t able to ship product to key customers from a certain plant. (That’s the smell.)
- Identify where the smell is coming from. This identification requires you to understand the business process that went awry, and to then further research the process to identify the critical data elements. Oh, and WRITE IT DOWN. Now you have started both a data inventory (including metadata like criticality and associated business process) and data policies. (SAP BusinessObjects Information Steward can help with this process.)
- Quantify how much the smell is costing you, and how much benefit you’d achieve from managing the offending information elements more holistically. Given that you started with a smelly business process, identifying the cost of that business process failure with the stakeholders from business should be relatively straightforward. Plus, these basic definitions of value will help you build metrics that show the success of your program. You’re still writing these down, right? And documenting consensus? Of course you are.
- Establish a current baseline for those critical elements. How much management of these critical elements is enough to achieve optimum value? Perfect data is many times not pragmatic, but it’s hard to know how far to go when you don’t know where you ARE. The current baseline also drives more accurate project estimates. (SAP BusinessObjects Information Steward can help with this process.)
- Create an improvement plan. This plan lays out steps from getting from here (current baseline) to there (efficient business process). Many times, companies use technology to help them scale this effort, like SAP BusinessObjects Data Services, SAP NetWeaver Master Data Management, SAP NetWeaver Business Process Management, SAP Extended Enterprise Content Management, SAP NetWeaver Information Lifecycle Management, or SAP Master Data Governance.
- Align Business and IT behind enterprise data standards, policies, and processes. Make sure to include data ownership, too. –WRITE IT DOWN— Many times the critical elements for one business process also impact another business process, and there’s absolutely no reason you should start from scratch when another whiff of information smell wafts in your direction.
- Define and control the full data lifecycle (definition to destruction). As you discover the data policies/baselines/etc. we talked about above, you are bound to also uncover create, read, update, and delete (CRUD) policies around those data elements. Again, write those things down as they come up, and you’ll have a start at documenting CRUD policies.
- Quantify results. With the baseline and quantified value stated above, you’ve nearly already designed your dashboard for your key stakeholders. Make sure you use some of your metrics, and some glowing testimonials about your sweet-smelling business process to publicize your success. (SAP BusinessObjects Information Steward can help with this process.)
- Identify what you can do to stop errors in the future. Can you establish a quality firewall at point of entry, for example? Can you check for duplicates before bad information enters your system. Many times, you can leverage the exact same transformation rules you used to fix the information in this project to establish these point-of-entry policies. (SAP BusinessObjects Data Services, SAP NetWeaver Master Data Management, and SAP Master Data Governance can help with this process.)
One of our great Global Data Managers likes to say that 90% of the problems in her company are master data problems. At least, when a business process starts to smell, that’s where people point the finger when the refrigerator door opens. Information governance practitioners not only dig through the refrigerator to track down the true source of the smell, but follow the key information governance best practices outlined above.
With this process, you’ve refactored your information foundation by leveraging techniques Dave uses to refactor code. Sweet!