Taxonomy: It sits in the critical path of …
June 9, 2008
Where is the excitement around this issue?
It seems that “Taxonomy” was my word for the week. This is my third post about it within 7 days. It’s not that I am in love with the word, rather, it’s just pretty darn important! With any big initiative, the first thing we look to is a solid foundation for communication. Think about it, we usually address taxonomy anywhere from casual discussions to formal governance groups for many initiatives – dare I say any initiative that strives to bring real change to an organization begins with taxonomy (either consciously or subconsciously). Thinking it through, here are some of my top-of-mind game changers that require a solid taxonomy:
• Master data management. By definition this is really an enterprise taxonomy that is the official reference data for an organization.
• Metadata management. Tagging data with information is best performed only after a taxonomy is well-established. Else, with-what-shall-I-tag-it plagues the process.
• Business Intelligence. Without a proper taxonomy, how do we bring together people from diverse business perspectives together to understand data from a central and enterprise viewpoint?
• Data Management. Well, of course we can’t properly manage data without knowing where things are in a hierarchy or what context the data should exist in.
• Data Quality. Here we are really measuring data against the taxonomy; whether implicit or explicit.
• Governance. Strategic decisions are made for specific purposes and they need to rely and depend upon a socialized and accepted taxonomy.
• Data Stewardship. This is the process of holding someone accountable for making tactical decisions to implement strategic direction.
• BPM. When we look to manage business processes, they depend upon real information. So, having a taxonomy to base these data points is crucial.
• SOA. Reusing software components and exposing them at the enterprise level demands a highly accepted understanding of the organizations data. Sure, this view is exposed as a group of web services that are published in a repository that is self aware – but without a canonical data model as your underlying foundation, consistency is not reached. A canonical data model is highly correlated to a mature taxonomy.
• Strategy, Solutions and Architecture. It’s near impossible to calibrate these three without a friction free flow of communication. Let’s not talk about what should’a, could’a and would’a – but let’s focus on the business problem at hand. Having a living taxonomy that is socialized, accepted and part of our DNA is key to gaining quick momentum as we put distance between us and our competition.
These are just some quickies that I bubbled up. What other initiatives need a solid taxonomy? Thinking about taxonomy, when you look to bring real change to an organization, what happens? From my experience, there are two choices:
1. Address taxonomy early and often. Realize that there are some things that are so important that we need to establish, socialize and enforce them.
-OR-
2. Jump to build a solution. Then realizing there are terms misaligned, misused and duplicated, go back and either fix the data models (and all subsequent diagrams and code – this rarely gets done) or create a lot of code to hide these issues. When we do this, we establish a short term brittle foundation that breaks when the next change comes or we end up with a bunch of custom spaghetti that tries to tie things together but really ends with just a lot of confusion.
Bottom line
• Embrace taxonomy within your natural collaboration style. When something is unclear, pause, ask, record, check for understanding, agree on the outcome and move forward. It’s not a development phase, don’t sell it to senior management. No one cares about it. It’s an expected minimum of doing business. Add it to your culture’s DNA.
• Don’t underestimate the issues when terms are not aligned. It wrecks havoc to your foundational infrastructure and the costs (both hard and soft) can be big. Know it is there and plan for it.
• Scope your risk. If you are working within a group or team, the risk is small. Plan for it and cross it off your list as you develop it. However, if you are aligning silos or working across divisions or bringing others into alignment, or working cross-culturally, or introducing new teams, these issues can be big. Again, plan for it – put someone in charge of its care and feeding.
• Use it as a way to create excitement and ownership. Once you work together, it is always good to look for accomplishments to celebrate. Depending upon your scope, it can also be a way to generate a new level of buy-in. Manage the group right and they walk away with the justified feeling that they had part in it – that they created it and it reflects their slice of the business. Trust me here – then they will socialize it and ensure that its followed in their domain!
Now that is exciting stuff!
~ Scott Felten
Comments
4 Responses to “Taxonomy: It sits in the critical path of …”
Got something to say?
Polls
Loading ...Latest Headlines
Recent Comments
- Christie Bautch on A Secret from an Old Man
- Jodie on A Secret from an Old Man
- Jodie on When You Have Trust
- Guy on A Secret from an Old Man
- Steven McWhorter on A Secret from an Old Man
[...] Continues at The Future Value of Business [...]
Right on Scott!
Taxonomy is also very important when you think about Information Architecture for portals, collaboration sites, etc. Without a well defined and well thought out taxonomy, it would result in much chaos - a good taxonomy helps to mitigate potential chaos. I also allude to this in my post around Enterprise 2.0 and Governance (found at http://thefuturevalueofbusiness.com/enterprise-20-and-governance.htm).
Raveen
You are correct. Taxonomy plays important role in the whole data management scenario. Especially when you take the data extracts out of the legacy systems, analyze it and then clenase it and then needs to classify, enrich your data. I worked on one of the classification thing earlier.
I guess the missing part in the above list is Data profiling. Thats simple yet valuable step in any data management / standardization project. I have described its importance, approach and solutions in my blog - http://manageyourdata.blogspot.com
Prashant,
Very good point! Taxonomy is to data profiling as good foot skills are to excellent football (soccer). One begats the other - its the chicken and the egg of data quality. Which comes first - a decent taxonomy or the effort of data profiling. I suppose that if we start with the business folks in a weak data infrastructure environment, we would purly speak about taxonomy. Given a somewhat mature data infrastructure or starting from IT, we would most likely want to begin our discussions on the interim data profiling results.
Either way, nice catch Prashant!
In your post “Approach to Start the Data Profiling” - I agree that if you are very familiar with the data and are equally talented in sql, then a db driven approach may be appealing. On the other hand, purchasing tools gives broader coverage of profiling ability. Like you are implying, you have to weigh the benefits of each to make the decision.
Take care,
~Scott Felten