I don’t know exactly what goes into a hot dog. I haven’t had to think much about it until recently when pressed by my kids. We try to buy good quality stuff but the best that I could do was say “well.. it’s meat..” — not good enough. Peppered with specific questions on specific ingredients I had to throw up my hands and profess that while I don’t know the exact ingredients it’s all safe and all fit for consumption.
As I danced around the specific questions related to hot dogs I realized how similar this scenario was to that which I had seen several times when a report is presented and the origins are brought into question. Where did this number come from? Why are the total sales not aligned with the other report? What is the definition of cash for this report? Who provided this number on the eastern region?
Organizations make decisions based on a diet of information. If the origins are opaque and unclear and the information is potentially flawed then so are the decisions. To have confidence in the decisions, the organization needs to have clarity on the sources and history of information.
The current state of most organization’s information architecture would roughly resemble that of a meat grinder. There are many inputs of a variety of types and quality levels. These get extracted, transformed and loaded into new data forms (sometimes re-factored several times) and are often manually adjusted for various reasons and purposes. These results and sources are then frequently re-factored for storage into a business intelligence platform or platforms. The information is then analyzed, re-assembled and packaged into reports, fact cubes, or other results. The clarity from source to consumer is often lacking.
Much of the reason for the information convolutions relates to the systems complexity sprawl: we continue to incrementally add things and bolt on solutions resulting in a myriad of systems and data models. Transformation programs are a great opportunity to unwind this complexity and information transformation can provide a clear lens on this progress.
Many transformation initiatives (outside of information centric domains such as risk or investment management) are not really focused on the information flow or design.
So how can this be addressed as part of a transformation initiative? Here are a few ideas:
- Make information transformation a key priority of the program: Have someone leading this and tasked with the end to end clarity and usability of the information.
- Question the past, envision the future: Many of the ways that data has been consumed in the past will no longer hold in the future. Consider reducing the number of reports, generated views, and making information more consumable than prepared.
- Maintain a view of the end to end information flows: Consider the impact of changes on the information design, balance these off of other vectors such as time/cost/complexity.
- As much as possible reduce the number of steps, layers, and processing that information goes through: Simplify the design.
- Try to get the information right at the source instead of spending time adjusting it further along the process: Reduce sources, validate input,
- Create transparency on information sources: have adequate meta-data so that people can understand and interpret the source further down the chain.
- Have clarity on the strategic information levers: Know which information elements are most important to the business outcomes and have a focus maintained on getting these right.
Making information flows a central theme during transformation helps to keep a focus on complexity reduction and data sprawl. Getting quality information is a major input to making good decisions so it’s worth the extra attention and focus. Instead of having information design be an after-thought of the transformation, having it established as a strong foundation can help to simplify the overall design and greatly improve business outcomes. With a simpler design and transparency comes confidence in the end product so that when the difficult questions on sources and origins materialize simple and direct answers can be provided.