Tips for Excel, Word, PowerPoint and Other Applications
Analytics and Modeling
Data modeling and analysis is an art, not a specific task that is done once in a while then forgotten about. It is an internalized, holistic way of thinking through a problem from the from time you begin to the time the ultimate question is answered to satisfaction. In any case, I'll use this section for larger macro sections of how I work through analytical problems because they usually involve multiple data sources, outputs, users, and applications.
All too often, the standard approach to an analytical problem is to start with the data at hand, churn about for a while and see what interesting numbers pop out, present the mass of data and then hope your stakeholder (general term for anyone who's asking the questions) can draw the insights required.
I'm a firm advocate of flipping that process on its head and working backwards:
- Sit down with your stakeholder and determine what answers are needed
- Based on the answers needed, figure out the appropriate questions to ask
- Structure your analysis to answer the questions
- Determine what data is required to support the analysis
This may sound strange, but how many of us have sat down with a stakeholder, received their input on what they want to see, then after about 5 iterations of analysis and meetings, we finally figure out what they really wanted. You may laugh, but those extra 5 iterations cost you in terms of time, stress, and productivity Had we done our homework and spent an extra 15 minutes getting to the heart of what questions need to be answered and how the data would be used, we could have saved ourselves work, stress, and aggravation. We would have been able to go home earlier more often.
In the larger context, not being crisp and precise about what questions we need to answer is costly and time consuming. We see this all the time in the Request for Proposal (RFP) process. While it may be easier to ask for everything under the sun, in the guise of, "we might potentially need it, so ask for it", it costs suppliers time and effort to answer all of your questions/provide all of your data and these costs are passed on to you in the form of higher prices. Your evaluation team has to sort through all of the information, which can get overwhelming and therefore lead to poor decisions.
Think about this in terms of adding an extension to your house. You would think someone was nuts if they just drove down to Home Depot, bought a truck load of supplies and just started building, hoping that the extension would take shape as the building progressed. No, you plan an extension in the same way you should plan your analysis. You would ask yourself:
- What do I want this extension for (game room)?
- In order to support the things I want to do in the game room (e.g., pool table) how large does it need to be?
- Based on the dimensions, what kinds of materials will I need and in what quantities?
- Depending on availability and delivery as well as other resources and prerequisites, what does the build schedule look like?
Pretty straight forward, right?