Tips for Excel, Word, PowerPoint and Other Applications
The Zen of Modeling
What is a model?
A model is a mathematical simulation of some state of being. It is our best approximation of the financial impact our decisions will have on revenue, cost, profitability, savings, etc.
Data analysis and modeling are a journey, not a discrete destination. We continue to build and refine models as we go, learning more with each iteration.
Characteristics of a model
- Simulates alternative courses of action or sets of conditions (e.g., savings from awarding PC contract solely to IBM)
- Capable of 'what if' analysis (e.g., how do savings change if contract is split between IBM and Compaq?)
- Is dynamic (e.g., incorporates revised data and assumptions and updates results)
- Synthesizes data collected throughout the sourcing project
- Clearly communicates the conclusions, supporting analysis, and underlying assumptions
- Stands on it own
- Is 'Manager' friendly in its presentation
So where do we use data models in our daily lives?
At home, this might be, "how much money will I have in the bank if I buy a new car vs. not buying a new car?"; At work, this may be, "If I switch suppliers, how much will the company save?"; or "If I need to reduce my budget by 10%, what expense levers can I tweak to bring my spending level down?"
Typical Data Models
- Savings. How much could my company save by switching from one supplier to another? What portions of my spending should I switch to the new supplier?
- Negotiations. Does one supplier's offer to provide free cell phones offset another supplier's offer to reduce cost per minute?
- Finance. How much will my company spend in cellular phones in the next fiscal year?
- Make vs. lease. Should I lease or buy a new car?
- Acquisition. Will the planned acquisition of a printer manufacture pay for itself within 5 years?
Why learn the skills of effective model making and data analysis?
Understanding how to build data models efficiently ultimately leads to better results with less effort. The resulting analysis is more robust and useful because you have a better understanding of the key variables which influence the conclusion and you are able to model a range of inputs as your needs change. These models take less time to construct because they are built more efficiently, are flexible enough to accommodate a range of different analyses, and are robust enough to deal with constant changes.
If you or your company pays for outside consultants to come in and help you with your analytical work, you will end up paying thousands of dollars when realistically, a lot of the analysis could be performed by your own staff. You could avoid this or you need to at least challenge the analysis to make sure it's right for you
Saves time, saves money, builds more credibility in your work
Notes
| Last updated | 9/2/07 |
| Application Version | NA |
| Author | Michael Kan |
| Pre-requisites | None |
| Related Tips | None |