Often when we request an analysis from an engineer we run with the results and don’t ask a lot of questions about the analysis itself. Having done a lot of analysis I am familiar with all the assumptions and estimations that go into making a calculation work. But that means that the results of the analysis are only relevant to those assumptions and estimations. The analyst may have to make the following “calls” without additional input or only a small fact-finding mission.
- Remove a suspect part of the data set that appeared contaminated by a non-relevant factor or collection error
- There is not a clear definition for how these two parts will be assembled yet and an assumption had to be made on boundary conditions
- There is no record of if variable “C” was controlled during the experiment
So the end user, likely not the analyst, will be reporting on the results and may not have all the answers to upcoming questions coming from other parties. In many cases this is the person/team that is making a large decision and these analysis results are apart of that process. They may even be the hinging factor.
It’s important to minimize the potential of misunderstandings and misapplication of results. From the information customer’s end this is what is needed.
- Management teams need to understand the story behind the results
- They need to consider the drivers behind a trend.
- They need to consider the risks underlying their decisions today.
- They need to look behind the numbers
This can be easily accomplished with a little work on both ends. The analyst can create a multi-layer document that allows a user to search for details on the assumptions and estimations. This can be done in anything from Word to Excel to a custom piece of software. The end user can do their part by taking a little time to, in advance, review the results and the reference doc. They should confirm their understanding of both to the analyst before incorporating the results into their work.
The true end game for analysis in many cases is assisting with making decisions. So how are decisions made?
- The person starts with a problem – determining which college they will be attending.
- They then look at the alternatives – the schools they can choose from.
- Now, they decide which is more beneficial by looking at their criteria – what they believe is most important in terms of choosing a school. This can be things like, the price of the college, the location of the college, etc. Things that are important to that particular person.
- People then compare their criteria to their alternative; they see which best fits their criteria. They see which alternative is the most beneficial for them.
- Finally, they make a decision.
So the boundary conditions for the decision process and the analysis must match. Whose responsibility is that? Everyone!
Looking at the big picture it becomes very clear that all parties involved with the analysis, data input, calculating, result interpretations, and final decision making are required to synchronize on the peripheral factors before, during, and after. It’s a guaranteed “miss” if everyone assumes they are on the same page.
A little effort can go a long way with making analysis results an effective contributor in program decision making.
-Adam
Leave a Reply