While most companies have too much data, people can always identify something they aren’t tracking that they should be tracking. Then they think they have to set up a whole system to collect the measurement. This is a mistake. You don’t know if the measurement is useful until you have collected some. Rather than wait for a measurement system, start today using the simplest of tools: a checksheet.
I’ve use these kinds of checksheets when I’m working with a team to implement Six Sigma. They find causes, I write them down and tally the number of times each occur. By the 30th data point, a pareto pattern appears that points us at the most common (i.e., root) cause of the problem.
Checksheet Data Collection
Nothing could be simpler than data collection with a checksheet. The QI Macros have a template in the improvement tools to get you started:
Simply print it out and start writing on it. What to write?
- In column A, write the first instance of any defect, problem or symptom you detect. For example, if someone is calling us for support and they have a problem with understanding GageR&R, then we’d write “GageR&R” in A3 and put a stroke tally in the day of the week (e.g., Monday):
- Then continue as the week goes on, adding defects, problems or symptoms. By the end of the week, we’ll have an interesting picture of support calls.
Just add up the number of calls and “Excel disabled macros ” jumps out as the majority of the calls. Create a pareto chart to better demonstrate this.
But now the questions arise:
- Does a specific version of Excel disable the macros?
- What are the SPC Questions about? Histograms? Control Charts? Which control chart?
- And so on.
Well, if we spent a lot of time and money setting up a measurement system to track just these items, it wouldn’t do us much good. Now we can tune up the checksheet to give us better data:
Use a checksheet to prototype your data collection efforts. Iterate until you start to understand what you really need to know to make improvements. A series of checksheets may be all you need to solve a pressing problem. If necessary, you can implement a measurement system to collect the data over time.
So please, don’t wait for a magical, all encompassing measurement system to deliver data. It’s not going to happen. And I often find it’s just an excuse to avoid making improvements (“I can’t because I don’t have the measurements I need.”)
