iSixSigma

How Do You Measure Variable Output?

Six Sigma – iSixSigma Forums General Forums General How Do You Measure Variable Output?

Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
    Posts
  • #53884

    Peter Aveyard
    Participant

    What techniques can you apply where the input and output of a process varies considerably.

    I\’m talking about the production of software and changes to software. Every requirements is different from the last so no two process outputs are the same.

    Can anyone offer any advice on how best to approach measuring the process?

    0
    #191757

    Ken Feldman
    Participant

    While no two outputs might be the same, are the underlying production and change processes different or are they relatively consistent across the different software products? You may also look at whether you need to focus on the absolute requirement or can you measure the deviation from requirements.

    0
    #191761

    Peter Aveyard
    Participant

    Hi thanks for getting back to me on this. The underlying process does not change. The activities carried out are all the same the only variable is the amount of work to be done.

    Think of it as a bakers shop, the first order is for a loaf of bread that weighs 1 kg the second is for a loaf that weighs 12kg, the third is for 7kg and so on. Its the size of the order that changes.

    0
    #191766

    john berilla
    Guest

    What are you trying to affect? I would think a Goal Question Metric approach would be a good place to start your research. In addition, any development process should create consistently profitable value streams, so metrics (actual and projected) like ROI (ROI = Earnings – Investment / Total Life * Investment) and market share change can be used to assess the effectiveness of a development process or proposed project effort. In addition, phased time components – which drive development cost and revenue capture – can also be used. For example, cycle times from concept to prototype, or prototype to testing are another option. Food for thought – nutritional content is debatable – good luck.

    0
    #191770

    Ken Feldman
    Participant

    Peter, you made my point. My analogy was the printing industry. Although each publication coming off the press is a different color, different size and with different words, the underlying process of setup and printing is the same. Therefore we focused on those activities and didn’t worry much about the output other than whether it met the specs or not. I think you have the insight, now go forth and improve.

    0
    #191780

    Mike Carnell
    Participant

    Just as an fyi if you have a 1lb loaf or a 12 lb loaf you can go to defects per pound. When you have a variable output you need to tie to something that drives changes in the metric that are related to things like volume. That is why there are metrics like dpu. If I count pure defects and the line volume drops so does the defect count. This isn’t indigenous to software.

    Just my opinion

    0
    #191785

    Peter Aveyard
    Participant

    Thanks for the feedback chaps.

    0
Viewing 7 posts - 1 through 7 (of 7 total)

You must be logged in to reply to this topic.