iSixSigma

using doe in actual customer application

Six Sigma – iSixSigma Forums Old Forums General using doe in actual customer application

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
    Posts
  • #35033

    gordo
    Participant

    Can anyone share any advice about things to watch out for or lesons learned in running a Design of Experiments in actual customer locations?
    (A little bit of background…My team is trying to test options for a new cleaning procedure to be done by our customers – in terms of how the procedure is communicated to the customers, the contents of a cleaning kit, which surfaces to clean, etc…). We are working to define very precise operational definitions for all the metrics and get stakeholder buy-in prior to the experiment but we want to be certain that management will believe the validity of the results [which is why my team is focused on getting real results from the end user versus having a bunch of internal people try to do the same procedures…] Any advice would be greatly appreciated :)
    Thanks
    Jason
     
     

    0
    #97403

    SSNewby
    Member

    Yep.   Don’t get so hung up on structuring an intriguing multivariate analysis, tweaking input factors, and assessing the impact to the system that you lose track of the fact that your issue is more than likely identifying sources of process variance and eliminating them.   A lesson learned…

    0
    #97906

    Kim Niles
    Participant

    Gordo:
    Here’s a couple points of advice:
    1.      Make sure you watch what happens as possible.  I’ve performed similar experiments thinking those taking the data are well trained only to find out they mixed the samples in the bags.
    2.      Throw in one variable that you already have a good feel for as possible.  That way if the results for that one variable look strange, then you’ll know something is wrong. 
    3.      Create pretty plots as possible as they help sell subsequent experiments and or the results.
    4.   Perform confirmation runs after the experiment.   
    Good luck either way,
    http://KimNiles.com
     

    0
    #97913

    Jansen
    Participant

    Gordo-
    Similiar to what Kim explained, have 2 runs of center points at the beginning, middle and end of the experiment. check to make sure that you get the same results for each run. If you dont, something changed over the length of the experiment.
    Hank

    0
Viewing 4 posts - 1 through 4 (of 4 total)

The forum ‘General’ is closed to new topics and replies.