Calculating Process Capability for Wait Time at a Restaurant

Six Sigma – iSixSigma Forums General Forums Methodology Calculating Process Capability for Wait Time at a Restaurant

Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
  • #53424


    I am new to Six Sigma, and am trying to find Process Capability for the wait time at a small barista-style coffee shop. The trouble is, because a good number of the wait times are zero and I have a small sample size, the distribution is skewing far to the right.

    The data is separated into two chunks, and was collected at the same time each day for a week. The first is the wait time to go from the back of the line to the front (if there is a line), and the second is the wait time from placing an order (a beverage) to receiving said beverage, both in seconds. The first data set for line wait is labeled Order, while the second is labeled Receive.

    I know that in order to calculate Capability, the distribution needs to be normal and in control, however this data is neither.

    I have attached the Minitab file to this message, and if anyone could walk me through this process, I would greatly appreciate it. Please be gentle. I am VERY new to this.

    Thank you,



    Sorry, here is the file….

    [file size=75862][/file]



    Have you considered approaching this problem from more of a Lean perspecive than a Six Sigma one? Both sides of the force are very strong :-)

    If you go Lean (to start) you can look at customer demand (by hour to handle “busy” times). You need to calculate Takt time to see how often the line needs to complete a new order. When you look at the cycle times of each “station” (take order, fill order, paymentrcvd) you can determine the bottlenecks (any longer than Takt time)… ya-da… ya-da

    Let me know if you are interested in this path…
    bobhody at yahoo dot com



    Actually, your data does not need to match a normal distribution for process capability. You’ll see in minitab an option for process capability – non-normal. If you know your distribution, you can select it in the non-normal option in a drop-down menu.

    If your data is not in control, understand and eliminate any special causes before you do the Process capability function.



    Hai Awylan,

    whenever I have a data set that contains a lot of 0’s I split my capability analysis of the CTQ in two.
    CTQ1= % non-zero’s; Target is very often 0; goal of project part1 to reduce average value and variation of CTQ1
    CTQ2 = CTQ if CTQ is non-zero. target = 0; goal of project2 to improve Cpk and reduce values.
    By splitting this up in tow problems the data is easier to investigate.

    Good luck,




    Thanx for your open advice. I was stuck in nearly a same problem from last a week and feel that it will solve using this technique. I have implement this and now looking for the result.

    Keep advising good ideas !



    You didn’t actually get any good advice except from Delannl.

    Process capability is not the tool you need and neither is takt time.

    Some simple facts – this is a queueing problem and arrival times and service times have expected distributions and neither is expected to be normal.

    Do some basic reading on queueing (start with infinite queue assumptions). Come back with an informed question and you’ll get help.

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

You must be logged in to reply to this topic.