iSixSigma

HopeOverExperience

Activity

  • I would start by try to persuade your COO that this isn’t a good idea but if you really need to.

    For each process decide how you can tell you’ve got a good product e.g. quality, on time delivery, time to produce, cost to produce etc, and turn this into a basic pass or fail.
    Combine all components into one large pass/fail measure, you may want…[Read more]

  • Not knowing your business I would try the following if at all possible and it’s a standard method. There are lots more though.

    Sponsor
    Get a key manager/managers to sponsor process improvement.

    Goal Setting.
    What is the startegy of company.
    Define measure to show implementation of measures.
    Which measures need to improve to allow you to…[Read more]

  • Why do you think a P of 0.0000 means there’s no significance?

    Put simply what this means that there is a 0.0000% chance of your data distribution appearing by chance. Therefore your two data sets are different. 10 years, 6 months ago

  • Good question. From what I’ve seen so far most have so many issues that they don’t need to get past Lean to get major improvements and where data is used, it tends to be used to prove they’ve made a difference, rather than to direct where they should be making a difference.

    Saying that in the past, 5 years ago or so, I’ve ran projects in the…[Read more]

  • @MMBinWI – welcome to the world of finance 10 years, 6 months ago

  • @Mike Carnell. I was also talking about serious data analysis and there was strong correlation even when taking out other factors e.g. age, credit score etc. I’m sure that you know there are always outliers which do not conform to the rules and as I said when I did the analysis it was four or five years ago so it may have changed as mobile phones…[Read more]

  • The credit score card which you’re talking about is one which is being looked at but out of my remit, another team is looking to see if the rules they have hold true with the current data. I am looking for indicators that people who are likly to go to recovereis so we can speed them through so we can concentrate on the people likly to pay up and…[Read more]

  • Thank you for the suggestions. The issue with collecting information going forward is the length of the process in recoveries to get to the point of writing off the debt or collecting on it.

    When you add in time for legal to get involved where required, for closing and selling property, the use of one or more outside agencies, customer going on…[Read more]

  • I’m currently working for a UK bank looking at their collections and recoveries departments.
    I was wondering if anyone had experience in this area and in particular have run projects looking at how the amount of […] 10 years, 7 months ago

  • I would go for a quick overview of the basic concepts and what the wastes are and then dive in. If you think they need to know a tool go through it then and get them to apply it to their problem.
    Teaching them tools they don’t need at this point would only confuse the issues. This is only the first pass remember what you need is some quick fixes…[Read more]

  • The visual aspect of standard work tends to have a major piece to play in the design of the completed solution. As this in most cases will dictate the ease of use, miss use (errors) and speed of use.

    What you need to do to test will be down to the number of times it’s in use and complexity and consequences of getting it wrong. Something which…[Read more]

  • Hi

    I would look at the process and equipment used for producing your SKU’s and use these to caclulate the OEE but do this from the suppliers point of view and take your SKU’s out of the equation.

    By improving the suppliers overall OEE you will help your own company.

  • Well done on getting this far then.

    Run your tests on the original data.

    Remember that if your tests show you something you didn’t expect check your data and you’ve used the correct test before acting on your findings.

  • 1, Yes

    2, On Orginal

    3, Stability is how the data changes over time more or less, and normality is the overall shaoe of your data and doesn’t take time into account.

    Out of curosity can I ask where you were trained?

  • On number of rows in minitab there used to be somewhere to set the maximum but I can’t find it now. Saying that I just created 10,000,000 rows of data will this be ok.

    I’m not suprised by the data being stable between weeks and not days.

    If you’ve got too much data to fit into Excel I’m guessing you have enough to start your project.…[Read more]

  • Basically my project is for collections..(Collect the money faster)…
    There are around 300,000 invoices per month.

    If your systems are electronic I would see no reason to sample your data.

    I would like to ask is how are you grouping your data to check for stability?

    I’ve worked in collections for late payers of loans and our data was…[Read more]

  • So, Hope: Let us in on the secret.

    Treat as a stability issue.

    D
    Objective = Make the process Stable

    M
    Process Map your process
    FMEA, Root Cause… to get what you think your X’s are
    MSA your data – is it real stability issues or measurement problems
    Collect Data

    A
    Split data by your different X’s and can you see stability within…[Read more]

  • In brief possible answers arel
    1 and 2. Set up a script/ decsion tree to guide the operator through the call and to make the correct decision. Work with them to make sure they understand it and it’s not too restrictive to them. If it could be electronic that wold be best seen these simpley set up in Excel if that’s all you’ve got.
    3, Look for…[Read more]

  • What’s missing from the above two answers?

  • You shouldn’t run a process sigma score on the data but you could do a product one.

    Hyp testing could tell you why your data is unstable. As could process mapping and any of the other root cause tools.

    yes you can run a DMAIC project. It’s goal should be to get a stable process/

  • Load More