Looking for Help with Demand Forecasting

Six Sigma – iSixSigma Forums General Forums Methodology Looking for Help with Demand Forecasting

Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
  • #244137

    Ikrame Mahrou

    Good afternoon all!

    Greetings! We are a group of students who recently chose to specialize in Lean/Six Sigma. For us to complete our course we are required to work on a project which is the following:

    We have to come up with a demand -forecasting model + re-order point for a company. The demand -forecasting model is for three main products. The data we have received are sales numbers of 2016-2017-2018 accompanied with purchase number and of course the Sales price and purchase price.

    We have recently bought minitab to use it for the model and re-order point but we have no idea what variable we need or how we can do it. Unfortunately our teacher isn’t really knowledgeable in forecasting methods.
    We could really use some help/guidance perhaps any of you could either help us out or steer us into the right direction.

    Thank you very much in advance!


    Robert Butler

    I think you will need to provide more information before anyone can offer much in the way of suggestions.

    1. You said, “The data we have received are sales numbers of 2016-2017-2018 accompanied with purchase number and of course the Sales price and purchase price.”

    2. You also said, “We have to come up with a demand -forecasting model + re-order point for a company.”

    Question: What is the level of detail of the sales numbers?  Monthly/weekly summaries? Actual sale-by-sale reporting for each of the three items and the day they were sold? Or…something else?

    Question: What is the criteria for the need to re-order? X number of items 1,2,3 in non-working inventory, JIT delivery with a known lag time for meeting JIT requirements? Or…something else?

    Question: Regardless of the level of detail, since it is time series data have you plotted the data against time?  If not, why not? If you have, what did you see?  Demand cycling? Steady state sales? etc.

    If you can provide answers to these questions perhaps I or someone else might be able to offer some advice.




    I think this is a poor choice for a six sigma training project.   I suppose that your instructor has his/her reasons, maybe that this is something management wants to address, so you’ll have to live with it.  If, as you say, your teacher isn’t really knowledgeable about forecasting methods, that’s a real bummer.  I’d advise you to look up information on Six Sigma for Demand Forecasting.  There’s some pretty good stuff there that you might find helpful.



    Usually Demand Forecasting Models deal with Lead Time variability taking in consideration each product supplier’s lead time and it’s comsumption history (data you received) variability, also you can organize your inventory using ABC (annual consumption volume * unit price) criteria and stabilish the replenishment policy acording to both above criterias!

    This should be the esiest part, it’s just math! What is not tottaly clear to me from your description is What is the problem you want to solve?

    Hope this helps with some insights.

    Best Regards,



    Dear Ikrame,


    I am currently studying a bachelor in industrial engineering and management.

    Maybe the formula from Camp is an option to take a look at.

    with this formula you can determine the reorder point and quantity.

    I have added the link to the wikipedia page.

    If you have any questions feel free to message me.

    The page is in dutch, but I hope I helped you and your group.


    Kind regards,

    Kjeld Stommels


    Mohamed Asif Abdul Hameed

    For a simple time forecast you could use Time Series Plot:

    Stat > Time Series > Time Series Plot

    However, when you want to forecast based on Model Type (Multiplicative/Additive) and Model Components (Trend/Seasonal) it is better to use Decomposition

    Stat > Time Series > Decomposition

    Advanced level of forecasting when you want to smoothen your data to reduce Mean Absolute Percentage Error, go for below forecasts:

    Single/Double Exponential smoothing or

    Winter’s method


    Go through below link for more clarification.



    Firstly, demand forecasting is an oxymoron, but what are you trying to forecast exactly? Most times when companies try to forecast or predict future sales, they are way off base. Instead look towards true Lean concepts (no six sigma required) such as Takt Time, cycle time and level loading. If you are trying to determine how many units to produce you will need to know what the rate of customer demand is where you would calculate Takt Time. (Takt Time = available working hours/# of units). From there you will be able to determine the amount of employees you will need based on Takt time (customer rate of demand) and cycle time (time in minutes it takes to make a product).

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

You must be logged in to reply to this topic.