In my previous post, I encouraged you to mingle with your new colleagues and make yourself known as the new go-to person for when planning is involved. While doing so, you have not only started building relationships, you’ve also raised some expectations. Now that you have an idea on what challenges your company might face, you can actually start planning.

Back in the day, plans heavily relied on experience which, in the best case, was shared from veteran to newbie. Fortunately, we have tools these days that can help capture the knowledge.

That’s nice but how do you start building a planning tool? Whatever software you will use, make sure to map out your business first. Refer to your informal meetings and ask what questions your model should help answer. Next to these questions, define what decisions you want the data to support. Every threat will have different alternatives but if pre-purchasing and storing raw materials is impossible, there are really no decisions to be made there and the tool should not be modeled for that. If on the other hand, warehouse space can easily be rented on a short term basis at different locations for different rates, you probably want to spend more effort in that area.

In any case, the first step to build a model is mapping your current situation. There are numerous valuable applications available but paper & pen or a whiteboard can also get you a long way. As you start drawing, continuously question what you do. What can cause disruptions on a certain production line, are there alternatives for a specific STO flow… make sure that once you have completed this exercise, you have a thorough understanding of how your business operates including the bottlenecks and potential threats. Once you have that, you need to define the data requirements. There is no supply chain management without data but you have to identify what is relevant for your problem(s). In general, a good planning system will require data in 4 categories:

  • Demand by time – product – customer
  • Transportation
    • Delivery lead times (location to customer)
    • Procurement lead times (supplier to location)
    • STO lead times ( location to location)
  • Inventory by product location
    • Current inventory
    • Standard cost
    • Target inventory
    • Storage capacity
  • Supply
    • BOM Run Rate
    • Operating hours by work center
    • Procurement by location – materials

For each of these categories, there is a lot of data available to you. Transportation to the client can have distance, cost, mode, energy consumption, supplier, possible time slots for delivery and more. Depending on the problem you want to solve, you need to decide what data is relevant, where it is stored and how you can get your hands on it.

Now is a good time to validate your data. We all tend to agree that clean data is mandatory for meaningful results. Yet, I have seen it happen too much that people start modeling and plan on cleaning data once the model is up and running. We all want results and we want them fast. My advice? Sit back, take a deep breath and tackle this now. Assess how accurate your data is and what is needed to clean it up. Don’t be a maniac but remember that Garbage In, Garbage Out will bite you if data errors are ignored now.

Finally, you can start modeling! The easiest way to continue is by identifying all valid product-location combinations. A product-location combination is valid when 1 of the following is true:

  • You have final demand for this product at this location
  • There is production defined for this product at this location
  • The product is being consumed as part of a production activity in this location

Once we have these listed, we can define the valid procurement types. For those that are produced, we can then identify the relevant work centers, the BOM and the production rates alongside with the inventory details.

As you have cleaned your data, you know where it usually goes wrong…try and incorporate some automated alerts if things go astray again. This will again make your life easier further down the line and help explain why certain results are what they are.

Like this blog? Please share with colleagues and also follow us on LinkedIn or Twitter and we will send you notifications on all future blogs.