This blog is inspired by a tweet last week from @CEOatSteelwedge. The tweet had a link to a blog from editor in chief from Supply Chain digest, Dan Gilmore. His blog was about rebuilding supply chains for the future. He refers to the ‘the smart folks at McKinsey’, which came up with a way to segment supply chains based on demand volatility and demand volume.

My supply chain mantra is and probably always will be focus, focus, focus. And how can you do that? The answer is; segment, segment, segment. And segmenting the sku portfolio is exactly what McKinsey did.

I’ll share with you 3 examples on how I used segmentation in my supply chain career. It’s nothing dramatic, nothing new, but the McKinsey solution isn’t new either. The segmentation examples I discuss are:

1. Margin and volume focus for finished goods
2. Raw material segmentation for sourcing strategy
3. Supply chain predictability segmentation

I hope you enjoy them and find them useful

1. Margin and volume focus for finished goods

Let’s start with the easiest segmentation example. In a food company I used a what I call a double Pareto analysis to define focus in finished goods. Once the data is available this is a reasonable simple exercise in Excel and can be done in half an hour per product group. Use the next steps:

1. Create Pareto (A, B, C analysis) for sku portfolio based on volume
2. Create a second Pareto based on gross profit for the same portfolio
3. Combine the 2 Pareto’s to define 9 segments of margin/volume focus

This will give a picture like this from a real life example of 99 sku’s

The segments can be used for different types of focus. For example:

Segment AA (high volume and margin) are hero sku’s. 10% of these sku’s deliver 46.7% of profit and 51.6% of volume. What to do with them?
– List them on a planning or IBP watch and prioritize list.
– Safety stock levels can be set higher as we get a good return on these products.
– Expect operation to be more flexible in short term change requests.
– Contracts for raw material in these products can be treated more strategic and long term to reduce risks.
– Sales field packs and marketing materials can be focused on these hero sku’s

Segments AB and AC are sku’s where we need focus and plans to improve margin
Segments BA and CA are sku’s where we need focus and plans to improve volume
Segment CC really needs to be scrutinized to optimize the portfolio and delist or substitute. This segment represent 47.5% of the sku’s and delivers only 6% of volume and 5.9% of margin

2. Raw material segmentation for sourcing strategy

In a beverage company, I did a segmentation to define different sourcing and inventory strategies. First I agreed in a workshop cross functionally the criteria for raw material purchasing complexity and criticality. The steps for this are:

1. Choose criteria for Complexity and Criticality
2. Assign Score to criteria
3. Assign Weight to criteria
4. Position product and define segment

This defines segments between low and high complexity and criticality. Highest focus will be raw materials that are both complex to procure and business critical.

In the next step create a Pareto analysis and combine this with the four segments. In this way you can segment high and low volume products that are critical to the business and complex to purchase. In this real life case a total of 11 segments. In this way we ended up with a cross functionally agreed set of raw material segments, on which we could apply different procurement and inventory strategies.

3. Supply chain predictability segmentation

In a similar approach as in example 2, we can segment based on demand predictability and supply predictability. In the article mentioned before, McKinsey used demand volume and demand volatility for segmentation. I would argue that demand volatility is only one criterion to define demand predictability. How many customers drive this erratic behaviour? Are these highly promoted sku’s? In which part of the lifecycle are these sku’s? Are there substitution opportunities? These are questions that go deeper then looking only at the end result, demand volatility.

Before we align our supply chain to volatile sku behaviour we first want to see if we can reduce demand volatility and answer some of these questions. Improve the forecasting process, communicate better between internal departments or with customers, decrease promotions, substitute high volatile with lower volatile products are only a few examples on how we can influence or shape demand and impact demand volatility.

Similar as the previous approach, the steps for this are:

1. Choose criteria for supply and demand predictability
2. Assign Score to criteria
3. Assign Weight to criteria
4. Position product and define segment

This can be done by product/source combination. It is important to agree criteria cross functionally, so when change is implemented after the analysis, there is no arguing on the data. Once scores are available on this level of detail they can be grouped to brand or vendor level. The picture shows two real examples of supply and demand predictability, mapped for a beverage brand and seven suppliers. The size of the bubbles can show volume or margin to assess a combination of risk and volume or value impact