RSKRisk, cost & sustainabilityStop 2 of 9
Supply chain segmentation
Running different supply chains for different products and customers, instead of one model that fits none of them well.
Read atChange any time
Most companies run one supply chain for everything: the same planning rules, inventory targets and delivery promises for a staple detergent and for a limited-edition festive gift pack. It fits neither well. The detergent needs low cost; the gift pack needs speed and flexibility.
Marshall Fisher’s classic idea makes the split clear:
- Functional products (staples like soap, rice, basic shampoo) have steady, predictable demand and thin margins. They need an efficient supply chain: low cost, high utilization, lean inventory, full trucks.
- Innovative products (new launches, fashion, electronics, seasonal packs) have uncertain demand and short lives. They need a responsive supply chain: buffer capacity, faster transport, postponement and quick reaction to early sales.
Segmentation builds a few distinct supply chain models and assigns each product, and often each customer group, to the one that fits. Typical dimensions:
- demand volume and predictability (the ABC-XYZ view),
- product lifecycle stage (launch, mature, end of life),
- customer service needs (key accounts with tight delivery windows versus small stores served weekly),
- channel (modern trade, general trade, e-commerce).
A common combination is lean for the predictable base, agile for the volatile rest, sometimes called “leagile”: steady volume is planned efficiently, while peaks and new products get flexible capacity and faster lanes.
The hard part is keeping it to a handful of segments that systems and teams can actually run.
Words to use
- Functional vs innovative product
- Predictable, low-margin staples versus uncertain, short-lived, higher-margin products.
- Efficient vs responsive
- Supply chains built for low cost versus built for speed and flexibility.
- Leagile
- Lean for the stable base volume, agile for the volatile remainder.
Segmenting a portfolio
Classified by volume (high or low) and demand variability (low or high):
| Segment | SKUs | Share of volume | Supply chain model |
|---|---|---|---|
| High volume, low variability | 80 | 70% | Lean flow |
| High volume, high variability | 40 | 15% | Base plus surge |
| Low volume, low variability | 180 | 10% | Periodic batch |
| Low volume, high variability | 200 | 5% | Make to order or central stock |
Sixteen per cent of SKUs carry 70% of volume, and 40% of SKUs carry only 5%. One set of rules can’t serve both well.
What changes by segment
| Policy | Lean flow | Base plus surge | Periodic batch | Make to order or central |
|---|---|---|---|---|
| Planning | Make to stock, level schedule | Base level plus flexible capacity | Make to stock in campaigns | Make to order, or stock in one place |
| Safety stock | Low (predictable) | Higher, plus postponement | Moderate | None locally, or pooled centrally |
| Production | Dedicated, high-speed lines | Flexible lines or co-packers | Shared lines, longer cycles | Flexible, small batches |
| Transport | Full trucks, cheapest modes | Mix of slow base and fast top-up | Consolidated | Parcels or express when needed |
| Service promise | Next-day from stock | Short lead times | Standard lead times | Quoted lead times |
Fisher’s test
Fisher’s question is whether the supply chain matches the product: an efficient chain for functional products, a responsive one for innovative products. The two classic mismatches are costly in opposite ways. An innovative product in an efficient chain runs out and then gets marked down; a functional product in a responsive chain pays for speed nobody needs.
Lean, agile and leagile
Christopher and Towill (2001) argued that lean and agile are complementary. Lean suits predictable, high-volume flows, where eliminating waste lowers cost. Agile suits volatile demand, where speed and flexibility matter more than utilization. Leagile designs combine them, typically lean upstream of a decoupling point and agile downstream, or lean for the base volume and agile for the surge.
Segmenting by customer behaviour
Products are only half the picture. Gattorna (2010) proposed aligning supply chains with how customers buy, for example customers who want low cost and consistent supply, customers who want fast response, customers who want innovative solutions, and customers in crisis who want anything that works now. Combining product and customer views gives segments that match both what is sold and how it is bought.
Making segments real
A segmentation that lives only in a presentation changes nothing. It becomes real when:
- Systems hold different parameters by segment: planning strategies, safety-stock methods, lot sizes, transport rules.
- KPIs differ: cost per unit for lean segments, responsiveness and availability for agile ones.
- Organization gives each segment an owner who manages it end to end.
- Migration rules move products between segments as they grow, mature or decline.
How many segments?
Each segment adds complexity in systems, processes and people. Most companies end up with three to six. Too few and the segments don’t fit; too many and nobody can run them consistently. Start with the biggest differences in volume, variability and customer needs.
Linking to cost-to-serve
Segments should also differ in what they cost and what they earn. Cost-to-serve analysis shows whether customers in a high-service segment pay for it, through price, volume or strategic value, and supports terms that match the service each segment receives.
Interview kit
Practise it out loud
Answer each question yourself before opening the model answer.
Supply chain segmentation means running a few distinct supply chain models instead of one. Following Fisher, functional products with predictable demand get an efficient chain, and innovative products with uncertain demand get a responsive one. I’d segment on demand predictability, lifecycle, customer service needs and channel, keep the number of segments manageable, and often combine lean for the base with agile for peaks and launches.
Why shouldn't a company run one supply chain for all products?
Because products differ in predictability, margins and service needs. One model built for cost under-serves volatile, high-margin products, causing stockouts and markdowns; one built for speed overspends on stable staples. Segmentation matches each product to the model that fits.
How would you segment a consumer goods company's products?
Start with volume and variability (ABC-XYZ) and lifecycle stage, overlay channel and customer service requirements, and group them into three to five segments, such as stable high-volume, seasonal or promoted, new launches, and long-tail. Then define planning rules, inventory targets and fulfilment model for each.
Connected stops
Where this idea feeds in, or gets its inputs from.
Sources and further reading
- Fisher, M. L. (1997). What is the right supply chain for your product? Harvard Business Review, 75(2), 105–116.
- Christopher, M. & Towill, D. (2001). An integrated model for the design of agile supply chains. International Journal of Physical Distribution & Logistics Management, 31(4), 235–246.
- Gattorna, J. (2010). Dynamic Supply Chains: Delivering Value through People (2nd ed.). FT Prentice Hall.