Bill of materialsBill of Materials (BOM)

A bill of materials (BOM) is the recipe of a finished product: the full list of components and materials that go into making one unit, with the quantity of each. It's the document that says exactly what one unit is made of.

For a beauty brand, the BOM for one unit of a serum might be 30ml of bulk formula, one glass bottle, one pump, one cap, one front label, one back label, and one carton. For a fashion piece, it's the metreage of each fabric, the trims, the zips, the buttons, the care label, and the polybag. For a food or CPG product, it's the ingredients, the bottle, the cap, the sleeve, and the outer.

The BOM does one job that nothing else does: it converts product into materials. Once you know you need 5,000 units of a SKU, the BOM tells you that means 150 litres of bulk, 5,000 bottles, 5,000 pumps, and so on. Without it, a finished goods forecast is just a number of units with no way to act on it upstream.

What's on a BOM

A single-level BOM is a flat list. Each line has:

• The component or material (the bottle, the label, the fabric)
• The quantity per finished unit (one bottle, 1.4 metres of fabric)
• The unit of measure (each, metres, litres, grams)
• Often a wastage or yield allowance — the bit you lose to offcuts, spillage, or fill tolerance

That wastage allowance matters more than it looks. If you assume 30ml of formula per unit but actually use 31ml after fill losses, a 5,000-unit run overshoots your bulk by 5 litres — and over a year that unaccounted gap becomes a stockout on the input you forecast most confidently.

Multi-level BOMs

Many products aren't made in one step. The bulk formula is itself manufactured from ingredients before it's filled into bottles. The garment is cut and sewn from fabric that was knitted from yarn. That structure is a multi-level BOM: the finished good is built from sub-assemblies, and each sub-assembly has its own BOM of raw inputs.

A two-level beauty BOM looks like this:

• Finished good — 1 unit of serum
  • 1 × filled bottle (sub-assembly)
    • 30ml bulk formula
    • 1 × bottle
    • 1 × pump
  • 1 × carton
  • 1 × label set

The reason multi-level matters for planning is lead time. The ingredients that feed the bulk may have a 12-week lead time, while the cartons arrive in two weeks. You can't plan the buy off the finished-good date alone — each level has to be exploded back through its own lead time so the longest-lead input gets ordered first. The work-in-progress between levels is the work in progress sitting at your manufacturer between a sub-assembly being built and the final unit being filled.

How a BOM connects a forecast to a buy

This is the BOM's real value in demand planning. The chain runs:

1. Demand forecast — how many units of the finished SKU you expect to sell.
2. Production plan — how many units you need to make, after accounting for stock on hand and safety stock.
3. BOM explosion — multiply the production plan by each BOM line to get gross material requirements.
4. Net requirements — subtract the raw materials you already hold, then round up to each component's MOQ.
5. The buy — the purchase orders that actually go to suppliers, by lead time.

Step three is called BOM explosion, and it's where a forecast in units becomes a shopping list in components. Get the BOM wrong and every number downstream is wrong — you'll order the right quantity of the wrong thing.

Why component MOQs complicate the buy

A clean BOM still doesn't give you a clean buy, because components have their own MOQ. You might need 5,000 bottles but the supplier's minimum is 10,000. Now you're holding 5,000 spare bottles against the next run — which is fine if the SKU is a steady seller and a problem if it's a one-season piece.

Shared components make this worse, in a good way: if three SKUs use the same pump, you can pool the demand across all three before you hit the MOQ, which is exactly the kind of consolidation a flat per-SKU buy misses.

When BOMs drift out of date

BOMs rot. A formula gets reworked, a label gets redesigned, a supplier swaps a component — and unless the BOM is updated, the explosion produces a buy for parts you no longer use and misses the ones you now do. The most common failure isn't a wrong BOM; it's a stale BOM that was right two seasons ago. Treat the BOM as a living record, versioned alongside the product, not a spreadsheet someone built at launch and never reopened.