What is direct spend analysis, and how do you run it?

What is direct spend analysis?

Direct spend analysis is the review of what a company pays for the materials and parts that go into its products. It works part by part: unit prices across plants and suppliers, the material cost of each finished product, and the few parts that drive that cost. The result tells buyers where to negotiate, standardise or re-source.

The Chartered Institute of Procurement and Supply (CIPS) defines direct spend as “the purchase of raw materials, goods and services that go directly into a product being manufactured or into the services being delivered by an organisation”. For a pump maker, that means castings, impellers, bearings, motors and seals. For a food company, it means ingredients and packaging. For a car maker, steel, tyres and electronic modules.

A typical first finding. A company with two plants makes the same pump in both. The analysis shows that the second plant pays 17% more for the same bearing, ordered under its own local agreement. Nobody had compared, because each plant’s buyer only saw its own orders. That is the kind of question direct spend analysis answers.

On the web, the term is mostly used by software vendors and consultants, as one type of spend analysis next to indirect, tail and category spend analysis. The general method (collect, clean, classify, analyse) is explained in spend analysis. This page covers what changes when the spend is direct materials. The difference between direct and indirect procurement itself is explained in indirect procurement.

How is direct spend analysis different from general spend analysis?

General spend analysis, as the US Government Accountability Office described it, shows who the buyers and suppliers are and how much is spent on which goods and services. It usually starts from invoices and payments. Direct spend analysis goes one level down, to the part, because a direct material is bought again and again, in large quantities, against a known need.

Two other views sit next to it. The ranking of spend by parent supplier, useful for direct and indirect spend alike, is in supplier spend analysis. The many small suppliers at the bottom of the ranking are the subject of tail spend management; CIPS describes tail spend as the bottom 20% of spend, usually 80% of suppliers.

How large is direct spend in a manufacturer?

In many manufacturers, materials are the largest cost. US Census figures for 2021 give an idea of the order of size, sector by sector.

Read these ratios with care. The Census measure of cost of materials is wider than direct materials: it also counts fuel, purchased electricity, contract work and goods bought for resale, and it includes freight paid to bring materials in. The figures are for the United States and one year. They still show why a small price gap on a direct part matters: it is repeated on every unit produced.

What data does a direct spend analysis need?

The heart of the data is the bill of materials (BOM). SAP’s glossary defines it as “a complete, formally structured list of the components that make up a product or assembly”, with the quantity and unit of measure of each component. Joined to purchase prices, the BOM turns part prices into a material cost per finished product.

The hard part is usually the part master. The same bearing may exist under two part numbers in two plants, with two descriptions. Before comparing prices, someone who knows the parts, often from engineering, has to confirm which part numbers are the same item.

How do you run a direct spend analysis, step by step?

1. Set the scope: which plants, which product lines and which period. Twelve months of purchase order lines is a common start.
2. Extract order lines, receipts, part master data and bills of materials, with the part number and the plant on every line.
3. Map identical parts across plants and part numbers, with engineering’s help. Group similar parts into part families, such as “deep groove ball bearings, 25 mm bore”.
4. Calculate spend per part: quantity received × unit price paid, converted into one currency and one delivery term.
5. Compare unit prices for the same part across plants and suppliers, and calculate the gap: (price paid − lowest price) × quantity.
6. Roll part prices up through the bill of materials to get the material cost per finished product, and rank parts by their share of it.
7. For the few parts that drive most of the cost, build a should-cost estimate and check price movements against the relevant commodity index.
8. Review the findings with category managers, engineering and finance, and write down one decision per part or part family.

Keep the part mapping and the currency and delivery-term rules with the results. Next year’s analysis must use the same rules, or a change in the figures may only reflect a change in method.

What does a direct spend analysis look like on a real product?

The example below is fictional. A company makes the same pump in two plants: North builds 10,000 pumps a year and South builds 6,000. Each pump uses one housing, one impeller, two bearings, one motor and two seals. The table compares the unit prices each plant pays.

Two findings come out. First, three parts have price gaps between plants, worth €31,800 a year. The quick decision is to extend North’s agreements to South, or to run one sourcing event for both. Second, and more important, the motor is €96 of the €159.50 material cost at North, about 60%, and both plants pay the same. A gap analysis will never question that price. Only a should-cost estimate and an index check can tell whether €96 is fair.

Check of the totals: North spends 10,000 × €159.50 = €1,595,000 and South spends 6,000 × €164.80 = €988,800, so €2,583,800 together. The gap is 6,000 × €5.30 = €31,800.

What is a should-cost model, and when is it worth building?

A should-cost model estimates what a part ought to cost, built up from its cost elements: material, processing, scrap, overhead, profit and transport. It answers a different question from a price comparison. A comparison shows that you pay more than another plant; a should-cost model shows whether everyone pays too much.

The idea is old in public procurement. The US Federal Acquisition Regulation describes should-cost reviews as “a specialized form of cost analysis” that does not assume a contractor’s historical costs reflect efficient operation, and says a program review evaluates direct costs such as materials and labour. In the United Kingdom, the government’s Sourcing Playbook asks for a should-cost model on complex outsourcing projects. In industry, buyers use the same logic on parts.

A should-cost model takes time and needs people who understand the manufacturing process. Build one only for parts with high spend, few suppliers or a price that seems out of line. For a 2 euro bearing bought from a catalogue, a price comparison is enough. For the €96 motor, which repeats on every pump, a model is worth the effort.

How do commodity indices help analyse direct spend?

Many direct parts follow the price of a raw material: cast iron, steel, copper in a motor, resin in a plastic part. A price index tracks that movement. Comparing a supplier’s price increases with the index shows whether the increases follow the market or go beyond it.

Contracts can tie prices to an index. The US Federal Acquisition Regulation lists three kinds of economic price adjustment: based on established prices, based on actual costs of labour or material, and based on cost indexes of labour or material named in the contract. The US Bureau of Labor Statistics publishes a guide on using its Producer Price Index (PPI) for this. It advises choosing an index that represents the costs of producing the item, not the price of the item itself, and naming a base period.

The guide’s own example: a base price of $1,000 is tied to an index value of 178.4 in December 2010. In December 2011 the index stands at 187.7, an increase of 5.2%, so the adjusted price becomes $1,052. The guide also warns that PPI data can be revised for up to four months after first publication; the value published four months later is final. The BLS does not recommend for or against price adjustment clauses and does not help write contracts.

• Price rises faster than the index: ask the supplier which other cost changed, or re-source.
• Index falls and the price stays: ask for a decrease, or add an index clause at renewal.
• No index clause and a volatile material: agree one, with a base period, a review frequency and a threshold below which prices do not move.

Which decisions come out of a direct spend analysis?

One price across plants

Extend the best agreement to every plant, or run one sourcing event for the combined volume.

Fewer part variants

Ask engineering to replace near-identical parts with one standard part, so volumes add up.

Second source

Qualify a second supplier for a single-sourced part that drives cost or carries risk. The steps are in strategic sourcing process.

Index clause

Tie the price of material-heavy parts to a named index, so increases and decreases both pass through.

Design or process change

Work with the supplier on cost reduction: a lighter casting, a cheaper finish, a wider tolerance where function allows.

Supplier in another country

Compare landed cost (the price plus freight, duties and other costs to deliver the part), not unit price, before moving a part abroad; see global strategic sourcing.

A lower price is not a saving if quality or delivery suffers. A cheaper seal that leaks raises warranty costs that the analysis of purchase prices will not show. Check each decision with quality: the supplier quality management guide explains how parts are approved and suppliers monitored. Decisions are then grouped by category in the category plan described in category management.

Tools range from a spreadsheet for one product line to dedicated software that cleans part data and reads bills of materials. How to choose one is covered in spend analysis software.