What Is a Spend Cube? How Spend Cube Analysis Works
A spend cube organizes spend by supplier, category and business unit so you can slice it any way you need. Here is how it works, what it is good for and where it falls short.
By Aditya Chavali
If you have been in procurement for a while, someone has shown you a spend cube. Usually in a consulting deck, with three axes and a lot of colored blocks.
The idea is older than most of the tools that use it. It is still the clearest way to think about spend.
What is a spend cube?
A spend cube is a way of organizing spend data along three dimensions at once: who you buy from (suppliers), what you buy (categories), and who is buying (business units, sites or cost centers). Every transaction sits at one point in that cube.
Once spend is organized that way, you can slice it along any face. All spend in one category, across every supplier and site. All spend with one supplier, across every category. Everything one site buys.
Many teams add a fourth dimension: time. That lets you see how each slice changes month over month.
Why the cube matters
The cube is useful because most procurement questions are really questions about two of the three dimensions.
- Category by supplier: how many suppliers do we use for this category, and how is spend split between them?
- Supplier by business unit: which sites buy from this supplier, and do they pay the same price?
- Category by business unit: which sites drive spend in this category, and does one buy very differently from the rest?
Each of those is a view of the cube. Without the cube, each one is a separate spreadsheet built by hand.
What spend cube analysis surfaces
When teams first see their spend organized this way, a few patterns tend to appear.
Fragmentation. One category spread across many suppliers, often because each site chose its own. This is the starting point for supplier consolidation. See What Is Supplier Consolidation?
Price differences. The same item bought from the same supplier at different prices by different sites. That is price variance, and it often has a simple cause. See What Is Price Variance?
Concentration. A few suppliers that the business depends on more than anyone realized.
A long tail. Many suppliers with small amounts of spend each. See What Is Tail Spend?
The cube is only as good as its edges
Here is the catch. A spend cube needs every transaction to land in the right place on each axis.
That means the supplier axis has to be normalized. If one supplier appears under five names, the cube shows five small suppliers. See What Is Vendor Normalization?
The category axis has to be classified consistently. If half the spend sits in "Miscellaneous", half the cube is empty. See What Is Spend Classification?
The business unit axis has to map cleanly across systems, which is harder than it sounds when every acquired company kept its own ERP.
Most failed cube projects did not fail at the analysis. They failed at the edges.
Where the cube falls short
A spend cube shows where money goes. It does not tell you what to do about it.
Knowing that a category has twelve suppliers does not tell you which three to keep, what the switching risk is, or whether a renewal is coming up next month. Those are decisions, and they need more than a slice of data.
That is why I think of the cube as the starting point, not the deliverable.
How SpendCraft uses the cube
SpendCraft builds the edges first. Vendor and item normalization settle who is who. Classification maps every line into your own taxonomy, with proprietary AI plus rules and a person reviewing the result. The result is the Corpus, a governed record of your spend that stays current as new data arrives.
From there, analytics show spend by vendor, category, entity and time. Savings Opportunity Scans surface patterns such as supplier consolidation and price variance as Min to Max ranges scoped to Processed Spend. Caddies take an opportunity from there to a decision you own.
What good looks like
A good spend cube is one nobody thinks about. Category managers slice their spend without asking an analyst. Finance trusts the totals. And when the data changes, the cube changes with it.
If your cube needs rebuilding every year, the problem is not the cube. It is the edges.