Note: A portion of the content listed above is only available in the email version of this newsletter. Don’t miss out! Sign up for free to get the next edition.
Table of Contents
Introduction
Nearly every procurement technology selection exercise I've supported started with demos. Procurement teams reached out to vendors they thought might help and started looking at solutions.
But before the demos, it should have started with a spreadsheet (blasphemy, I know…).
You’ve been here: the vendor's AI agent drafts an RFQ in eight seconds in the demo. Everyone in the room nods. A few months later the tool is live, and nobody can say whether it moved the needle, because nobody wrote down where the needle was in the first place.
The fix is a tool I rarely see procurement teams build: a spend profile (not to be confused with a spend cube). It maps what you spend, how often you transact, and which purchasing channels you use to do it, spend category by spend category. Build one and your highest-ROI technology opportunities show themselves clearly.
This article covers what a spend profile is, how to build one, and how to read it before going back to asking for the demos.
What is a spend profile?
A spend profile is a breakdown of your organization's spend and transaction volume by spend category and purchasing channel. It maps how you buy goods and services in each category today and how you want to buy it in the future to generate ROI.
Let’s take MRO parts as an example. A spend profile tells you how much you spend on MRO (like a spend cube). But, it also shows every route those dollars take: how much flows through vendor-managed inventory (VMI), how much through punch out catalogs, and how much through quotes your buyers chase by email.
A standard spend analysis rarely goes that deep on the second part.
Spend profile vs. spend cube
A spend cube is a multidimensional model of your spend data, usually sliced by category, supplier, business unit, and time. It is primarily used to help procurement achieve its traditional objectives (e.g. savings, cost avoidance, ESG goals, etc).
A spend profile overlays your purchasing channels on that data, at the level of detail technology decisions need, and adds how you want each purchasing channel to evolve. It’s a procurement digitalization tool.
Let's be fair to the cube. A well-built spend cube already holds transaction data from AP, P2P systems, P-cards, etc. Many count POs and invoices alongside dollars. The better ones even flag PO-backed versus non-PO spend, or P-card versus invoice.
The gap is purchasing channel granularity. "PO versus non-PO" is a compliance view. A technology decision needs to know which kind of PO:
Did MRP generate it against a fixed source of supply, or did a buyer have to chase quotes?
Was it a catalog item linked to your material master, or a non-stock catalog item?
Was it a release against a VMI framework order, or a free text order?
Those variations change from one company to the next, and they don’t live in a single ERP field. You usually reconstruct them from a combination of business rules related to document type, item category, source list, catalog flags, and account assignment (at least in SAP systems).
That's why off-the-shelf spend cubes don't carry them to the level of granularity needed out-of-the-box: someone has to define the purchasing channel taxonomy for your business first.
Picture two categories with the same $10M in annual spend, both 100% PO-backed. In the cube, they look the same. The first runs on a few dozen releases against framework agreements. The second runs on tens of thousands of manually quoted POs. Only the spend profile shows you that one of them is screaming for digitally enabled capabilities to drive additional efficiency and savings.
Spend cube | Spend profile | |
|---|---|---|
Core dimensions | Category, supplier, business unit, time | Category and purchasing channel, often by site |
Purchasing channel | Coarse at best (PO, non-PO, P-card) | Operational level, defined for your business |
Measures | Spend and transaction counts | Spend and transaction volume by purchasing channel, always paired |
Time frame | Historical and planned (at least in market leading solutions) | Same as your spend cube |
Best for | Sourcing, category strategy, compliance (Pure Procurement... Pun intended) | Technology roadmap, business case, deployment scope and sequence |
The spend profile sits on top of your spend cube. If your cube is in good shape, you already have most of the data you need to make good technology decisions.
The purchasing channels to map in your spend profile
A purchasing channel (also called a buying channel) is the end-to-end route a requirement takes, from the moment someone needs something to the moment the supplier gets paid. (The full directory lives in Purchasing Channels Explained.)
For an MRO category, the channels in your spend profile will look something like:
Vendor-managed inventory (VMI). The supplier manages stock on your site (whether by refilling bins or with vending machines). You invoice against a framework PO.
Non-stock vendor catalogs. Requesters buy from supplier catalogs. The item doesn't exist in your inventory management system or ERP. It is consumed to a cost center at purchase.
Stock vendor catalogs. Catalog items linked to your own material numbers, so purchases can feed inventory and MRP calculations (or go straight to expense).
MRP-based material master purchases on contract. MRP generates the requisition, and a fixed source of supply turns it into a PO.
MRP-based material master purchases requiring quotes. MRP generates the requisition, but the source of supply varies, so a buyer has to go get quotes.
Free text orders. The requester types out the need by hand. (Your data quality's worst enemy…)
Everything else. PCards, non-PO invoices, and the channels nobody approved but people use anyway.
Your exhaustive list of channels will change by category. Services bring service POs and contract invoices. Software brings contract invoices. Capture every route your money takes, including the ones your policy pretends don't exist.
Track dollars and transactions for every channel
Spend tells you where your negotiating leverage sits. Transaction volume tells you where your people spend their days.
You need both numbers:
Low spend with high transaction volume points to labor problems. Labor problems are what digitally enabled capabilities solve.
High spend with low transaction volume is usually a sourcing, a contracting or in intelligence problem. A slicker workflow won't fix it.
High spend and high volume are both interesting places to start looking, but with different mindsets.
How to build a spend profile in 5 steps
You don't need new software for this. Excel works fine. (Yes, the same Excel you're trying to get rid of… 😅)
Pull 12 to 24 months of spend by category. Use your commodity code taxonomy at level 2 or 3. Level 1 is too coarse to show channel differences.
Tag every transaction with its purchasing channel. PO document types, item categories, account assignments, and catalog flags in your ERP get you most of the way. Expect some manual cleanup and adding columns to make your profile your own. Ensure you capture the business rules you used to split the spend by channel.
Count transactions per spend category and channel. PO lines or invoice lines both work depending on context.
Add a target channel column. For each row, write down how you realistically want to buy it in the future. This column is where your roadmap starts (and where you need to understand the "art of the possible" to get it right).
This is when attending conferences and watching demos becomes most useful… Because you have holes you’re trying to fill instead of aimlessly looking for ways to “leverage AI”…
Adjust for planned spend. More on this below.
Here's one category from an illustrative MRO spend profile (the numbers are made up for the example, not a benchmark):
Purchasing channel | Annual spend | Annual PO lines | Target channel |
|---|---|---|---|
VMI (invoice against framework PO) | $6M | 1,200 | Keep |
Non-stock vendor catalogs | $4M | 18,000 | Keep and expand |
Stock vendor catalogs | $2M | 6,000 | Expand |
MRP material master, on contract | $14M | 22,000 | Keep |
MRP material master, requiring quotes | $9M | 41,000 | Shift to stock catalogs or automate quoting |
Free text orders | $5M | 12,000 | Shift to catalogs |
Repeat for every spend category that matters. If the complexity requires it, add a column for business units or sites, so you know where each channel runs.
Adjust for the spend that's coming
At this point, your spend profile is probably built on historical spend. History doesn't know about the plant you're opening next year or the upcoming M&A activity.
A profile based only on the past points your technology roadmap at last year's problems. Before you size a business case, add what you know is coming: new sites, capex programs, acquisitions, categories you plan to insource or outsource.
Reading your spend profile: where procurement technology ROI hides
Once the table exists, the opportunities are hard to miss.
Example: the MRO quote treadmill
Look at the table above again. The biggest transaction volume sits in MRP-based material master purchases requiring quotes. Every one of those 41,000 lines means a buyer emailing suppliers, comparing answers in a spreadsheet, and keying the winner into the ERP (or forgoing to do any of that because you don’t have the capacity….).
That's the row where technology moves the needle for MRO because prices are so volatile, and you have two ways to attack it.
The first is to shift the volume to stock vendor catalogs. Link supplier catalog items to your material numbers and MRP-driven demand converts into catalog orders at current pricing, with no quote needed.
The second is to automate the quote. An automated "3 bids and a buy" capability sends the RFQ to qualified suppliers, collects the bids, and turns the winner into a PO with minimal buyer involvement.
Either way, you know which capability to look for and which 41,000 lines your business case rests on. Compare that to "we need AI for sourcing."
Channels cut across categories
The next thing that becomes evident as you stare at your profile is that purchasing channels are shared. The automated quoting you build for MRO can also serve packaging, lab supplies, or IT peripherals if those categories rely on the same quote-heavy channel.
Read your spend profile in both directions. The category view shows where the biggest opportunity sits, and the channel view shows how many spend categories a single capability can reach.
One capability deployed across five spend categories makes for a far easier business case than five tools for five categories.
Categories give you your minimum viable scope
The category view also keeps you honest about where to start. You don't roll a new capability out to every category on day one.
You pick the category where the profile shows the biggest return and where the stakeholders are ready to adopt. That's your minimum viable scope: the smallest deployment that lets you test, fail safely, and prove value before you scale.
Where your procurement view meets your technology view
The spend profile works because it puts two views of your business in a single table.
The first is the pure procurement view (pun intended): categories, spend, suppliers, sites, and strategy.
The second is the technology view. Every purchasing channel is a business process (request, approve, source, order, receive, pay) that runs on systems (ERP, P2P suite, marketplace, sourcing tool, intake layer), fed by data (material masters, catalog content, contracts, sources of supply, pricing).
Change the process, the system, or the data and you've changed the channel. That makes every technology decision a decision about which channels you'll open, close, or improve, and for which categories.
Your spend profile shows where those decisions meet your money.
Mining the "art of the possible"
Of course, you can only write down target channels you know exist, and you don't know what you don't know…
Before reading this article, you may not have known you can buy from vendor catalogs while linking their items to your material masters in the back end, without anyone maintaining the mapping by hand. If you didn't know, that option never made it into your target channel column.
That should change how you attend conferences, webinars, and virtual events.
Right now, without a plan, attending conferences is more confusing than anything else because every ProcureTech software vendor sounds the same.
When attending events, you should go mining for new purchasing channels to add to your target state, new capabilities inside the channels you already run (automated quoting, catalog-to-material-master linking), and value levers you weren't measuring.
Your spend profile can also include "purchasing channel adjacent" capabilities, for example: faster supplier onboarding and shorter SRM cycle times count as value even though they never show up as a P2P purchasing channel (although they might slow one down).
Everything you find goes into your spend profile, where you can test it against real volume.
"AI agent" or not, ask the same three questions
If a vendor wants to call it an AI agent, fine. The label doesn't change how you evaluate it. Ask:
Which process steps does it execute? Name them. "Collects three bids and creates the PO" is a process step. "Transforms sourcing" isn't.
What data does it consume? Material masters, catalog content, supplier records, contracts. Are the data sources internal (yours) or external (purpose-built data sets). If the data isn't there, the agent won't deliver.
What new business capability does it give you? Then go back to your spend profile. Does that capability touch enough volume, in enough categories, to generate outsized returns in your business?
If the answers don't map to a row in your spend profile, the pitch is entertainment.
Mistakes that make a spend profile useless
Stopping at spend. Without transaction volume or adjacent processes, you'll chase big-dollar categories where digitally enabled capabilities have little to offer.
Mapping channels too broadly. "PO" isn't a channel. There are at least 11 buying channels hiding behind that label. An MRP-based PO on contract and one requiring quotes are two different problems with two different fixes.
Leaving the target channel column empty. Without it, the profile describes last year and tells you nothing about what to fix. The hard work of digitalization is figuring out what goes in this column.
Treating history as the future. Adjust for planned spend before you size anything.
Building it once. Refresh it as things evolve, and every time your category strategy or system landscape changes.
Start with your spend profile, then look at technology
A spend profile shows how you buy each spend category and how much work each route takes. Add your target channels and the highest-ROI technology opportunities become obvious, along with the category to start in.
After that, every demo gets measured against your numbers instead of the vendor's script.
Ready to map your first category? Get the Spend Profile Template and start this week.
🔒 PREMIUM ONLY: YOUR FREE COPY
Premium members get the Spend Profile Template for $0. Upgrade and the checkout code shows up right here, in this box.