How to create a spend taxonomy

Spend data usually arrives as thousands, or millions, of transactions in different formats and with different attributes. Labeling and classifying that data is one of the ways to structure the ocean of records and get it ready for insight generation.

The spend classes and their relationships are referred to as the spend taxonomy: a hierarchical structure you can use to slice and dice all transactions and generate different overviews, the foundation for both insight generation and execution.

The primary questions procurement and supply chain professionals face are deceptively simple:

• Where to start? With years of accumulated spend, how do you choose the baseline for spend analysis?
• What are the best practices? How do you run spend analysis in a timely manner and prevent the project from taking years before it yields results?

Here is our tip, based on our own experience working with many companies and on best practices from the likes of Gartner. It comes down to three moves.

1. Define the purpose of the taxonomy for your organization

This is potentially the most critical decision, and a few factors drive it. The first: does your business require you to exchange or share your spend sources with others? Companies in logistics often sit in that box. If that's you, you should be thinking about a highly standardized taxonomy, so you can easily compare and share your spend analysis with partners, competitors and suppliers across your industry.

There are several standardized schemes out there. The most popular are UNSPSC and eCl@ss, both created with the same goal: the ability to compare and share spend on common terms.

However, if your business is going to use spend analysis only internally, which is where most businesses fall, then the purpose of the taxonomy is to truly represent your products and services, and those your current supplier base serves today. In this scenario you'll build your own tailor-made taxonomy. It carries tremendous advantages over standard schemes, but it also needs real time investment from the firm.

Owned taxonomies are sourcing-based. They're tuned on the complexity, criticality and volume of your business by experts who have worked in it for years, usually inside the strategic sourcing group, but sometimes close functions like Advanced Operations or R&D. No wonder spend analysis fueled by an owned taxonomy can deliver much deeper descriptive and prescriptive insight, the kind that's understood and respected across the business.

In both cases, the taxonomy acts as a fingerprint for spend analysis. For all but a minority of businesses, the more specific it is to you and to your suppliers, the higher the potential value spend analysis can capture.

2. Find a baseline aligned with the purpose

At first this is overwhelming, especially when there's no taxonomy in place. If that's the case, our suggestion is to start from one of the standard schemes, like UNSPSC, and tweak it toward your sourced products and services. That takes the initial burden off your teams and helps them structurally scope and begin tuning the spend tree.

It is highly recommended to assign this responsibility to a single team or group, so all efforts stay coordinated and protected, with nothing lost along the way.

The process of tweaking the tree into your organization's spend signature can be lengthy, and it requires a broad range of people to share years of domain knowledge. That's exactly why the continuity of this activity, and visible leadership support, is most essential. (For more on sustaining it, read our companion piece on continuous value generation with spend analysis.)

3. Pick a software tool to allocate the taxonomy to transactions

The previous steps draw a picture of your particular spend profile. The logical next step is to allocate your spend transactions into the boxes you created, your taxonomy. This is straightforward if you have very few products, services and vendors. But for most companies, with hundreds or thousands of products and vendors and new transactions every day, allocation becomes a genuinely complicated operation. So it's worth evaluating this decision carefully, right from the beginning, and revisiting it every couple of years.

At Mithra-AI we're passionate about facilitating exactly this, in a novel way, not only for larger enterprises but for smaller teams and even individuals responsible for a single spend portfolio. The same engine that classifies a global manufacturer's spend can put a clean taxonomy within reach of a one-person category team.

Where to go from here

Building your first taxonomy, or rethinking the one you have? Have a chat with us, and let us inspire you.