Spend Analysis in 3 Lessons: Visualizing Spend Data (Lesson 2)

Analytics Rawpixel.com/Adobe Stock

This three part series (see Part 1) covers three spend analysis lessons. I continue today with the opportunity and challenges associated with the visual display of quantitative information.

Lesson 2: The Visualization of Spend Data and Quantitative Information

I have a confession to make. The early spend analysis tools I was involved with at FreeMarkets, and later reviewed as an analyst at Spend Matters, would have gotten a “F” in “Data Visualization 101” for their ability to enable business users to work with and collaborate around spend data.

Granted, some were fabulous at enabling the power user to manipulate data and pursue the creative destruction of as many spend cubes as possible. But generally speaking, these solutions were aimed at one or two people in a company tasked with running reports and hunting for savings, cost management and working capital opportunities from their own cubes (pun intended). In short, they rarely served as a broader platform for collaboration and coordination with different business stakeholders.

McKinsey’s Gene Zelazny (the grandfather and grand poobah of good charting/graphics taste) would have had a heart attack with the reports most of these first generation tools spat out. Incidentally, Zelazny, who started his career in design and consulting over 50 years ago, is as relevant as ever in today’s digital age — and his books most deeply influenced my own design and presentation skills, especially those involving data (sometimes I wish my colleagues had read them!). If you’ve not read his work, run, don’t walk, to Amazon, and be prepared to learn everything you’ve been doing wrong with graphics, charts and underlying datasets, and how best to persuade an executive audience with them.

But I digress. In spend analysis today, the visual display of information is as important as ever. A picture can indeed tell a thousand words.

For example, a base level overview of spend data should center on overall metrics and health — call it a procurement or spend scorecard if you will. This type of data should show core information and show it well in the right charts, dashboards and underlying drillable datasets. As I’ve written before, basic information shared in a manner that is easy for users to drill into should include answers to questions such as:

  1. What is my total spend?
  2. Who are my largest suppliers (parts, spend, categories)?
  3. What are my largest spend segments?
  4. What parts are growing in total spend? Shrinking?
  5. What parts have the largest price inflation (over a given period of time)?
  6. Am I paying more from one supplier than another for part X?
  7. Where can I quickly cut costs by taking action?

Other basic (let’s call them “foundational plus”) questions to answer can fall into the other areas that I’ve also noted in past papers:

  1. Contract Management
  • Am I paying the contract price?
  • How much am I buying off-contract?
  • Why? (e.g., non-compliance, expediting, etc.)
  1. Buyer Management
  • Who is managing the most items/spend?
  • Who is managing this contract?
  1. Spend Disbursement
  • Percent of spend from Low Cost Countries?
  • Percent of spend from MWE?
  1. Time Variance
  • What has changed over the past year?
  • Why has the variance occurred (e.g., restocking vs. demand-driven replenishment based on a pull model)?
  1. Management, Leverage and Planning
  • Who should own commodity X?
  • How can we best leverage similar items (but potentially with different SKU/part numbers and suppliers) across operating units?
  • What items will be growing?

Beyond this, foundational performance data should include areas such as drill downs on compliance and P2P insights (e.g., cycle times, payment discount uptake, payment terms/working capital impact, etc.). And that’s just a start.

But it’s not just the type of information you share: It’s how you share it. In fact, if you just tackled the above areas with a user interface that anyone could walk up to and use, you’d be ahead of most of your peers in the market. Yet the next level, as we might describe it, of spend analytics and spend visualization is incrementally more valuable than the first.

Toward Advanced Insights and Supply Analytics

While transitioning from spend to broader supply analytics can (and should) be the longer-term focus of procurement managing larger and larger data sets, there are a number of interim steps to take, nearly all of which include taking existing data sets and adding in new types of views/visualizations or related information to extend, often dramatically, what basic data can tell us.

Here are some of my favorites (shared recently by a friend and colleague Stefan Dent from Simfoni Analytics):

  • Category Strategy Views — At the most basic, this can include plagiarizing (with the best of intent) a drillable dashboard version of Peter Kraljic’s famous 2/2 matrix of “non-critical, leverage, bottleneck and strategic” spend categories and opportunities, represented with a bubble overlay with sizes of circles in each matrix sized based on whatever additional dimension an analyst wants to emphasize (e.g., size of opportunity (savings), ease of implementation, etc.)
  • Second-Level “Spend Overviews” — Another drillable dashboard example that shows not only the basic spend data (outlined above) but additional insights and fields such as geographic mappings with tier one (and even multi-tier supplier spend data, if available), spend fragmentation, heat maps of categories, etc.
  • 80/20 Analysis — This is a great consultant’s view into data that essentially provides a dashboard of spend based on the classic 80/20 or Pareto rule. It can be as granular as a spend analyst desires — e.g., by geography, P&L, etc. — and should include spend by category, cost center, number of vendors, total spend value, etc.
  • Should Cost Monitoring — My friends who do should cost modeling and cost tear downs for a living would laugh at the exceptionally rudimentary breakdowns in even the better spend analysis tools today, but I say pish-posh to that. Even basic details can be helpful (and we’re not all engineers!). Breaking down spend by underlying commodity components tied to indexes, supplier value-add, logistics/transportation components and other areas adds value, even if it’s high-level
  • Risk/Reputation Analytics — This can include sentiment and other analysis on suppliers based on any range of factors (financial, quality, labor, etc.)
  • Operations/Financial metrics — This can include views into working capital-related metrics (beyond what’s outlined above) including overdue POs, invoices, goods received but not invoiced, goods invoiced but not received, etc.

Below, we’ve included some example screen “grabs” that showcase some of the above examples — and others.

Regardless of which types of dashboards and visualizations you decide to use, keep in mind that a top-level spend dashboard in a spend analytics application should approximate — albeit with more detail — what a good McKinsey slide looks like (think: mutually exclusive, clearly exhaustive, or MECE). Ideally, detailed information should be a click away so the user is not overwhelmed with the first page view. This is what makes data actionable for users.

Of course, the power-user has different needs (and requires different information) than a business user. Which will bring us to the third installment of this research series.

Share on Procurious

First Voice

  1. Don Hoeppner:


    Those early FreeMarkets and other tools may not have been very good compared to the latest tools, but they were a lot better than going through Accounts Payable file cabinets and making copies of suppliers invoices … and then guessing at the total volume of a category when the company bought several categories from the same supplier … or the same commodity from several suppliers!!! And that was better than not looking at AP at all and just doing 3 bids for each requisition when Procurement was all about generating POs rather than Adding Value to a business!! Thanks for new technology … when it works!!!

    Cheers, Don

Discuss this:

Your email address will not be published. Required fields are marked *