What Procurement Transformation Actually Looks Like

Lessons from Sonoco, Veolia, and SDI on data, adoption, scale, and what happens when a transformation plan meets reality.

Procurement transformation always looks cleaner on a slide.

Find the opportunity. Build the business case. Implement the technology. Drive adoption. Capture the savings. Scale.

In practice, it rarely happens in that order.

At SOAR26, procurement leaders from Sonoco, Veolia, and SDI shared three very different transformation stories. They came from different organizations, started from different places, and are at different stages of their journeys.

But there was a common thread across all three: the original plan is only the starting point.

Real transformation exposes things you couldn't see before. Messier data than expected. Buying behaviors that don't match the documented process. Workarounds you didn't know existed. Assumptions that turn out to be wrong.

The organizations that make progress aren't necessarily the ones that avoid those discoveries. They're the ones that learn from them.

Here are five lessons that stood out.

1. Don't Wait for Perfect Data to Start

“Clean your data first” sounds like sensible transformation advice.

Sonoco discovered it wasn't quite that simple.

As the company began modernizing tactical sourcing across its highly decentralized plant network, the team assumed cleaning its item data would need to happen before the new process could scale.

Instead, the process itself started improving the data.

To competitively source an item, buyers needed an accurate description, manufacturer information, part number, and unit of measure. Running real sourcing events forced those issues to be corrected transaction by transaction. Rather than completing one massive cleanup project before getting started, Sonoco began improving its data through the work itself.

SDI's experience shows what becomes possible as that foundation matures.

Its MRO model starts by understanding what customers actually buy, cleaning and enriching that information, and turning it into something procurement can use for category strategy, supplier selection, catalogs, and inventory decisions.

The lesson isn't that data quality doesn't matter. Quite the opposite.

You don't always need perfect data before transformation begins. You need a process that continuously makes the data better.

That's particularly important as procurement moves toward more AI and automation. AI can do increasingly sophisticated work with good context. Bad inputs simply allow confusion to move faster.

2. Design Around How People Actually Buy

A business case describes how a process should work. Transformation reveals how it actually works.

Veolia experienced that firsthand.

The team identified $26.3 million in potential opportunity and initially expected roughly $2–3 million in savings. The model looked compelling: intercept requisitions, route the appropriate activity through sourcing, and capture the opportunity.

Then reality showed up.

As Veolia dug deeper, it discovered that roughly half of the activity it was looking at was tied to purchases that had effectively already happened. People were buying first and creating the necessary records afterward. The opportunity on paper wasn't the same as the opportunity procurement could actually influence.

That could have been interpreted as a failed implementation. Instead, Veolia treated the new visibility as information.

The team changed its approach, moved sourcing capability closer to the locations where purchasing was actually happening, and began building from there.

There's an important lesson in that.

Sometimes transformation doesn't immediately fix the process. First, it shows you what the process really is.

That visibility has value of its own, because you can't redesign a workflow you don't understand.

3. Adoption Isn't Something You Launch

All three stories challenged another common transformation assumption: if the business case is compelling enough, people will adopt the new process.

They won't. At least not automatically.

Sonoco had savings potential, better visibility, and a simpler way for plant buyers to test the market. But the team still saw the quiet workarounds that happen whenever established behaviors are disrupted.

So adoption became an operating discipline.

Executive leadership reinforced expectations. High-performing plant buyers became change champions. The procurement team visited facilities, listened to buyers, monitored activity, and put performance data in front of plant leadership regularly.

Veolia's experience was even more direct.

Cindy Solomon described meeting with participating teams every week and seeing continued engagement. When those meetings stopped, usage dropped. She restarted them.

SDI approached adoption similarly, but at greater scale. The organization didn't attempt to change every behavior and supplier relationship on day one. It brought users and suppliers into the new process, made their activity visible, and then used that visibility to address where the process was being bypassed.

The takeaway is more specific than “change management matters.”

Adoption isn't an implementation milestone. It's something you operate.

You measure it. Reinforce it. Listen to the people doing the work. Correct behaviors. And keep doing that long after go-live.

4. Automate the Right Process, Not Just the Existing One

It's tempting to look at a manual procurement process and immediately ask how AI or automation can make it faster.

But if the underlying workflow is inefficient, automation can simply help you execute the wrong process more efficiently.

SDI's transformation is a good example.

Before modernizing its purchasing operation, 80 buyers were effectively operating like independent franchises. Quotes arrived through emails, phone calls, online forms, and even notes. Buyers then manually entered information into the ERP, creating downstream errors and rework when invoices didn't match purchase orders.

One of SDI's clearest transformation goals wasn't simply “save more.”

It was to eliminate unnecessary touches.

Instead of moving manual work from one interface to another, SDI focused on integration so information could flow from quoting through award and into its ERP without buyers repeatedly re-entering it.

The results show what happens when that operating model scales. SDI moved from roughly 1,200 PO lines per buyer per quarter to more than 9,000, while its business grew without corresponding purchasing headcount growth.

Sonoco offers another version of the same lesson.

Not every purchase should follow the same sourcing strategy. Some categories benefit from long-term supplier relationships and strategic agreements. Others are volatile, comparable, and switchable enough that testing the market at the time of purchase makes more sense.

Its model isn't about automating everything the same way. It's about applying the right sourcing motion to the right type of spend.

Before asking what can be automated, get clear about what the process should actually be.

5. The Best Transformation Creates the Next Opportunity

Perhaps the most interesting thing about these three stories is that none ends with “and then we were done.”

Sonoco's tactical sourcing activity began revealing patterns in its transactional spend. Recurring activity could signal where the next strategic sourcing event should happen. Tactical sourcing wasn't replacing strategic sourcing—it was feeding it.

Veolia's journey has taken a different path than originally planned, but the organization now has much greater visibility into its purchasing behavior and is continuing to evolve its model based on what it has learned.

And SDI shows what a more mature version of the journey can look like.

It has grown from fewer than one million PO lines annually to more than 2.4 million and reduced its supplier base from more than 6,000 to 2,879. With more standardized processes, integrated technology, and better visibility in place, the organization is now looking toward AI-driven supplier insights, exception management, and further optimization as its next phase.

That's an important perspective as procurement thinks about AI.

You don't arrive at autonomous or agentic procurement by simply switching it on.

You build toward it.

Better processes create better data. Better data creates visibility. Visibility allows you to standardize and automate. Automation creates capacity. And that foundation gives AI something meaningful to work with.

Transformation Doesn't Have to Follow the Original Plan

There isn't one blueprint for procurement transformation.

Sonoco found that deployment could help solve the data problem it thought needed to be solved first.

Veolia discovered that its initial opportunity looked different once it had visibility into actual buying behavior—and adapted.

SDI spent years building the process, integration, supplier participation, and operating discipline necessary to scale before moving into its next phase.

Different journeys. Different maturity levels. Different lessons.

But they point to the same conclusion:

The most successful procurement transformations aren't necessarily the ones that perfectly execute the original plan. They're the ones that create enough visibility to learn, adapt, and keep moving.

And as AI changes what's possible in procurement, that ability to learn may become even more important.

Because the technology will continue to change.

The harder—and more valuable—work is building an organization that can change with it.

Hear the Stories Behind the Lessons

Watch the SOAR26 customer sessions on demand to hear leaders from Sonoco, Veolia, and SDI share the decisions, surprises, challenges, and lessons behind their procurement transformations.

Watch SOAR26 on demand

One registration gives you access to the complete SOAR26 on-demand session library.