Daniel Perkins shared something this week worth sitting with — an account, from Kraft Heinz’s new CPO at DPW’s Recode event, of what a digital procurement journey actually looks like from the inside. As Daniel put it, it was one of the most honest accounts he’d heard on a stage in a long time. He’s right, and three moments stood out.
The 2022 state: local sites empowered to buy their own tools and run their own playbooks. It felt like agility. It was chaos — disconnected systems, duplicated spend. The hardest question wasn’t technical; it was “who owns supplier data?” — the same supplier standing up three times across three ERPs, three sets of terms, three deals. And the line I keep returning to: you can’t train an agent on a process that exists eighty-five different ways. So before deploying AI, they’re writing the playbook. Standardize first, then automate.
Credit where it’s due — to Daniel for the clarity of the write-up, and to the CPO for the candor. Honesty on a stage is rare. But I have to tell you where my mind went, because it went somewhere specific: I have watched this exact conversation happen at this exact company before — seventeen years ago, when it was still Kraft.
In September 2009 I published a piece arguing that Kraft — the company that would later become Kraft Heinz — had bought into what I called the mirage of vendor rationalization. Kraft had announced a sweeping reduction of its supplier base, savings projected up front. My argument wasn’t that consolidation is bad. It was that consolidation without governance produces the opposite of value — that you can reduce your supplier count and increase your complexity, centralize your systems and keep your chaos. The distinction was between consolidation and governed consolidation. Seventeen years later, Kraft Heinz’s CPO is describing that exact distinction from the inside: not “we centralized,” but “we built a strong center — one strategy, one data backbone — and governed it.” Same company. Same lesson. Second era.
And the year before that, in 2008, answering a reader’s question about procurement in mergers, I’d written something plainer still: no amount of data or spend-management technology will make a difference if the organization is found wanting in how it engages and governs itself — and if that weren’t true, 85% of these initiatives wouldn’t fail. The question the CPO says was hardest — who owns supplier data? — is a decision-rights question. It was the 2008 question. It has not changed.
Here is the part that matters, and I want to be careful how I say it, because the weak version is “I predicted it.” I didn’t predict Kraft Heinz. Prediction isn’t the point, and it was never the value.
The value is that the pattern has been the constant. In 2008 the technology in question was ERP. In 2026 it’s AI agents. The technology changed completely; the determining variable didn’t move an inch. Both eras arrive at the same structural truth — the operating logic has to be in place before the technology can deliver on it — and the CPO stated it more cleanly than any framework could: you can’t train an agent on a process that exists eighty-five different ways. Technology-last isn’t a slogan. It’s what the record shows, across two eras, at one company — the constant I’ve come to call Invariant Physics™: technologies keep progressing; the requirement that the operating logic be in place first has not once failed to hold.
That is what an archive is for. Not to say “I said it” — anyone can write a retrospective once the outcome is known. The value is that the observations were made contemporaneously, in public, dated, before the outcome arrived — and that the same pattern keeps reappearing as each new technology becomes the lens everyone views it through. You can dispute an interpretation. You cannot retrofit a date.
So my thanks to Daniel, and to DPW for putting a practitioner on a stage instead of a theory. The account was honest precisely because it was lived. And the reason it sounds familiar isn’t that anyone is repeating anyone — it’s that the foundations don’t change. The technology does. The physics of implementation stay remarkably still while everything above them keeps moving.
Truth Is Believing. Accuracy Is Knowing. Outcome Is Proof.™
-30-
Related
The Foundations Don’t Change: From Kraft to Kraft Heinz
Posted on July 6, 2026
0
Daniel Perkins shared something this week worth sitting with — an account, from Kraft Heinz’s new CPO at DPW’s Recode event, of what a digital procurement journey actually looks like from the inside. As Daniel put it, it was one of the most honest accounts he’d heard on a stage in a long time. He’s right, and three moments stood out.
The 2022 state: local sites empowered to buy their own tools and run their own playbooks. It felt like agility. It was chaos — disconnected systems, duplicated spend. The hardest question wasn’t technical; it was “who owns supplier data?” — the same supplier standing up three times across three ERPs, three sets of terms, three deals. And the line I keep returning to: you can’t train an agent on a process that exists eighty-five different ways. So before deploying AI, they’re writing the playbook. Standardize first, then automate.
Credit where it’s due — to Daniel for the clarity of the write-up, and to the CPO for the candor. Honesty on a stage is rare. But I have to tell you where my mind went, because it went somewhere specific: I have watched this exact conversation happen at this exact company before — seventeen years ago, when it was still Kraft.
In September 2009 I published a piece arguing that Kraft — the company that would later become Kraft Heinz — had bought into what I called the mirage of vendor rationalization. Kraft had announced a sweeping reduction of its supplier base, savings projected up front. My argument wasn’t that consolidation is bad. It was that consolidation without governance produces the opposite of value — that you can reduce your supplier count and increase your complexity, centralize your systems and keep your chaos. The distinction was between consolidation and governed consolidation. Seventeen years later, Kraft Heinz’s CPO is describing that exact distinction from the inside: not “we centralized,” but “we built a strong center — one strategy, one data backbone — and governed it.” Same company. Same lesson. Second era.
And the year before that, in 2008, answering a reader’s question about procurement in mergers, I’d written something plainer still: no amount of data or spend-management technology will make a difference if the organization is found wanting in how it engages and governs itself — and if that weren’t true, 85% of these initiatives wouldn’t fail. The question the CPO says was hardest — who owns supplier data? — is a decision-rights question. It was the 2008 question. It has not changed.
Here is the part that matters, and I want to be careful how I say it, because the weak version is “I predicted it.” I didn’t predict Kraft Heinz. Prediction isn’t the point, and it was never the value.
The value is that the pattern has been the constant. In 2008 the technology in question was ERP. In 2026 it’s AI agents. The technology changed completely; the determining variable didn’t move an inch. Both eras arrive at the same structural truth — the operating logic has to be in place before the technology can deliver on it — and the CPO stated it more cleanly than any framework could: you can’t train an agent on a process that exists eighty-five different ways. Technology-last isn’t a slogan. It’s what the record shows, across two eras, at one company — the constant I’ve come to call Invariant Physics™: technologies keep progressing; the requirement that the operating logic be in place first has not once failed to hold.
That is what an archive is for. Not to say “I said it” — anyone can write a retrospective once the outcome is known. The value is that the observations were made contemporaneously, in public, dated, before the outcome arrived — and that the same pattern keeps reappearing as each new technology becomes the lens everyone views it through. You can dispute an interpretation. You cannot retrofit a date.
So my thanks to Daniel, and to DPW for putting a practitioner on a stage instead of a theory. The account was honest precisely because it was lived. And the reason it sounds familiar isn’t that anyone is repeating anyone — it’s that the foundations don’t change. The technology does. The physics of implementation stay remarkably still while everything above them keeps moving.
Truth Is Believing. Accuracy Is Knowing. Outcome Is Proof.™
-30-
Share this:
Like this:
Related