The determining variable doesn’t change. The cases keep confirming it — including the ones that don’t make the cut.
A second analyst group recently published its own version of the Gartner Supply Chain pyramid, and Logitech was on it. That caught my attention, because I had written about Logitech’s procurement story back in early 2020 — a “results-first” approach in which the company brought its indirect procurement function back in-house after an outsourcing arrangement that cost half a million dollars a year and never delivered. It was, on its face, a readiness-before-technology story.
The Supply Chain Now webcast, in which the Supply Chains to Admire winners were announced, and in which Lora Cecere and Dave Winstone talked through the methodology.
When I say Logitech reminds me of Virginia, I don’t mean the Virginia most people know — the twenty-plus-year benchmark whose eVA program is cited the world over. I mean Virginia in 2007, when I first wrote about it: six years into its arc, showing the right signals, doing the process-first work, but with the long proof still ahead of it and nothing yet certain. That is where Logitech sits now. And there is one factual difference worth naming, because it cuts in Logitech’s favor without proving anything: Virginia’s early years unfolded in relatively stable conditions. Logitech’s consolidated through 2020 — into the most violent supply-chain disruption in modern memory. It faced a global crisis at a stage Virginia’s early years never had to.
So I held it up against the same standard I hold every case to — the one that has decided, for nearly three decades, whether something belongs in the evidence base or stays an interesting anecdote. And here is what I want to be transparent about: Logitech did not clear the bar as a proof case. Not because the pattern wasn’t there, but because the independent, measured outcomes weren’t, and because I couldn’t establish that its 2026 standing was actually continuous with the 2020 work. The story was real. The proof was not complete. So it stays an illustration — and the discipline that keeps an illustration from masquerading as proof is the same discipline that makes everything else in the record trustworthy.
That is the part worth understanding before anything else. A framework you can build on is not one that claims every case that flatters it. It’s one that declines the cases that don’t meet the standard — even when they’d have been useful. The foundation doesn’t get stronger by accumulating illustrations. It gets stronger by holding its line.
What the case did confirm
There was one signal in the Logitech story worth keeping. The executive most associated with its 2020 transformation died in a tragic accident in 2024, and Logitech remained on the list in 2026 regardless. If a result is lodged in one talented individual, losing that person tends to expose it. That this one held is consistent with the entire argument: talented people are necessary, but they are never the determining variable. The determining variable is structural — it lives in how the organization actually operates, not in who happens to be operating it. I’ll hold even that honestly: two years is an early signal, not the two-plus decades Virginia’s eVA program has sustained. But it points exactly where the foundation says it should.
The thing the foundation already contained
Here is what the case brought into focus — and I want to be precise, because it is easy to mistake for the framework changing. It didn’t change. A strand that was always part of it simply came into clearer view.
It would be easy to hear “the technology is rarely the determining variable” and conclude that the provider doesn’t matter. That’s wrong — and it’s worth saying exactly why, because both halves are true at once.
The platform’s capability is not the deciding factor. Virginia and Ontario ran the same Ariba software to opposite outcomes; that settles the question of whether the software decides. It doesn’t.
But a provider’s commitment — its willingness to design for a client’s real operating conditions, and to walk away from a deal that isn’t ready rather than chase the sale — is a genuine contributing factor. The cleanest example I know is a provider that scored at the very top of an independent analyst’s ratings and still walked away from a major account, because the client met them with the most dangerous phrase in this business: “we have always done it this (our) way.” That provider wasn’t at fault. Walking away was the proof of their commitment. They brought everything a provider can bring — and it still wasn’t enough, because the client wasn’t ready. A capable, committed provider is necessary. It is not sufficient.
So a good provider is never the reason a project succeeds — and never a non-factor either. Both statements hold, and they only seem contradictory if you’re still looking for a single hero or a single villain.
You can’t blame the car for the lake
Patrick Marlow, a GenAI champion at Google, put it better than I could, in an exchange we had in 2024: “Imagine buying a new Corvette, running it into the lake, then blaming Chevy for it not being able to float.” The car isn’t the problem. A Corvette is a superb machine. But the machine’s quality was never the determining variable — and that cuts both ways. You can’t blame the technology for the lake, and you can’t credit the technology for the track.
That is worth sitting with, because it’s the part the market keeps getting wrong in both directions. When a deployment fails, the technology takes the blame. When one succeeds, the technology takes the credit. Both are the same error — crediting or blaming the machine for what the driver and the conditions decided. A committed provider can build a superb car, and can even refuse to hand the keys to someone about to drive it into a lake. But what determines the outcome is whether the organization behind the wheel was ready — and whether anyone bothered to find out before the keys changed hands.
Why this should reassure rather than unsettle
I’ll close on the point that matters most right now. In a moment when the technology itself is moving fast enough to make everyone uncertain, the instinct is to look for stability in the tools. That’s looking in the wrong place — the tools are exactly what won’t hold still. The stability is underneath them, in the determining variables that haven’t changed across ERP, e-procurement, cloud, and now agentic AI.
That’s why a case like Logitech is reassuring rather than unsettling, once you see what it’s doing. The framework didn’t bend to absorb it. It tested it, declined to overclaim it, confirmed the determining variable through it, and clarified — without changing — the role the provider plays. That is what a foundation does. It doesn’t get rebuilt every time a new case arrives. It absorbs the case, and stays where it is.
The technologies will keep changing. The determining variable — whether the people behind the wheel are ready, and whether the provider was honest enough to check — will not.
Truth Is Believing. Accuracy Is Knowing.
-30-
Related
In 2020, Logitech Repatriated Procurement Into a Pandemic. By 2026, They Remind Me of Where Virginia Started
Posted on June 28, 2026
0
The determining variable doesn’t change. The cases keep confirming it — including the ones that don’t make the cut.
A second analyst group recently published its own version of the Gartner Supply Chain pyramid, and Logitech was on it. That caught my attention, because I had written about Logitech’s procurement story back in early 2020 — a “results-first” approach in which the company brought its indirect procurement function back in-house after an outsourcing arrangement that cost half a million dollars a year and never delivered. It was, on its face, a readiness-before-technology story.
The Supply Chain Now webcast, in which the Supply Chains to Admire winners were announced, and in which Lora Cecere and Dave Winstone talked through the methodology.
When I say Logitech reminds me of Virginia, I don’t mean the Virginia most people know — the twenty-plus-year benchmark whose eVA program is cited the world over. I mean Virginia in 2007, when I first wrote about it: six years into its arc, showing the right signals, doing the process-first work, but with the long proof still ahead of it and nothing yet certain. That is where Logitech sits now. And there is one factual difference worth naming, because it cuts in Logitech’s favor without proving anything: Virginia’s early years unfolded in relatively stable conditions. Logitech’s consolidated through 2020 — into the most violent supply-chain disruption in modern memory. It faced a global crisis at a stage Virginia’s early years never had to.
So I held it up against the same standard I hold every case to — the one that has decided, for nearly three decades, whether something belongs in the evidence base or stays an interesting anecdote. And here is what I want to be transparent about: Logitech did not clear the bar as a proof case. Not because the pattern wasn’t there, but because the independent, measured outcomes weren’t, and because I couldn’t establish that its 2026 standing was actually continuous with the 2020 work. The story was real. The proof was not complete. So it stays an illustration — and the discipline that keeps an illustration from masquerading as proof is the same discipline that makes everything else in the record trustworthy.
That is the part worth understanding before anything else. A framework you can build on is not one that claims every case that flatters it. It’s one that declines the cases that don’t meet the standard — even when they’d have been useful. The foundation doesn’t get stronger by accumulating illustrations. It gets stronger by holding its line.
What the case did confirm
There was one signal in the Logitech story worth keeping. The executive most associated with its 2020 transformation died in a tragic accident in 2024, and Logitech remained on the list in 2026 regardless. If a result is lodged in one talented individual, losing that person tends to expose it. That this one held is consistent with the entire argument: talented people are necessary, but they are never the determining variable. The determining variable is structural — it lives in how the organization actually operates, not in who happens to be operating it. I’ll hold even that honestly: two years is an early signal, not the two-plus decades Virginia’s eVA program has sustained. But it points exactly where the foundation says it should.
The thing the foundation already contained
Here is what the case brought into focus — and I want to be precise, because it is easy to mistake for the framework changing. It didn’t change. A strand that was always part of it simply came into clearer view.
It would be easy to hear “the technology is rarely the determining variable” and conclude that the provider doesn’t matter. That’s wrong — and it’s worth saying exactly why, because both halves are true at once.
The platform’s capability is not the deciding factor. Virginia and Ontario ran the same Ariba software to opposite outcomes; that settles the question of whether the software decides. It doesn’t.
But a provider’s commitment — its willingness to design for a client’s real operating conditions, and to walk away from a deal that isn’t ready rather than chase the sale — is a genuine contributing factor. The cleanest example I know is a provider that scored at the very top of an independent analyst’s ratings and still walked away from a major account, because the client met them with the most dangerous phrase in this business: “we have always done it this (our) way.” That provider wasn’t at fault. Walking away was the proof of their commitment. They brought everything a provider can bring — and it still wasn’t enough, because the client wasn’t ready. A capable, committed provider is necessary. It is not sufficient.
So a good provider is never the reason a project succeeds — and never a non-factor either. Both statements hold, and they only seem contradictory if you’re still looking for a single hero or a single villain.
You can’t blame the car for the lake
Patrick Marlow, a GenAI champion at Google, put it better than I could, in an exchange we had in 2024: “Imagine buying a new Corvette, running it into the lake, then blaming Chevy for it not being able to float.” The car isn’t the problem. A Corvette is a superb machine. But the machine’s quality was never the determining variable — and that cuts both ways. You can’t blame the technology for the lake, and you can’t credit the technology for the track.
That is worth sitting with, because it’s the part the market keeps getting wrong in both directions. When a deployment fails, the technology takes the blame. When one succeeds, the technology takes the credit. Both are the same error — crediting or blaming the machine for what the driver and the conditions decided. A committed provider can build a superb car, and can even refuse to hand the keys to someone about to drive it into a lake. But what determines the outcome is whether the organization behind the wheel was ready — and whether anyone bothered to find out before the keys changed hands.
Why this should reassure rather than unsettle
I’ll close on the point that matters most right now. In a moment when the technology itself is moving fast enough to make everyone uncertain, the instinct is to look for stability in the tools. That’s looking in the wrong place — the tools are exactly what won’t hold still. The stability is underneath them, in the determining variables that haven’t changed across ERP, e-procurement, cloud, and now agentic AI.
That’s why a case like Logitech is reassuring rather than unsettling, once you see what it’s doing. The framework didn’t bend to absorb it. It tested it, declined to overclaim it, confirmed the determining variable through it, and clarified — without changing — the role the provider plays. That is what a foundation does. It doesn’t get rebuilt every time a new case arrives. It absorbs the case, and stays where it is.
The technologies will keep changing. The determining variable — whether the people behind the wheel are ready, and whether the provider was honest enough to check — will not.
Truth Is Believing. Accuracy Is Knowing.
-30-
Share this:
Like this:
Related