Thirty years of removing the technical barrier to supplier participation — and what happened to participation.
Truth Is Believing. Accuracy Is Knowing. Outcome Is Proof.™
For thirty years, the explanation for why suppliers do not join a buyer’s platform has been broadly the same: it is too hard and it costs too much.
That explanation has the great advantage of being testable. If technical difficulty is the cause, then removing technical difficulty should move participation.
It has been removed, repeatedly, by four generations of technology. Here is what happened.
The barrier, generation by generation
Electronic data interchange required a value-added network subscription, mapping software, technical staff most small suppliers did not employ, and a per-transaction cost that made every document a small commercial decision. Joining was a capital project.
XML removed most of that. An open standard, no VAN, transport over the public internet, and a technical floor low enough that a supplier with a competent developer could clear it.
Web services, cXML and punchout removed more. Supplier portals and networks removed more again — no integration at all for a small supplier, just a browser and a login. Cloud onboarding removed what was left.
By 2014, joining an automotive data exchange in India meant filling in a form. Compared with getting EDI running in 1995, that is not an incremental improvement. It is the difference between a project and an afternoon.
If the barrier were technical, participation should have followed the barrier down.
It did not
I have written about this repeatedly across the period, because I kept expecting the next generation to be the one that closed it.
In July 2014 I asked how IBM’s AutoDX was truly different from Covisint — the automotive exchange that had preceded it by fourteen years, on entirely different technology. My answer at the time was that the difference would not be found in the platform. It would be found in whether the thousands of small suppliers outside the core consortium came along, because a standard set by a centrally driven buyers’ group tends to be exclusionary rather than inclusive regardless of how easy it is to adopt.
Eight years later, in March 2022, ACMA was still holding sessions to bring component manufacturers and suppliers onto AutoDX. The success figure being quoted at that session was one million transactions in the first year — a 2014 number, cited in 2022.
AutoDX did not fail. That is what makes it useful. It is still running. It has had a decade of iteration on capable cloud technology. And the participation question is the same one it was in 2014.
The floor kept dropping. The room did not fill.
The strongest objection, and what survives it
The obvious response is that I am describing the barrier too simply, and it is a fair response.
XML did not make everything easier. It removed the cost of transport while introducing difficulty elsewhere — semantic interoperability, mapping to internal systems, exception handling, versioning, and the absence of the governance that EDI had achieved precisely by refusing to permit variation. Anyone who has reconciled two organisations’ interpretations of the same XML schema knows that some of what EDI solved by rigidity was un-solved by flexibility.
Accept all of it.
It still does not explain what happened, for one reason: that difficulty landed on the buyer’s side. Semantic mapping, ERP integration and exception handling are integration work, and integration work sits with the party operating the system, not with the supplier deciding whether to register.
The supplier’s barrier to joining fell sharply and stayed down through every subsequent generation. That is the barrier the participation problem has always been attributed to. It moved. Participation did not move with it.
A second objection is better. Network effects — suppliers join when enough buyers are already there — and accumulated distrust from earlier generations both plausibly slow adoption independent of technical ease. I think both are real. Note what they are, though: relational conditions, not technical ones. Conceding them concedes the argument.
If not the technology, then what?
Having set the technical explanation aside, the question becomes what suppliers were actually responding to. I have put that question to them, and to the people who represent them, across two decades — usually in their own settings rather than in a buyer’s.
PTDA Keynote (2005). Roughly two hundred delegates in Calgary, 52 distributors and 91 manufacturers, briefed on the mechanics of reverse auctions — the mechanism being deployed at them. Their concern was never how to connect. It was what happens to a relationship when price is the only variable the buyer’s tool is designed to see.
CFIB Radio Interviews (2009). The small business perspective on government procurement, and the structural barrier nobody counts: contract bundling. A supplier can be fully registered on a flawless platform and still be outside the addressable market, because the contract has been aggregated beyond anything they could bid on. Listener responses went further, describing bundling shaped by individual preference rather than by requirement.
2011 Judy Bradt Interview. Al Gordon — former aide to New York Governor Mario Cuomo — had told me that ninety percent of winning bidders in the US government market are effectively decided before the RFP is issued, through legitimate relationship-building. I put that to Judy Bradt, who at the time had spent fifteen years as a Canadian Embassy trade commissioner and two decades advising suppliers on both sides of that process. She did not dispute the number. She reframed it: think of the Winter Olympics, where ninety percent of what decides the winner happens before anyone crosses the starting line. Which means that if most of the decision precedes the formal process, joining the platform does not put a supplier in the game.
Bravo/Province of Ontario Interview and Series (2016). A platform implementation followed through to an Auditor General’s report — third-party corroboration rather than my assessment, on a system whose supplier-facing performance was not what the business case had described.
Twenty-one years, four technology generations, and four separate accounts of what stood between a supplier and participation.
Not one of them is technical.
The case that ran the other way
There is a counter-example, and it is instructive because it happened in the same era, on the same generation of technology, under harder terms.
Virginia’s eVA program launched in 2001. In the year before launch, between 5,000 and 6,000 of roughly 20,000 registered suppliers were receiving orders. By 2006, 14,371 of 34,000 were. Registration grew, and — the part that matters — the share of registered suppliers actually winning work grew with it.
Now consider the terms that growth happened under.
eVA charged suppliers to participate: a reverse funding model, initially one percent of order value capped at $500 per transaction. Incumbent large suppliers lobbied against the consolidation, correctly seeing it as a threat to established revenue with individual agencies. Those are harder conditions than Covisint ever imposed, on comparable technology, in the same years.
Participation grew anyway.
Bob Sievert, then Director of Virginia’s eProcurement Bureau, described the real story as the quality of participation rather than the registration count — the suppliers had become, in his phrase, the fuel for the procurement engine.
And the mechanism was stated plainly at the time. Virginia’s ability to sustain supplier interest was directly linked to the expanding distribution of business across the supply network. The more suppliers that won work, the more the Commonwealth looked like a genuine revenue opportunity rather than a compliance obligation.
Cost and benefit were aligned for the party being asked to participate. Not because the technology was easy — the technology was 2001 vintage and the suppliers were paying for access — but because joining improved the supplier’s position rather than only the buyer’s.
And then it held
Those figures are from 2007. A fair reader would ask what happened next, because a counter-example that stops nineteen years ago is a snapshot rather than evidence of anything lasting.
So in June 2026 I asked Bob Sievert directly: since the 2007 interviews, how many times has the technology platform changed or been modified?
His answer was that the system had modifications and improvements done just about monthly, carried out by CGI on the basis of requests Virginia made for user experience improvements, fixes and additional functionality. As for the platform itself, that changed once in the entire period — from the Ariba/CGI Advantage platform to Ivalua, a few years ago.
Read those two numbers together. Monthly revision. One platform replacement in roughly two decades.
That is the part the 2007 figures cannot show. Alignment at launch is a design decision, and plenty of programmes have got that far. Alignment nineteen years later is a cadence — the system kept being changed in response to what the people using it actually asked for, so the distance between what it did and what the operation needed never grew wide enough to require starting over.
Which is worth setting against everything else in this post. The pattern elsewhere is a platform implemented, then left to diverge from the operation until the divergence becomes intolerable and a replacement programme is launched — arriving into the residue of the last one. Virginia did the opposite, continuously, and needed the replacement exactly once.
What the comparison isolates
Set the two side by side.
AutoDX, 2014 onward. Trivially easy to join. Free at the point of entry. Modern cloud platform, a decade of improvement. Participation still the open question eight years on.
eVA, 2001 onward. Harder to join. Suppliers paying a transaction fee. Technology of its era. Participation grew, and the share of suppliers winning work grew with it.
If ease of access were the operative variable, those results are the wrong way round.
What differs is not the technology. It is whether the party being asked to participate had a reason to.
Why this is rarely discussed
I want to be careful here, because it is the one place in this argument where I could easily substitute a motive for an explanation.
The question receives remarkably little attention, and the most likely reason is structural rather than sinister. Every technology generation naturally directs attention toward what is newly possible. EDI and XML are settled infrastructure now — nothing to license, no category to track, no conference programme — so commercial and research attention moved elsewhere.
But if supplier participation did not resolve through those generations, then the new opportunity is carrying an old unsolved problem forward. The attention moved. The problem did not necessarily move with it.
That is a weaker claim than accusing anyone of avoidance, and I think it is the true one.
The question worth asking
This matters now because the same claim is being made again, about AI-assisted supplier onboarding, agentic data exchange and autonomous supplier management. The pitch is that the friction is finally low enough.
The friction has been getting lower for thirty years.
So before the next onboarding programme is approved:
What evidence do we have that this time the participation follows?
And the harder version, which is the one eVA actually answers:
What would a supplier gain by joining that they do not gain today?
Thirty years of the “technology barrier” dropping is not evidence. It is the control condition.
Keep the human at the wheel. Everything else is just faster.
-30-
This analysis draws on the Procurement Insights archive — an independent record carrying zero vendor sponsorships, published openly since 2007 and consolidating documented client work, lectures, and articles reaching back to 1998. Every claim is held to the Provenance Ledger™: a verify-before-publish discipline that traces each assertion to a primary source and never quietly edits the record once posted. That record is the evidence base for two working lenses — Invariant Physics™, the constant that however far the technology advances the operating logic must be in place first, and Implementation Physics™, its per-engagement application. Phase 0™ identifies and examines the unique and collective attributes within a wide range of seemingly disparate strands — the Strand Commonality™ theory, funded by the Government of Canada’s Scientific Research and Experimental Development program.
Getting it right rather than being right.
Jon W. Hansen, FCIPS — Procurement Insights | Hansen Models™
Related
Orchestration Will Improve Supplier Onboarding: How Will It Benefit The Suppliers?
Posted on September 3, 2026
0
Thirty years of removing the technical barrier to supplier participation — and what happened to participation.
Truth Is Believing. Accuracy Is Knowing. Outcome Is Proof.™
For thirty years, the explanation for why suppliers do not join a buyer’s platform has been broadly the same: it is too hard and it costs too much.
That explanation has the great advantage of being testable. If technical difficulty is the cause, then removing technical difficulty should move participation.
It has been removed, repeatedly, by four generations of technology. Here is what happened.
The barrier, generation by generation
Electronic data interchange required a value-added network subscription, mapping software, technical staff most small suppliers did not employ, and a per-transaction cost that made every document a small commercial decision. Joining was a capital project.
XML removed most of that. An open standard, no VAN, transport over the public internet, and a technical floor low enough that a supplier with a competent developer could clear it.
Web services, cXML and punchout removed more. Supplier portals and networks removed more again — no integration at all for a small supplier, just a browser and a login. Cloud onboarding removed what was left.
By 2014, joining an automotive data exchange in India meant filling in a form. Compared with getting EDI running in 1995, that is not an incremental improvement. It is the difference between a project and an afternoon.
If the barrier were technical, participation should have followed the barrier down.
It did not
I have written about this repeatedly across the period, because I kept expecting the next generation to be the one that closed it.
In July 2014 I asked how IBM’s AutoDX was truly different from Covisint — the automotive exchange that had preceded it by fourteen years, on entirely different technology. My answer at the time was that the difference would not be found in the platform. It would be found in whether the thousands of small suppliers outside the core consortium came along, because a standard set by a centrally driven buyers’ group tends to be exclusionary rather than inclusive regardless of how easy it is to adopt.
Eight years later, in March 2022, ACMA was still holding sessions to bring component manufacturers and suppliers onto AutoDX. The success figure being quoted at that session was one million transactions in the first year — a 2014 number, cited in 2022.
AutoDX did not fail. That is what makes it useful. It is still running. It has had a decade of iteration on capable cloud technology. And the participation question is the same one it was in 2014.
The floor kept dropping. The room did not fill.
The strongest objection, and what survives it
The obvious response is that I am describing the barrier too simply, and it is a fair response.
XML did not make everything easier. It removed the cost of transport while introducing difficulty elsewhere — semantic interoperability, mapping to internal systems, exception handling, versioning, and the absence of the governance that EDI had achieved precisely by refusing to permit variation. Anyone who has reconciled two organisations’ interpretations of the same XML schema knows that some of what EDI solved by rigidity was un-solved by flexibility.
Accept all of it.
It still does not explain what happened, for one reason: that difficulty landed on the buyer’s side. Semantic mapping, ERP integration and exception handling are integration work, and integration work sits with the party operating the system, not with the supplier deciding whether to register.
The supplier’s barrier to joining fell sharply and stayed down through every subsequent generation. That is the barrier the participation problem has always been attributed to. It moved. Participation did not move with it.
A second objection is better. Network effects — suppliers join when enough buyers are already there — and accumulated distrust from earlier generations both plausibly slow adoption independent of technical ease. I think both are real. Note what they are, though: relational conditions, not technical ones. Conceding them concedes the argument.
If not the technology, then what?
Having set the technical explanation aside, the question becomes what suppliers were actually responding to. I have put that question to them, and to the people who represent them, across two decades — usually in their own settings rather than in a buyer’s.
PTDA Keynote (2005). Roughly two hundred delegates in Calgary, 52 distributors and 91 manufacturers, briefed on the mechanics of reverse auctions — the mechanism being deployed at them. Their concern was never how to connect. It was what happens to a relationship when price is the only variable the buyer’s tool is designed to see.
CFIB Radio Interviews (2009). The small business perspective on government procurement, and the structural barrier nobody counts: contract bundling. A supplier can be fully registered on a flawless platform and still be outside the addressable market, because the contract has been aggregated beyond anything they could bid on. Listener responses went further, describing bundling shaped by individual preference rather than by requirement.
2011 Judy Bradt Interview. Al Gordon — former aide to New York Governor Mario Cuomo — had told me that ninety percent of winning bidders in the US government market are effectively decided before the RFP is issued, through legitimate relationship-building. I put that to Judy Bradt, who at the time had spent fifteen years as a Canadian Embassy trade commissioner and two decades advising suppliers on both sides of that process. She did not dispute the number. She reframed it: think of the Winter Olympics, where ninety percent of what decides the winner happens before anyone crosses the starting line. Which means that if most of the decision precedes the formal process, joining the platform does not put a supplier in the game.
Bravo/Province of Ontario Interview and Series (2016). A platform implementation followed through to an Auditor General’s report — third-party corroboration rather than my assessment, on a system whose supplier-facing performance was not what the business case had described.
Twenty-one years, four technology generations, and four separate accounts of what stood between a supplier and participation.
Not one of them is technical.
The case that ran the other way
There is a counter-example, and it is instructive because it happened in the same era, on the same generation of technology, under harder terms.
Virginia’s eVA program launched in 2001. In the year before launch, between 5,000 and 6,000 of roughly 20,000 registered suppliers were receiving orders. By 2006, 14,371 of 34,000 were. Registration grew, and — the part that matters — the share of registered suppliers actually winning work grew with it.
Now consider the terms that growth happened under.
eVA charged suppliers to participate: a reverse funding model, initially one percent of order value capped at $500 per transaction. Incumbent large suppliers lobbied against the consolidation, correctly seeing it as a threat to established revenue with individual agencies. Those are harder conditions than Covisint ever imposed, on comparable technology, in the same years.
Participation grew anyway.
Bob Sievert, then Director of Virginia’s eProcurement Bureau, described the real story as the quality of participation rather than the registration count — the suppliers had become, in his phrase, the fuel for the procurement engine.
And the mechanism was stated plainly at the time. Virginia’s ability to sustain supplier interest was directly linked to the expanding distribution of business across the supply network. The more suppliers that won work, the more the Commonwealth looked like a genuine revenue opportunity rather than a compliance obligation.
Cost and benefit were aligned for the party being asked to participate. Not because the technology was easy — the technology was 2001 vintage and the suppliers were paying for access — but because joining improved the supplier’s position rather than only the buyer’s.
And then it held
Those figures are from 2007. A fair reader would ask what happened next, because a counter-example that stops nineteen years ago is a snapshot rather than evidence of anything lasting.
So in June 2026 I asked Bob Sievert directly: since the 2007 interviews, how many times has the technology platform changed or been modified?
His answer was that the system had modifications and improvements done just about monthly, carried out by CGI on the basis of requests Virginia made for user experience improvements, fixes and additional functionality. As for the platform itself, that changed once in the entire period — from the Ariba/CGI Advantage platform to Ivalua, a few years ago.
Read those two numbers together. Monthly revision. One platform replacement in roughly two decades.
That is the part the 2007 figures cannot show. Alignment at launch is a design decision, and plenty of programmes have got that far. Alignment nineteen years later is a cadence — the system kept being changed in response to what the people using it actually asked for, so the distance between what it did and what the operation needed never grew wide enough to require starting over.
Which is worth setting against everything else in this post. The pattern elsewhere is a platform implemented, then left to diverge from the operation until the divergence becomes intolerable and a replacement programme is launched — arriving into the residue of the last one. Virginia did the opposite, continuously, and needed the replacement exactly once.
What the comparison isolates
Set the two side by side.
AutoDX, 2014 onward. Trivially easy to join. Free at the point of entry. Modern cloud platform, a decade of improvement. Participation still the open question eight years on.
eVA, 2001 onward. Harder to join. Suppliers paying a transaction fee. Technology of its era. Participation grew, and the share of suppliers winning work grew with it.
If ease of access were the operative variable, those results are the wrong way round.
What differs is not the technology. It is whether the party being asked to participate had a reason to.
Why this is rarely discussed
I want to be careful here, because it is the one place in this argument where I could easily substitute a motive for an explanation.
The question receives remarkably little attention, and the most likely reason is structural rather than sinister. Every technology generation naturally directs attention toward what is newly possible. EDI and XML are settled infrastructure now — nothing to license, no category to track, no conference programme — so commercial and research attention moved elsewhere.
But if supplier participation did not resolve through those generations, then the new opportunity is carrying an old unsolved problem forward. The attention moved. The problem did not necessarily move with it.
That is a weaker claim than accusing anyone of avoidance, and I think it is the true one.
The question worth asking
This matters now because the same claim is being made again, about AI-assisted supplier onboarding, agentic data exchange and autonomous supplier management. The pitch is that the friction is finally low enough.
The friction has been getting lower for thirty years.
So before the next onboarding programme is approved:
What evidence do we have that this time the participation follows?
And the harder version, which is the one eVA actually answers:
What would a supplier gain by joining that they do not gain today?
Thirty years of the “technology barrier” dropping is not evidence. It is the control condition.
Keep the human at the wheel. Everything else is just faster.
-30-
This analysis draws on the Procurement Insights archive — an independent record carrying zero vendor sponsorships, published openly since 2007 and consolidating documented client work, lectures, and articles reaching back to 1998. Every claim is held to the Provenance Ledger™: a verify-before-publish discipline that traces each assertion to a primary source and never quietly edits the record once posted. That record is the evidence base for two working lenses — Invariant Physics™, the constant that however far the technology advances the operating logic must be in place first, and Implementation Physics™, its per-engagement application. Phase 0™ identifies and examines the unique and collective attributes within a wide range of seemingly disparate strands — the Strand Commonality™ theory, funded by the Government of Canada’s Scientific Research and Experimental Development program.
Getting it right rather than being right.
Jon W. Hansen, FCIPS — Procurement Insights | Hansen Models™
Share this:
Like this:
Related