Buyers Will Not Change Behavior

Every coordination system is only as real as its adoption, and adoption is where I have watched good industrial ideas go to die. A buyer under schedule pressure already has a way of sourcing that works well enough. The last thing he wants is a new portal to learn, a new format to satisfy, and a new login to forget by Thursday.

So the objection is blunt: none of this matters, because buyers will not change their behavior, and a national coordination layer that depends on heroic behavior change is a whiteboard fantasy. I think that objection is right about buyers and wrong about what adoption actually requires. The domestic path has to become the path of least resistance, not a virtuous detour. Claim: buyers adopt a coordination layer only when the new path is faster than their current one and safer than a blind search, and only when it meets them in the tools they already use — email, PDFs, the drawings they already have — extracting structure for them rather than demanding it from them; a system that requires a new interface and clean uploads before it delivers value will be declined, and the failure shows up as buyers routing around it, back to the inbox.

The distinction worth holding onto is between changing behavior and reducing work. Nobody adopts a system that asks them to work differently for the system's benefit. People adopt systems that do their existing work faster, and that never feels like change. It feels like relief.

The Objection at Its Strongest

The objection is strong because buyer behavior is genuinely sticky and genuinely self-directed, and the data is unflattering to anyone building a portal. Thomas's 2019 survey of industrial buyers found that seventy-one percent vet fewer than five suppliers before deciding, which means the buyer is not shopping the network at all — he is confirming a short list already in his head. Buyers increasingly refuse hand-holding, too: Gartner reported in 2026 that sixty-seven percent of B2B buyers prefer a rep-free experience. And their attention is scattered across everything at once, with McKinsey's "rule of thirds" finding B2B buying split roughly evenly across in-person, remote, and self-service, and B2B customers now using ten or more channels, up from five in 2016. A buyer who vets four suppliers, wants no salesperson, and moves across a dozen channels is not going to reorganize his working life around a new manufacturing portal. Why would he?

There is a harder edge here, and it cuts against the obvious fix. Pure self-service makes outcomes worse. Gartner found that while most buyers want a rep-free experience, self-service purchases are far more likely to end in regret, and that buyers are 1.8 times more likely to complete a high-quality deal when they use digital tools in partnership with a human rather than alone. So the naive answer — build a slick self-serve portal and wait — fails twice. It does not get adopted, and it degrades the decisions of the people who do adopt it.

How This Fails Under Bad Design

The bad version makes adoption the buyer's job. It stands up a portal and waits. It requires clean, structured inputs before it returns anything useful, which is exactly backwards, since the buyer's inputs are messy for the same reasons the coordination problem exists in the first place. It forces a workflow change up front in exchange for a benefit that arrives later, if at all. And it tries to replace the inbox instead of reading it.

Under those terms the buyer does the rational thing. He routes around the system, back to email and a supplier he already knows, and the network never sees the demand it was built to coordinate.

The Governed Design

A governed layer meets the buyer inside the workflow he already has. It accepts an email with a PDF drawing attached, exactly as he sends it today, and does the structuring work itself — pulling out the constraints and flagging the unknowns, as Paper 14 described — so his first experience of the system is less work rather than more. That is the entire adoption strategy, and I do not have a second one hiding behind it.

From there it earns the next step by being faster than the status quo and safer than a blind search, which in practice means surfacing verified, relevant suppliers instead of a directory he has to vet by hand. When the inputs are genuinely ambiguous — a scanned print with a revision block written over in pen, a note that says "per previous" — it falls back to human review and clarification rather than breaking or guessing. That is not a concession. It is the hybrid path the data says produces better deals.

Adoption then compounds one reduced-effort interaction at a time, which is slower than a launch and considerably more durable. The control point is a single rule: the system adapts to the buyer's artifacts, and the buyer never adapts to the system's.

Operational Test

This is falsifiable in the adoption data, and the data arrives early. Measure how much demand enters through the buyer's existing channels — email, forwarded drawings — versus a form the buyer had to learn; early on the former should dominate, and if it does not, the system is selecting for the small population of buyers who enjoy new software. Track the time from a raw inbound request to a structured, routable one, and whether that time is falling.

Measure buyer workload directly, in clarification loops avoided and hours not spent vetting. Watch the fallback rate to human review as the signal that the system is choosing safety over false automation instead of confidently guessing at a tolerance nobody specified. Then watch whether they come back. If buyers try the path once and return to the inbox, the value arrived too late or cost too much effort, and no amount of explaining will fix that.

Implications

If adoption requires buyers to change how they work, the critic is right and the network stays empty while the demand stays in inboxes where it has always been. If adoption requires only that buyers do their existing work with less effort, behavior change stops being a precondition and becomes a byproduct.

You cannot exhort a national coordination layer into existence. It has to win the buyer's next RFQ on speed and safety, then the one after that, one email at a time, until the domestic path is simply the easiest path available — at which point nobody will describe what happened as adoption. They will say sourcing got easier.

The practical failure mode is non-adoption dressed up as a technology rollout. The objections so far have been practical. The next one I take up is legal, and it is the one that makes founders and their counsel nervous: coordinating buyers and suppliers at scale sounds, to certain ears, like antitrust risk.

Questions to Ask

1. Can a buyer use this by sending the same email and PDF they send today?
2. Does the buyer's first interaction create more work, or less?
3. Is the new path measurably faster than the current one, and safer than a blind search?
4. When inputs are ambiguous, does the system break, guess, or route to a human?
5. What share of demand enters through existing channels versus a new interface?
6. After one use, do buyers come back, or route around us to the inbox?