Specify the vendor pricing and tier structure #34
Reference in New Issue
Block a user
No description provided.
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Question
Directed change of course, not a surfaced question. The author reversed the free-bidder model
mid-session: there is no free tier. A nominal $100/month access fee is the floor, and the map
needs the tier structure, upsell axis, and pricing to match.
This supersedes What do bidders actually pay for? and decision 1 of
the entitlement and subscription model.
Resolve: where the paywall sits relative to a retailer's invitation; what the tiers are and what
axis a vendor upgrades along; what the base tier includes; and what happens to a vendor who lapses
mid-participation.
Parent: #1
Resolution
This reverses a settled decision. What do bidders actually pay for? held that seeing RFPs
and submitting a bid is free forever, and entitlement decision 1 built the whole
model on it. Both are superseded here by author direction. What follows is the replacement, and the
costs are recorded rather than argued away.
0. Access is paid. There is no free tier.
$100/month is the floor. A vendor cannot see or do anything without an active subscription.
The concern raised before deciding, recorded because it does not go away: $1,200/yr sits above
the ISNetworld/Avetta band ($450-900/yr) that the ecosystem research found 38% of
subcontractors cite as an enrollment barrier. The sharp edge is not vendor grumbling — supplier
diversity is a real retail procurement mandate, so MBE/WBE/DBE suppliers who cannot clear the fee
drop out of the retailer's pool and the retailer's own diversity reporting degrades. That converts a
vendor-side price objection into a retailer-side adoption risk, on the side that is supposed to be
free.
And #6's structural argument is now a live cost rather than an avoided one: gating access
suppresses bid count, and bid count is exactly what makes the free retailer side valuable. The two
sides are now in tension by construction. This is accepted deliberately; it is not solved.
1. The gate is hard — subscription required to see anything
An invited vendor with no subscription sees only that a retailer has invited them. No teaser, no
scope, no timeline.
The visibility model is re-layered, not reversed. Its teaser and acknowledgement gate
exist for confidentiality, not billing, and still apply — they simply now sit inside the
paywall:
What it costs. The retailer becomes the messenger for our invoice, and a retailer importing 400
suppliers needs 400 conversion events before their first sourcing event draws a full field. This is
the Ariba mechanic the research documented as the most resented in the ecosystem. Mitigation is
deliberately deferred — see below.
2. Three tiers, on seats and AI capability
1x / 3.5x / 10x. The top tier deliberately undercuts Responsive (
$14k/yr) and Loopio ($20k/yr).Never volume, and never per-retailer. Entitlement rejected volume caps as access-gating;
that objection gets stronger here, not weaker — a vendor cut off mid-sourcing-event after
paying is a refund conversation and a furious retailer. Per-retailer pricing was rejected because
it is "pay to talk to your customer" priced per customer: the single most precisely resented
mechanic in the entire research, and applying it on top of an access fee compounds it.
What it costs. Revenue does not track value at the small end — a one-person vendor bidding
constantly pays the same as one bidding twice a year.
3. The rail is in Base; the assistant is the upsell
Every paying vendor gets the durable vault, cross-retailer reuse, the acceptance graph, and the
portable profile — the one thing the ecosystem research identified as genuine white space.
Pro sells the AI.
Why not put the cross-retailer profile behind Pro, where it is the obvious upgrade lever: a Base
vendor would then pay $1,200/yr for something narrower than ISNetworld at $450-900/yr, and the
network flywheel would only spin among Pro subscribers. The rail has to be the thing everyone is on.
Note this inverts entitlement decision 1, which made persistence the paid feature
against a free tier. Persistence is now in the floor tier; the paid axis moved up to capability.
What it costs. Pro carries the entire upgrade burden on AI and seats alone.
4. On lapse, live participations run to completion
Any RFP the vendor is already a participant in stays fully accessible through award —
clarifications, finalist rounds, revised pricing. Lapse blocks only new participation.
This is the rule the vault decision 4 already set for document expiry: in-flight
participation is never silently revoked; a vendor does not lose a live bid to a clock. Applying it
to billing keeps one rule across two mechanisms.
The governing principle from entitlement needed restating, since entitlement now gates
access rather than convenience. The refined form:
Retailer-facing correctness still wins: a submitted bid is never withdrawn for non-payment, the
expiry scheduler still runs for every vendor regardless of billing, and eligibility never depends
on subscription state — a lapsed vendor stays eligible, is still selected into rule-defined
audiences, and simply sees the subscribe wall.
What it costs. A vendor can ride a long sourcing cycle unpaid; award-to-close can be months.
Deferred by explicit direction — pinned, not decided
Onboarding, trials, retailer-sponsored seats, and sales friction as a set. Weighed and pinned
for a later pass. Sponsorship in particular is no longer merely a growth lever — with a hard gate it
is the only proposed answer to the diversity-mandate problem in decision 0, which means
the Out-of-scope ruling on it is now downgraded from declined to deferred. Recorded in
Not yet specified.
Constraints handed to other tickets
reference #20 kept becomes load-bearing if sponsorship returns; participation rows need enough
state to answer was this vendor a live participant when they lapsed.
context, asserted per feature. It now also carries an access-level assertion at the
participation boundary, and the assertion sites remain the derived pricing page.
lapsed vendor still owns every row they owned yesterday, and eligibility is explicitly independent
of billing.
runs straight through the vendor workspace — the AI affordances must be visible-but-locked rather
than absent, or Pro has no upgrade moment.