You already have one side of the marketplace.
That changes almost everything about how the build should be sequenced.
Procurement-friendly intake, starting with an NDA.
Thirty minutes with our enterprise lead. Bring the internal constraints as well as the idea: the systems it has to talk to, the functions that have to agree, and the process we will need to pass to be allowed to start.
A founder starting from nothing has the chicken-and-egg problem. You do not. You have suppliers, or you have customers, or you have both under a contract that was never written with a marketplace in mind. What you have instead is a harder version of a different problem: turning a channel you already control into a market where you no longer control every transaction.
The three problems that do not show up in a startup launch.
Your existing supply relationships were priced for a different model.
If you currently buy from your suppliers and resell, you have margin. If you move to a marketplace where they sell directly and you take a commission, that margin becomes a take rate, and the number your finance team expects is often not a number the supplier will accept. This is a commercial redesign wearing the costume of a technology project, and it is the single most common reason internal marketplace programmes stall after the platform is built.
The work happens before the build. Which suppliers move first, what the rate is, what they get in exchange for it, and what happens to the ones who stay on the old terms while both models run in parallel.
Your existing customers did not ask for more choice.
A marketplace introduces variance. Different sellers, different fulfilment times, different service quality, different returns experience. For a customer who has been buying a curated, first-party assortment from you for years, that variance reads as a drop in quality unless the trust layer is doing real work. Seller standards, dispute handling and a clear position on who is accountable when an order goes wrong are not phase-two features here. They are the thing that protects the relationship you already have.
Someone internally owns the P&L you are about to change.
This is the part nobody writes about. In an established business the marketplace competes with an existing channel, and that channel has a leader whose numbers move. Whether the programme survives its first bad quarter usually has less to do with the platform than with whether that person was in the room when the model was designed.
Worth settling early.
Because the two need different platforms and different teams.
If you want more assortment under your own brand, your own pricing and your own fulfilment, that is a drop-ship or wholesale extension. It is a commerce problem and your existing commerce platform can probably carry it.
It is a marketplace when third parties set their own prices, manage their own inventory, and are visible to the customer as distinct sellers. That brings in seller onboarding, split payments, commission accounting, per-seller performance management and a dispute process where you are the adjudicator rather than the counterparty. Different problem, different build.
If every seller disappeared tomorrow, would customers notice that a seller had gone, or only that a product had gone? If it is only the product, you are extending your catalogue and you should say so plainly, because it is the cheaper and faster answer.
Four things, before anyone opens a design file.
A model and margin session before any build.
Which side pays, what the rate is, and what it does to the existing channel's numbers. We would rather find out in week one that the commercial model does not clear than in month five.
A supply cohort, not a supply strategy.
Pick the ten to twenty existing relationships that move first, and design the onboarding for those specific companies. Generic supplier onboarding built for a supplier you have not met yet is how you end up rebuilding it.
The trust layer scoped against your current standard, not against a startup's.
Your customers have an expectation set by years of first-party service. That is the bar, and it is higher than the one a new marketplace has to clear.
Integration reality, early.
ERP, PIM, finance, existing identity. In an established business the integration surface is usually the largest single piece of the build and it is the piece most often underscoped, because it is invisible until someone asks how a commission-bearing order posts to the ledger.
Different constraints from a founder engagement.
Procurement compatibility, security review, and a named program manager for the duration. A different contract shape too: defined scope, defined approval gates, and documentation your own teams can carry forward after we step back.
Longer discovery
Six to eight weeks instead of two. Stakeholder mapping, internal user research, integration scoping with existing systems.
Procurement compatibility
MSA, SOW templates, security review packages, vendor onboarding. We have run this gauntlet before.
Named program manager
A dedicated PM owning your account from week one through year two. Single point of accountability.
We work with the channel owner in the room rather than around them, because a marketplace that the existing business is quietly hoping fails is not a technology problem we can solve.
What we cover under one engagement.
Design
Brand-compatible design system, accessibility, multi-stakeholder review cycles, design ops.
Development
Sharetribe, Nautical, Carro, Randevu, or fully custom. Picked for your constraints, not ours.
Integrations
SSO, ERP, PIM, payment processors, identity providers, internal data warehouses.
Data & analytics
Event schemas, warehouse pipelines, BI dashboards, cohort instrumentation from day one.
Compliance
SOC 2 readiness path, data residency options, GDPR, accessibility audits.
Post-launch
Named program manager, on-call response, quarterly business reviews, roadmap continuity.
The boring stuff your security team will ask about.
IP ownership
You own everything we ship: repo, designs, documentation, and test suites. Assigned in the MSA.
SOC 2 readiness path
We work to SOC 2 Type II controls and can deliver into your audit pipeline. Letter of attestation available on request.
Data residency
Canada, US, EU. Picked at architecture phase, locked at launch.
Confidentiality
Mutual NDA on first call. Background-checked engineers on all enterprise engagements.
If you are earlier than a build and still testing whether the model clears, that is a pre-development conversation. If the category is the harder part, marketplace types covers what changes vertical by vertical.