30. Marketplace Business Models

A marketplace connects two or more sides of a market and makes transactions, discovery, trust, or coordination easier. Marketplaces can become powerful because liquidity and network effects can compound. They can also burn years because both sides wait for the other side to show up.

The first marketplace question is not “Can this be a platform?” It is: “Can we create a small pocket of liquidity where buyers and sellers repeatedly get a better outcome than the current alternative?”

Marketplace Basics

Core concepts:

Concept	Meaning	Founder test
Supply	Sellers, service providers, inventory, professionals, or assets	Can you onboard enough quality supply before demand arrives?
Demand	Buyers, users, employers, renters, patients, brands, or businesses	Is the demand urgent and repeated?
Liquidity	The chance that demand finds useful supply quickly	Can a user get a good result in a narrow segment today?
Trust	Proof that both sides will behave reliably	What reduces fraud, poor quality, delay, and fear?
Matching	How the right sides find each other	Is matching simple, searchable, curated, or managed?
Take rate	Your share of transaction value	Can you charge without encouraging bypass?
Disintermediation	Users taking the transaction outside your platform	Do you add enough trust, workflow, payments, or convenience to stay involved?
Network effects	More users make the marketplace more valuable	Does more supply improve demand outcomes, and vice versa?

Marketplaces are about density. A small dense market usually beats a large empty one.

The Marketplace Equation

A marketplace becomes a business when three things happen together:

Demand can find relevant supply.
Both sides trust the transaction.
The platform captures enough value to pay for acquisition, operations, support, and risk.

Founders often obsess over the third point too early. Take rate matters, but if buyers do not get good matches or sellers do not get real demand, there is nothing to monetize.

The first version of a marketplace can be ugly and manual. The key is whether the transaction works. If a founder manually calls suppliers, curates options, negotiates price, and resolves disputes, that is not failure. It is learning. The product can automate later what the founder first understands manually.

Types of Marketplaces

B2C marketplace: consumers buy from businesses or sellers.
B2B marketplace: businesses discover, purchase, finance, or transact with other businesses.
Services marketplace: buyers find skilled providers.
Labor marketplace: companies hire or contract talent.
Managed marketplace: the startup controls quality, fulfillment, or service delivery more deeply.
Vertical marketplace: focused on one category or industry.
Local marketplace: value depends on geography and proximity.

The more trust and quality matter, the more “managed” your marketplace may need to be early. Pure self-serve marketplace dreams often fail because early users need curation.

Marketplace Revenue Models

Take rate

The platform takes a percentage of transaction value. This works when transactions are visible, payments can flow through the platform, and the marketplace adds enough value to justify staying inside.

Watch for disintermediation. If buyer and seller can easily transact outside after discovery, you need workflow, payments, financing, guarantees, dispute handling, compliance, or convenience that keeps the platform useful.

Listing fees

Supply pays to be listed or promoted. This can work when demand is valuable and sellers trust that visibility leads to revenue. It is dangerous before liquidity because sellers may feel charged for empty traffic.

Subscription

One or both sides pay recurring access. This can work in B2B, talent, real estate, data, or procurement markets where discovery itself has recurring value.

Lead fees

The platform charges for qualified leads. This can work in services, education, finance, real estate, and high-ticket categories. The risk is lead quality disputes.

Managed margin

The platform buys, manages, bundles, or controls delivery and earns margin. This can increase trust and quality, but also turns the business more operationally heavy.

Financing or value-added services

Marketplaces can monetize credit, insurance, logistics, verification, software, analytics, or compliance. In India, B2B marketplaces often need these layers because pure discovery is not enough.

Cold Start Strategy

The cold start problem is brutal: supply does not join without demand, demand does not return without supply.

Ways to start:

Approach	When useful
Seed supply first	Demand has clear search intent and supply quality matters.
Seed demand first	Supply is easy to recruit when demand is proven.
Manual matching	The market is complex and you need to learn matching rules.
Single geography	Local density matters.
Single category	Category depth beats broad emptiness.
Anchor customers	A few large buyers can attract supply.
Community-led liquidity	Trust and identity matter more than pure transaction speed.
Managed service first	You need to perform the transaction manually before software can scale it.

Do not hide manual work too early. Manual matching can teach you what the product must eventually automate.

Choose The Atomic Market

A marketplace does not begin with the total market. It begins with an atomic market: the smallest segment where enough supply and demand can meet repeatedly.

An atomic market has:

One clear demand use case.
One supply type.
One geography, community, category, or workflow.
A transaction that can happen soon.
A trust layer that is sufficient for the risk.
A way to measure liquidity.

Examples:

Broad idea	Atomic market
Marketplace for business services	GST accountants for early-stage startups in Bengaluru
Healthcare marketplace	Dermatology consultations for young urban professionals
B2B procurement	Packaging suppliers for D2C brands shipping under a specific volume
Talent marketplace	Contract UI designers for funded SaaS startups
Used equipment marketplace	Refurbished restaurant equipment in one metro area

If you cannot create liquidity in an atomic market, a national launch will usually increase the mess. More categories mean more empty shelves. More geographies mean thinner supply. More user types mean weaker matching.

The Liquidity Loop

Marketplace growth should create a loop:

Demand arrives with a specific need.
The platform produces a useful match quickly.
The transaction is completed with enough trust.
The buyer returns or refers.
Supply earns enough value to stay active.
Better supply improves future demand outcomes.

If any step breaks, growth becomes expensive.

Broken step	Symptom	Possible fix
Demand is vague	Many visitors, few requests	Narrow the use case and landing promise
Supply quality is weak	Buyers browse but do not transact	Curate supply or raise onboarding standards
Matching is slow	Users leave before result	Manual matching, filters, concierge, better data
Trust is weak	Users negotiate offline or avoid payment	Verification, guarantees, reviews, escrow, support
Supply earns too little	Providers churn	Improve demand quality, pricing, utilization
Buyer does not repeat	One-time use only	Add workflow, reminders, subscription, adjacent needs

This loop should be reviewed by category or geography. Overall marketplace averages hide liquidity pockets and dead zones.

Liquidity Metrics

Measure liquidity narrowly, not at the total marketplace level.

Useful metrics:

Search-to-result rate.
Match rate.
Time to first useful match.
Fill rate.
Quote response time.
Transaction completion rate.
Repeat transaction rate.
Supply utilization.
Cancellation or no-show rate.
Dispute rate.
Buyer NPS and seller NPS by category or geography.

If the marketplace has 10,000 listings but buyers cannot find a good match in their category, liquidity is weak. If sellers join but never receive meaningful demand, supply will decay. If buyers get matches but do not transact, trust, price, quality, or workflow may be broken.

Marketplace Unit Economics

GMV can be seductive. It is not the same as revenue, margin, or cash.

Track the economics of a transaction:

Layer	Question
GMV	What is the total transaction value?
Take rate or fee	What portion do we keep?
Payment cost	What do payments, refunds, and disputes cost?
Operations cost	What manual work is required to complete the transaction?
Support cost	How much help do both sides need?
Fraud or failure cost	What losses come from disputes, quality issues, or cancellations?
Contribution margin	What is left after variable costs?
Repeat rate	Does the same buyer or seller transact again?

A marketplace with high GMV and negative contribution margin is not automatically a bad business, but founders must know what improves with scale. If scale adds more support, more disputes, more refunds, and more manual work, the model may not become better later.

For managed marketplaces, operations may be the product. That is acceptable if pricing and margins reflect it. Do not pretend a managed marketplace has pure software economics.

Trust Layers

Marketplaces often fail because users fear the other side. Trust can be built through:

Verification.
Reviews and ratings.
Curated onboarding.
Guarantees.
Escrow or payment protection.
Refund and dispute process.
Quality standards.
Insurance.
Identity checks.
Transparent pricing.
Responsive support.

The right trust layer depends on risk. Buying a low-cost product, hiring a caregiver, sourcing industrial materials, booking a doctor, and lending money are not the same trust problem.

Disintermediation Defense

If your marketplace only introduces buyer and seller, users may bypass you after the first transaction. To stay in the transaction, the platform must add value beyond discovery.

Defenses include:

Payment protection.
Escrow or milestone release.
Quality assurance.
Dispute resolution.
Logistics or fulfillment.
Compliance and documentation.
Insurance or guarantees.
Workflow tools.
Financing.
Reputation history.
Repeat ordering.

The best defense is not blocking phone numbers in chat. It is making the platform the safest, easiest, and most valuable place to transact.

Supply Strategy

Supply is not just quantity. It is quality, availability, price, response time, and willingness to transact through your platform.

Founders should know:

Why would supply join before demand is abundant?
What supply is hardest to acquire?
What supply attracts the best demand?
What incentives keep supply active?
What quality standards are enforced?
What happens to bad supply?

In many marketplaces, the best early move is to recruit supply manually and make them successful, not to build a self-serve onboarding flow first.

Demand Strategy

Demand must have enough urgency to try a new way. Weak demand says, “interesting.” Strong demand says, “can you help me solve this today?”

Good early demand sources:

A narrow buyer community.
Anchor customers.
Search intent.
Existing offline pain.
A repeated procurement need.
A high-friction service workflow.
A category where trust is broken.

Demand quality matters. A flood of low-intent users can waste supply and damage trust.

India Angle

India has strong marketplace opportunities because of fragmentation: suppliers, small businesses, informal networks, regional diversity, and inconsistent trust infrastructure. But fragmentation also increases operational load.

Indian marketplace founders must think hard about:

Trust and verification.
Payments and collections.
Refunds, disputes, and fraud.
Regional supply quality.
Logistics or fulfillment.
Offline relationships.
Price sensitivity and bargaining.
Support in local context.

For B2B marketplaces, the real moat may be procurement workflow, credit, fulfillment reliability, and supplier trust, not just listing inventory. For services marketplaces, quality control and dispute resolution can decide the business.

India also has a strong offline relationship layer. Many sellers and buyers already transact through brokers, WhatsApp groups, local agents, community references, or long-standing supplier networks. Your marketplace must be better than that existing trust system, not just more digital.

In local categories, geography can be destiny. A marketplace that works in one city may fail in another because supply behavior, pricing, language, regulation, or logistics differ.

Common Mistakes

Launching nationally before one segment has liquidity.
Measuring signups instead of successful matches.
Assuming network effects before retention exists.
Charging take rate before value is clear.
Ignoring disintermediation.
Treating supply as a commodity when quality varies.
Building software before manually learning the transaction.
Monetizing discovery before proving transaction value.
Ignoring disputes and refunds until they become public complaints.
Onboarding supply that demand does not want.
Optimizing GMV while contribution margin is negative.
Assuming network effects before repeat usage exists.

Founder Decisions

Before scaling, decide:

Which side is harder to acquire?
Which side should be subsidized first?
What is the first liquidity pocket?
What trust mechanism is non-negotiable?
What transaction must stay on-platform?
What take rate or monetization path is realistic?
What work must be managed manually before software can automate it?
What category, geography, or buyer type are we refusing for now?

Liquidity Diagnostic

Marketplace liquidity is not traffic. It is successful matching.

Track:

Metric	Why it matters
Search-to-match rate	Demand finds relevant supply.
Time to match	Liquidity is fast enough to trust.
Fill rate	Supply can satisfy demand.
Repeat transaction rate	The market has ongoing need.
Failed match reason	Reveals supply, price, trust, or UX problem.
Contribution margin per transaction	Growth is not hiding losses.

If liquidity is weak, do not scale geography or categories. Fix the atomic market.

Trust And Dispute System

Trust is part of the business model.

Design:

Verification of supply and demand.
Ratings or references where useful.
Clear pricing and fees.
Cancellation/refund rules.
Dispute resolution owner.
Fraud or abuse monitoring.
Support expectations.
Off-platform transaction policy.

In many Indian marketplaces, offline trust already exists. Your system must make the transaction safer, faster, more reliable, or more transparent than the offline alternative.

Marketplace Manual Ops Plan

Early marketplaces often need manual operations.

Write:

Which matches will be manual?
What quality checks happen before listing?
Who resolves disputes?
Which side receives concierge support?
Which process will be automated after it repeats?
Which manual work should never be automated because it creates trust?

Manual work is acceptable when it teaches the market. It becomes dangerous when it hides negative unit economics.

Take Rate And Incentive Design

Marketplace monetization must respect incentives. A take rate that feels fair in one category may break another.

Evaluate monetization with this table:

Question	Why it matters
Who receives the most visible value: supply, demand, or both?	The side receiving value can usually bear more cost.
Does the platform reduce risk, increase demand, or save time?	Fees are easier when value is obvious.
Is the transaction recurring or one-time?	Recurring markets can support different economics than one-off leads.
Can users bypass after discovery?	High bypass risk may require trust, workflow, payments, or managed service.
Does the fee distort quality?	Bad incentives can flood the market with low-quality supply or fake demand.
Is payment captured on-platform?	Monetization is easier when the transaction naturally flows through you.

Possible models:

Model	Works when	Risk
Take rate	Transaction value is clear and payment can stay on-platform.	Users bypass if platform adds little after discovery.
Subscription	Supply or demand needs ongoing access and tools.	Customers churn if liquidity is weak.
Lead fee	Leads have clear commercial value.	Quality disputes can damage trust.
Listing/promotion fee	Supply wants visibility.	Can reward whoever pays, not whoever is best.
Managed margin	Platform improves reliability, quality, or fulfillment.	Operations can crush margins.
Financing/value-added services	Platform owns valuable transaction data or workflow.	Can distract from core liquidity too early.

Do not monetize before users trust the market. But do not postpone monetization forever either. The right question is: where does the platform create value that users would miss if it disappeared?

Marketplace Density Plan

Marketplace growth should start with density, not ambition.

Define one atomic market:

Dimension	Example decision
Geography or niche	One city, one online community, one buyer cluster.
Category	One service, SKU type, worker type, or supplier class.
Demand job	One repeated reason buyers arrive.
Supply promise	One quality standard supply must meet.
Liquidity target	Time to match, fill rate, or repeat transaction.
Trust mechanism	Verification, guarantee, escrow, reviews, managed support.
Manual ops owner	Who makes the market work before software does.

Only expand when the first pocket has a repeatable reason to return. A marketplace without density is a directory with hope.

Liquidity Budget

Every marketplace spends something to create liquidity. It may spend money, founder time, discounts, guarantees, manual ops, trust, or brand. Make the budget explicit.

Liquidity Cost	What It Looks Like	Danger
Demand subsidies	Discounts, cashback, free delivery, credits.	Users leave when incentives stop.
Supply subsidies	Guaranteed minimum earnings, onboarding support, listing help.	Supply quality may depend on payment, not demand.
Manual matching	Founder or ops team creates matches by hand.	Software appears to work only because humans hide friction.
Quality assurance	Verification, inspection, training, dispute handling.	Trust cost rises faster than take rate.
Inventory or working capital	Platform fronts stock, payment, or fulfillment.	Marketplace becomes balance-sheet-heavy.
Customer support	Humans resolve exceptions and trust gaps.	Support cost can erase margin.
Community building	Events, groups, moderation, education.	Community can be noisy without transactions.

For the first liquidity pocket, write:

We will spend ______ to create liquidity until ______ happens. We will stop or change strategy if ______.

Marketplaces die when subsidy becomes identity. Manual effort and incentives are useful only if they teach you how liquidity can eventually repeat with better economics.

Subsidy Exit Plan

If you use incentives, design the exit before launch.

Incentive	What Must Improve Before We Reduce It	Exit Signal
Buyer discount	Repeat purchase, trust, selection, speed.	Buyers return at lower incentive.
Supplier guarantee	Organic demand, utilization, earning predictability.	Supply earns enough from transactions.
Free listing	Supply density and catalog quality.	Supply sees real leads or revenue.
Free managed service	Trust and workflow learning.	Process becomes standardized or paid.
Founder concierge	Learn matching and objections.	Ops playbook can be run by team or software.

An incentive is healthy when it buys learning, liquidity, or trust that remains after the incentive reduces. It is unhealthy when it merely buys fake GMV.

Marketplace Quality Decay

Marketplaces often get worse as they grow because weak supply, low-intent demand, fraud, spam, or low-quality transactions enter the system.

Watch for quality decay:

Decay Signal	What It Means	Response
More listings, worse conversion	Supply quantity is not quality.	Tighten verification, ranking, and category focus.
More demand, slower fulfillment	Supply density or reliability is weak.	Improve supply activation before more demand spend.
High GMV, low repeat	Transactions are not satisfying.	Measure repeat by cohort and reason for non-repeat.
Rising disputes	Trust system is underbuilt.	Add rules, evidence, support, and penalties.
Off-platform leakage	Users do not value platform after discovery.	Add payments, guarantees, workflow, financing, or tools.
Supplier churn	Supply economics or treatment is weak.	Improve utilization, earnings, tools, or communication.

The goal is not maximum listings or maximum traffic. The goal is repeated successful transactions with improving trust and economics.

B2B Marketplace Caution

B2B marketplaces are often harder than consumer marketplaces because procurement, credit, quality, relationships, logistics, and trust are heavier.

Before building a B2B marketplace, ask:

Is the transaction standardized enough to compare?
Does the buyer want choice or a reliable vendor?
Does the supplier want leads, orders, financing, tools, or distribution?
Can pricing be transparent, or is every deal negotiated?
Who handles credit, payment delay, returns, disputes, or quality claims?
Does the platform reduce risk enough to justify a fee?
Can the transaction move online, or is online only a lead source?

Many B2B “marketplaces” are actually workflow software, financing businesses, managed procurement, or lead-generation businesses. That is fine. The danger is pretending liquidity exists when the real value is operational trust.

Marketplace Unit Economics Review

Marketplace founders often celebrate GMV before understanding whether each transaction is healthy. GMV is activity. Unit economics show whether the platform can become a business.

Review economics by transaction type or category:

Item	Question
Gross transaction value	What is the total value moving through the marketplace?
Take rate or revenue	What does the platform actually keep?
Payment and processing cost	What does payment, wallet, escrow, refund, or settlement cost?
Acquisition cost	What did it cost to acquire buyer and supply?
Incentives	What discounts, credits, guarantees, or commissions were used?
Ops cost	How much manual matching, verification, support, dispute handling, or quality control was needed?
Losses and leakage	What refunds, fraud, off-platform transactions, or failed fulfillment appeared?
Repeat behavior	Did either side return without the same incentive?

Then classify each transaction cohort:

Cohort Type	Meaning	Founder Response
Healthy liquidity	Repeat transactions, improving trust, acceptable ops cost.	Invest more.
Subsidized learning	Incentives buy useful data or behavior.	Continue only with an exit plan.
Manual illusion	Transactions work only because founders hide friction.	Productize or narrow scope.
Bad liquidity	Volume grows with poor quality, disputes, leakage, or losses.	Stop or redesign the market.

A marketplace can look alive while losing money on every transaction. The founder’s job is to find the pocket where liquidity, trust, and economics improve together.

Marketplace Constraint Map

Every marketplace has a binding constraint. If you misread the constraint, you will spend on the wrong side.

Constraint	Symptom	Better Action
Demand shortage	Supply waits, listings go stale, sellers lose interest.	Narrow use case, improve demand acquisition, or seed anchor buyers.
Supply shortage	Buyers search but cannot find enough reliable options.	Build supply density, guarantee availability, improve supplier economics.
Trust shortage	Both sides exist but hesitate to transact.	Add verification, reviews, guarantees, escrow, support, or managed service.
Match quality shortage	Transactions happen but satisfaction is weak.	Improve categorization, ranking, qualification, and human matching.
Repeat shortage	First transaction works but users do not return.	Study frequency, habit, price, quality, and post-transaction workflow.
Take-rate shortage	GMV grows but revenue does not.	Add value beyond discovery: payments, tools, logistics, financing, workflow.
Ops-cost shortage	Founders manually save every transaction.	Narrow the atomic market, standardize process, automate only repeated work.
Leakage shortage	Users bypass after discovery.	Increase platform value after match: trust, payment, insurance, records, financing, convenience.

Do not try to scale both sides blindly. Ask:

What is the one constraint that, if improved, would make the next 20 transactions easier?

Then design the next experiment around that constraint. A marketplace is not a website with two user types. It is a liquidity system with a bottleneck.

Reader Action

Define your first liquidity pocket:

One geography or online niche.
One category.
One demand use case.
One supply type.
One success metric: match rate, time to match, repeat purchase, GMV, fill rate, or retention.
One trust mechanism.
One reason users will not bypass you.

If you cannot create liquidity in a narrow pocket, a broader launch will usually make the problem worse.

Marketplace Leakage Audit

Marketplaces leak when buyers and suppliers meet on the platform but move future transactions off-platform. Leakage is not only a policy problem. It is usually a value problem.

Audit leakage by asking why users would stay:

Platform value	If strong	If weak
Discovery	Users find better matches faster.	Users leave after first contact.
Trust	Verification, reviews, escrow, guarantees, or dispute help reduce risk.	Users rely on direct relationship instead.
Workflow	Booking, communication, documents, status, or delivery is easier on-platform.	WhatsApp and phone calls take over.
Payments	Payment, credit, invoicing, settlement, or refunds are safer.	Users transact directly to save fees.
Records	History, compliance, warranties, ratings, or proof matter.	No reason to keep the transaction recorded.
Repeat	Platform makes reordering, scheduling, or recurring demand easier.	Users store contact details and bypass.
Financing or insurance	Platform offers something hard to recreate directly.	Take rate feels like tax.

If leakage is high, do not only add stricter rules. Increase platform value after the match.

Leakage diagnosis

Use this table:

Leakage cause	Signal	Founder move
Take rate too high for perceived value	Users complain about commission or negotiate offline.	Add workflow/payment/trust value or change pricing.
Relationship matters more than platform	Repeat buyers contact supplier directly.	Own records, scheduling, quality, financing, or guarantees.
Platform slows transaction	Users switch to WhatsApp/calls.	Improve transaction workflow and response speed.
Trust is external	Users trust personal networks more than ratings.	Build verification, references, local proof, and support.
Payment is easier offline	Cash, bank transfer, credit, or invoice process wins.	Improve payments, settlement, invoice, credit, or collections.

The goal is not to trap users. The goal is to make the platform more useful than bypassing it.

Marketplace Trust Cost Ledger

Trust has a cost. Verification, support, escrow, dispute resolution, refunds, guarantees, moderation, manual review, field checks, and quality control all cost money. A marketplace business model should price and operate with those costs visible.

Create a trust cost ledger:

Trust activity	Cost driver	Who pays?
Supply verification	Manual review, document checks, phone calls, site visits.	Platform, supplier, buyer, or embedded in take rate.
Quality control	Sampling, audits, rework, ratings, delisting.	Platform or supplier.
Dispute handling	Support time, evidence review, refunds, chargebacks.	Platform, buyer, supplier, or shared.
Payment protection	Escrow, settlement delay, refunds, fraud checks.	Platform fee or transaction fee.
Guarantees	Replacement, refund, insurance, warranty.	Explicit fee or higher take rate.
Compliance records	GST invoices, contracts, identity, delivery proofs.	Often platform-funded until scale.

If the marketplace takes a fee but does not fund trust, users will eventually ask what the platform is for. If the marketplace funds trust but does not price for it, margins will collapse.

Marketplace Subsidy Rules

Subsidies can help cold-start liquidity, but they should have an exit plan. Subsidizing without a learning goal creates fake traction.

Before subsidizing, define:

Question	Answer
Which side is subsidized: supply, demand, or both?
What behavior are we trying to create?
What metric proves the subsidy worked?
What will happen when the subsidy is reduced?
What is the maximum budget and end date?
Which segment will not receive subsidy?

Healthy subsidies reveal whether liquidity, trust, or repeat behavior can emerge. Unhealthy subsidies buy transactions that disappear as soon as incentives stop.

Marketplace Expansion Gate

Do not expand to a new category, city, or segment just because the current one feels slow. Expansion multiplies operational complexity.

Use this gate:

Gate	Minimum evidence
Liquidity	Current atomic market hits match-rate and time-to-match targets consistently.
Trust	Dispute, refund, fraud, and quality issues are understood and manageable.
Unit economics	Contribution margin after support, refunds, incentives, and payment cost is visible.
Supply repeatability	Supply acquisition and activation can be repeated without founder heroics.
Demand repeatability	Demand source produces buyers with real transaction intent.
Operations	Manual processes are documented enough for another person to run.

Expansion should make the marketplace stronger, not spread weak liquidity across more empty rooms.

Marketplace Payment Timing And Working Capital

Marketplaces often look asset-light until payment timing appears. Refunds, settlement delays, escrow, supplier payouts, credit, failed payments, chargebacks, cancellations, and guarantees can create cash pressure even when GMV is growing.

Map payment timing:

Payment event	Founder question
Buyer pays	Is payment upfront, after delivery, milestone-based, cash, invoice, UPI, card, bank transfer, credit, or wallet?
Platform fee	Is the fee deducted before payout, invoiced later, or charged separately?
Supplier payout	When does supply expect money, and what proof is required before release?
Refund or dispute	Who funds refund while the dispute is investigated?
Failed payment	Who owns collection, retry, cancellation, or replacement?
Credit terms	Is the platform financing the buyer, supplier, or both?
Tax/invoice records	Who issues invoice, collects tax information, and maintains transaction proof?

Then ask:

Can the marketplace survive a month where GMV grows but collections, refunds, or payouts move against us?

Working capital risk is especially important in B2B, logistics, services, rentals, commerce, and any marketplace where suppliers need fast payment but buyers pay slowly. Growth can increase cash stress if timing is not designed.

Marketplace Governance Model

A marketplace is not only software. It is a set of rules that participants believe are fair enough to keep using.

Define governance early:

Rule area	What to decide
Admission	Who is allowed to join supply and demand?
Ranking	How are listings, suppliers, jobs, bids, or offers ordered?
Quality	What behavior improves or reduces visibility?
Pricing	Who controls price, discounts, commissions, and fees?
Cancellation	What happens when buyer or supplier cancels?
Disputes	What evidence is accepted and who decides?
Fraud and abuse	What behavior leads to warning, hold, suspension, or delisting?
Repeat relationships	What is allowed off-platform and what must remain on-platform?

Good governance should protect the best participants. If rules mainly protect the platform while serious users feel exposed, quality will leave. If rules are too loose, low-quality behavior will flood the market.

Run a monthly governance review:

Question	Evidence
Are good suppliers getting enough demand?	Response rate, repeat transactions, earnings, churn.
Are serious buyers getting reliable outcomes?	Match rate, completion, dispute rate, repeat demand.
Are rules being gamed?	Fake reviews, spam, bypassing, refund abuse, duplicate accounts.
Are support decisions consistent?	Dispute notes, refund reasons, escalation outcomes.

Marketplace trust compounds when good participants feel the rules are clear, fair, and enforced.

Marketplace Supply Health Review

Marketplace founders often chase demand too early. If supply quality is weak, unavailable, slow, unreliable, overpriced, or poorly governed, demand generation creates disappointment instead of liquidity.

Review supply health weekly in the atomic market:

Supply signal	Founder question
Active supply	How many suppliers are actually available now, not merely listed?
Response rate	How quickly and reliably do they respond?
Completion rate	Do accepted jobs/orders/transactions finish successfully?
Quality distribution	How many suppliers create most of the good outcomes?
Earnings or benefit	Is supply making enough money, utilization, leads, reputation, or savings to stay?
Cancellation rate	Who cancels and why?
Dispute rate	What quality, price, timing, fraud, or expectation issues repeat?
Concentration risk	Does liquidity depend on a few suppliers or one geography/category?

Use this diagnosis:

If supply is	Then
Plentiful but inactive	Improve demand quality, communication, and expected earnings.
Active but low quality	Tighten admission, training, ranking, reviews, and enforcement.
High quality but scarce	Narrow geography/category, increase supply acquisition, or improve economics.
Responsive but not converting	Diagnose pricing, buyer intent, trust, and match quality.
Converting but churning	Improve earnings, workflow, support, settlement speed, and fairness.

For Indian marketplaces, supply-side reality can be informal: phone calls, WhatsApp, cash expectations, local trust, GST/invoice differences, regional language, and relationship-driven behavior. Do not assume a listing equals supply. Count only supply that can deliver the promised outcome repeatedly.

The best early marketplace work is often unglamorous: calling suppliers, verifying quality, improving payouts, setting rules, fixing disputes, and removing weak participants. That work is the product.