Marketplace Glossary

Marketplace

A platform where two or more separate groups of people transact with each other, and the platform takes a cut for making it possible. The defining feature is that you do not own the thing being sold. If you own the inventory, you have a shop.

Two-sided marketplace

The standard shape: one group supplies something, another group wants it, and the platform sits between them. Airbnb, Uber, Etsy. Both sides have to be present for either to get value, which is what makes the early stage hard.

See also: Peer-to-peer marketplaces explained

Multi-sided marketplace

Three or more distinct groups. A food delivery platform has diners, restaurants and couriers, each with their own app, their own economics and their own reasons to leave. Every side you add multiplies the coordination work.

Side

One of the distinct groups on your platform. Naming the side is the most useful discipline in marketplace product work, because almost every feature serves one side at the mild expense of another, and saying which one out loud prevents a lot of confusion later.

Supply side

The people or businesses providing what is being sold. Sellers, hosts, drivers, providers, practitioners. On most marketplaces this is the harder side to acquire and the side that gates everything else.

Demand side

The people buying. Usually easier to acquire, and usually where founders instinctively start, because demand acquisition is a marketing problem with familiar tools attached.

Provider

A supply-side participant who delivers a service rather than shipping a product. A cleaner, a tutor, a physiotherapist. Providers have calendars, travel time and capacity limits, which makes a service marketplace a different build from a goods marketplace.

See also: Service marketplaces

Managed marketplace

The platform actively controls who can join, what gets listed and often how the transaction is delivered. Higher quality, higher operating cost, slower to scale. Most successful marketplaces end up more managed than they set out to be.

See also: Learning and coaching

Open marketplace

Anyone can join and list, and the market sorts out quality through reviews and search ranking. Scales faster, harder to keep clean.

Vertical marketplace

Focused on one category, and usually built around the specific way that category works. A marketplace only for equipment rental, or only for legal services.

See also: Marketplace types

Horizontal marketplace

Spans many categories. eBay, Craigslist, Facebook Marketplace. Very hard to start from zero now, because breadth is where the incumbents are strongest.

Wedge

The narrow first slice of a market you go after, chosen because you can reach saturation in it. "Etsy for X" names a category. A wedge names a group small enough that you can plausibly get most of them.

Aggregator

Collects listings from elsewhere and sends the user off-site to transact. Comparison sites, job boards. Easier to launch because you do not need supply to sign up, and structurally harder to monetise, because the transaction happens somewhere you cannot see.

First-party inventory

Stock you own and sell yourself. Some marketplaces carry a little of it early to make the platform look full while real supply is still coming on. It works, and it does need a plan for winding down, because it competes with the sellers you are recruiting.

Multi-vendor

A technical description of a platform where many independent sellers list under one storefront, each managing their own products, pricing and fulfilment. Often used interchangeably with marketplace, though it usually implies goods rather than services.

Peer-to-peer (P2P)

Both sides are individuals rather than businesses. Adds friction on trust, because neither party has a company reputation to lose.

See also: Peer-to-peer marketplaces

B2B marketplace

Both sides are businesses. Longer sales cycles, larger order values, approval workflows, contract pricing and invoicing rather than card payments.

See also: B2B and procurement and Build a B2B marketplace

A-CAR

Attract, Convert, Adopt, Retain. Marketplace Studio's framework for sequencing marketplace work. The load-bearing idea is that it is a sequence rather than four parallel workstreams: at any moment one of the four is the actual constraint, and work outside it tends not to move anything.

Cold start problem

The state a marketplace is in before it has enough of either side to be useful. No buyer wants an empty marketplace and no seller wants a marketplace with no buyers, so nothing happens until something breaks the deadlock.

See also: The first twenty listings are hand-sourced

Chicken and egg problem

The same thing, described from the founder's chair. Which side do you build first when each one only has a reason to show up once the other is there.

See also: Supply before demand

Constrained side

The side that is currently limiting the marketplace. Identifying it correctly is most of the work, because effort spent on the unconstrained side produces very little movement and looks like progress the whole time.

Liquidity

The probability that a listing finds a buyer, or a buyer finds a match, inside a window short enough that they stay. The central concept in marketplace work, and the reason a marketplace is not just a website with listings on it.

See also: The marketplace liquidity playbook

Density

How much supply exists inside the boundary the buyer actually cares about. For a local service that boundary is a neighbourhood, not a country. Two hundred providers in one city is a functioning marketplace. The same two hundred spread across twelve cities usually is not.

Match rate

The share of buyer searches that return at least one result the buyer would plausibly accept. Distinct from returning any result at all, which is the version most analytics tools give you by default.

Time to match

How long from a buyer's request to an accepted match. In on-demand categories this is the number that decides whether the product works, because a buyer with an urgent problem will not wait for a callback.

See also: Home and local services

Fill rate

The share of supply-side capacity that gets used. Empty calendar slots on a booking marketplace, unsold nights, unbooked hours. Low fill rate is how you lose supply quietly, because a supplier with an empty calendar concludes the platform does not work.

See also: Rental marketplaces

Search-to-fill rate

The share of searches that end in a completed transaction. The denominator is searches, not sessions or buyers, which is why the number looks worse than most founders expect the first time it is measured properly. A falling rate is the earliest signal that supply, matching or pricing has drifted.

See also: The marketplace liquidity playbook

Utilization

The share of listed supply that actually gets used: hours booked against hours offered, or nights filled against nights available. It is the number suppliers judge you on, and low utilization is the reason they leave long before they tell you they are leaving.

Supply acquisition

The work of getting the supply side onto the platform. Usually manual, usually unglamorous, and usually the thing that determines whether launch works.

Demand generation

The work of bringing buyers to the platform: search, paid acquisition, content, referrals, partnerships. The expensive version of this mistake is running it before supply is ready to serve what it brings, which turns a marketing budget into a bad first impression at scale.

Hand-sourcing

Recruiting your first listings personally, one conversation at a time, rather than through a signup funnel. It does not scale and it is not supposed to. It is how you learn what the funnel needs to say.

Single-player mode

Giving one side something useful before the other side exists. A tool, a template, a directory. It gives you a reason to talk to supply while you have no demand to offer them.

Seeding

Deliberately pre-loading one side of the market before opening the other. Often combined with restricting launch to a single city or category so the seeded supply looks dense rather than thin.

Conversion

On a marketplace, two separate things. Supply conversion runs from signup to a complete, live, high-quality listing. Demand conversion runs from arrival to a first transaction. They break for different reasons and respond to different fixes, so a single blended conversion number hides both.

Activation

The point at which a new user has done the thing that makes them a real participant. For supply that is usually a published listing, not a completed signup. Choosing this definition well matters, because everything downstream is measured against it.

Search-to-book conversion

The share of searches that end in a booking. Meaningless without its denominator: all sessions, all searches, or only searches that returned a result. Those three produce very different numbers from the same underlying behaviour.

Denominator

The bottom of the fraction. Most disagreements about marketplace metrics turn out to be disagreements about denominators, which is why every rate should carry its own.

Listing completion rate

The share of supply-side signups that end with a live listing. The gap between signup and published listing is usually the largest single leak on the supply side, and it is invisible unless you measure the two events separately.

White-glove onboarding

Manually walking early suppliers through setup: writing their listings, taking their photos, setting their prices. It does not scale, and at the start that does not matter. It is how most of the first twenty listings on a working marketplace got live.

See also: The first twenty listings are hand-sourced

Empty state

What a buyer sees when a search returns nothing. The most consequential screen in an early marketplace and typically the last one designed.

Zero-result search

A search that returned nothing. Worth logging as its own event with the query attached, because it is the cheapest supply research available: buyers telling you exactly what they wanted and you did not have.

Trust signal

Anything on the page that makes a stranger comfortable transacting. Reviews, verified badges, response times, photographs, a visible refund policy. On a new marketplace you have none of the ones that come from history, so the ones you design carry more weight.

Time to first transaction

Elapsed time from arrival to first completed transaction, measured per side. On a regulated marketplace this is often gated by verification rather than by anything in the product.

See also: Regulated industries

RFQ (request for quote)

A transaction pattern where the buyer describes what they need and suppliers respond with priced offers, rather than buying at a listed price. Common in B2B and in custom work. It trades instant checkout for flexibility, and it lives or dies on how fast the first quote comes back.

See also: B2B marketplace development

Second transaction

The most under-watched number in marketplace analytics, and a better predictor of whether the business works than anything about the first. A first transaction can be bought with a discount. A second one cannot.

See also: Why the second transaction decides it

Repeat rate

The share of users who transact more than once, inside a window appropriate to the category. Weekly for food delivery, annually for wedding services. Using the wrong window makes a healthy marketplace look broken.

Cohort

A group defined by when they joined, tracked over time. Marketplace metrics reported in aggregate are close to useless, because a platform adding users every month will show a blended figure that describes no actual group of people.

See also: Marketplace KPIs that actually matter

Frequency

How often a user transacts in a given period. Largely set by the category rather than by anything you do, and it determines almost everything about your economics, because low frequency means every transaction has to carry more margin. Weekly grocery ordering and once-a-year holiday booking are different businesses wearing the same interface.

See also: Travel and experiences and Food and beverage

U-factor

The rate at which participants who arrived on one side start using the other. A renter who becomes a lister, a buyer who becomes a seller. When it happens naturally, it reduces the cost of the harder side considerably.

Stickiness

Whether the platform accumulates something that makes leaving costly. Saved preferences, transaction history, a rating, a following. Distinct from satisfaction: people leave platforms they like.

See also: Community-driven marketplaces

Network effect

The property where each additional participant makes the platform more valuable to everyone already there. The reason marketplaces are winner-take-most, and the reason the second entrant into a mature category has such a hard time.

Direct network effect

Value increases with more users on the same side. Mostly a social-platform property, and rarer in marketplaces than pitch decks suggest.

Indirect network effect

Value increases for one side as the other side grows. More drivers make the service better for riders. This is the one most marketplaces actually have.

Local network effect

The effect operates inside a geography or category rather than globally. A platform can be dominant in one city and worthless in the next one over, which is why national expansion so often makes the product worse everywhere at once.

Multi-homing

When participants use several competing platforms at the same time. Drivers running two apps, freelancers listing on three sites. High multi-homing caps how much take rate you can hold and how defensible the position is.

See also: Talent marketplaces

Disintermediation

When the two sides meet through your platform and then transact off it, so you get the introduction cost and none of the revenue. Common in high-value, repeat-relationship services, where the second booking is worth going around you for.

Leakage

The measurable version of disintermediation: transactions you facilitated but did not capture. Hard to see directly, and usually visible in the pattern of a strong first booking followed by silence.

Churn

The rate at which participants stop being active. Measured separately per side, because supply churn and demand churn have different causes and very different costs to replace.

K-factor

How many new users each existing user brings in. Above one and the platform grows by itself. Almost nothing is above one, and quoting a K-factor without saying over what period is a common way to make a number look better than it is.

Defensibility

What stops a well-funded competitor from doing this to you. In marketplaces it is usually density in a specific market plus accumulated trust data, rather than anything about the software.

GMV (gross merchandise value)

The total value of everything transacted through the platform. Not your revenue. It is the number quoted in fundraising announcements and it says nothing about whether the business makes money.

AOV (average order value)

The average size of a transaction, before your cut. AOV multiplied by frequency and realized take rate is what one active buyer is worth to you, and that is what sets the ceiling on what you can afford to spend acquiring one.

Net revenue

What you actually keep. GMV multiplied by your realized take rate, minus payment costs. The number that matters.

Take rate

The share of each transaction the platform keeps. There is no correct number, only a number that follows from a decision about which side creates the value and which side can bear the fee.

See also: Take-rate is a strategy

Headline take rate

The percentage you publish. Usually lower than what you actually earn.

Realized take rate

Net revenue divided by GMV. What you actually earn once payment margin, promoted listings, subscriptions and fees are counted. The gap between headline and realized is where most marketplace revenue models live.

Commission

A take rate charged per transaction. The most common model, and the one that aligns your revenue with the market working.

Listing fee

A charge to publish, regardless of whether the item sells. Filters out low-effort listings and suppresses supply volume at the same time.

Subscription fee

A recurring charge, usually to the supply side, for access or better terms. Smooths revenue and disconnects it from whether transactions are happening, which cuts both ways.

Freemium

Free to use the core product, paid for something above it. In marketplaces it tends to work on the supply side, free to list and paid to be seen. It tends to fail as a replacement for a transaction fee, because the participants getting the most value are the ones you have decided not to charge.

Lead-generation fee

A charge per enquiry rather than per completed transaction. Attractive when you cannot observe the transaction, and it puts the risk on the supplier, who pays whether or not the work materialises.

Promoted listings

Paid placement in search results. A meaningful revenue line at scale, and one that trades against buyer trust in your rankings.

See also: Creator and content marketplaces

Payment processing margin

The difference between what your processor charges you and what you charge through. Small per transaction, invisible to both sides, and a real revenue line at volume.

Unit economics

What one transaction is worth to you after the costs directly attributable to it. If this does not work at one transaction, volume does not fix it.

Contribution margin

Net revenue per transaction minus the variable costs of servicing it. Support time, payment fees, insurance, refunds.

CAC (customer acquisition cost)

What it costs to acquire one participant. Calculated separately per side. A blended CAC across both sides of a marketplace obscures the only useful comparison, which is supply CAC against demand CAC.

LTV (lifetime value)

Total contribution margin expected from a participant over their life on the platform. Highly sensitive to the retention curve it assumes, and on a young marketplace that curve is still a projection rather than an observation.

Payback period

How long before a participant has generated enough margin to cover what you paid to acquire them. More useful than the LTV to CAC ratio on an early platform, because it does not require guessing a retention curve.

Split payment

One buyer payment divided automatically between the seller and the platform. Sounds trivial and is the single most common source of unplanned build cost, because tax, refunds and multi-seller orders all land here.

Escrow

Holding the buyer's funds until a condition is met, then releasing them. Necessary wherever the buyer cannot verify what they are getting at the point of payment, which is why it is standard in resale and authentication categories.

See also: Resale and recommerce

Payout

Moving money to the supply side. Timing is a product decision with real consequences: faster payouts are a genuine supply-side benefit and they increase your exposure to fraud and refunds.

Merchant of record

The legal entity responsible for the transaction, including tax and chargebacks. Whether that is you or your seller is a foundational decision that reaches into tax, compliance and dispute handling.

Trust and safety

The set of systems that let strangers transact. Identity, verification, reviews, moderation, disputes, fraud, payouts. Not a compliance function. On a marketplace it is a core part of the product.

Verification

Confirming something claimed by a participant is true. Identity, ownership, insurance, qualification. Each type has its own cost, its own refresh cycle and its own effect on how fast supply can come on.

KYC (know your customer)

The identity checks required before you can move money to someone. Triggered by payments regulation rather than by anything you choose, and it sits directly in your supply onboarding path.

Credentialing

Verifying a professional qualification and current licence to practise. Distinct from identity verification, because a licence is a state that expires rather than a fact that stays true.

See also: Regulated industries

Two-way review

Both sides rate each other. Improves behaviour on both sides and suppresses honest negative feedback, because people fear retaliation. The usual fix is publishing both reviews simultaneously after a cut-off.

Dispute

A transaction where the two sides disagree about what happened. You are the adjudicator, not a party, and the policy you write determines both your support cost and which side trusts you.

Chargeback

A buyer reversing a card payment through their bank. Expensive, weighted against you, and the cost usually lands on the platform rather than the seller.

Moderation

Reviewing listings, messages and profiles against your standards. Someone does this work whether or not you have planned for it.

Idea validation

Testing whether both sides actually want the exchange before building anything to host it: interviewing suppliers, pre-selling to buyers, and running the first transactions by hand. Weeks of this routinely save months of building the wrong thing.

See also: Pre-development

MVP (minimum viable product)

The smallest version that lets a real transaction happen end to end. On a marketplace this is larger than it sounds, because both sides plus the transaction have to work at once for anything to be learned.

Prototype

Something that looks and behaves like the product but does not transact. Useful for testing whether people want it. Not useful for testing whether the economics work.

Transaction engine

The logic that moves a transaction through its states: requested, accepted, paid, delivered, reviewed, refunded. Booking, order and quote marketplaces need different ones, which is why this is the decision that shapes most of the rest of the build.

No-code

Building with visual tools rather than written code. Viable for a real marketplace MVP now, with real ceilings on transaction volume, custom logic and cost at scale.

See also: The no-code marketplace builder guide

Sharetribe

A hosted platform purpose-built for marketplaces, shipping listings, the transaction engine, user management and Stripe payments out of the box, and extendable with custom code. One of several in its category, alongside Nautical, Carro and Randevu.

See also: The no-code marketplace builder guide

White label

Buying a pre-built marketplace platform and putting your brand on it. Fast and cheap to start, and constrained to whatever the vendor built, which matters most in the parts specific to your category.

Headless

An architecture where the customer-facing site is separate from the systems behind it. Relevant to you mainly because it affects how easily the front end can be changed later without touching the transaction logic.

Stripe Connect

The payments layer most marketplaces build on. It routes the buyer's payment to the seller, splits out the platform's fee automatically, and handles identity checks, payouts and tax reporting across countries. Choosing it also settles who your merchant of record is.

Marketplace bloat

Features that do not help supply meet demand, accumulated through competitor envy and edge-case requests. Bloat costs twice: once to build, and again every time it sits between a user and a transaction.