Price Discovery in Marketplaces

By Brian Tr · Updated Oct 7, 2026

Every marketplace has to answer a question that a normal store never faces: who decides what something costs? A store owns its inventory, so it sets the price. A marketplace sits between people who each have a view on value. The seller knows what the item cost or what the hour of work is worth. The buyer knows what it's worth to them. The platform's design decides how those two views meet.

That meeting is price discovery. This article covers the main ways marketplaces do it, what each one trades away, and what research and regulators have said about the choices. We won't cover how much the platform charges on top. That's a separate subject, covered in take rate and revenue models. Here the question is only how the transaction price itself gets set.

Why price discovery is the platform's problem

Economists have long treated prices as more than a number to pay. Friedrich Hayek's 1945 essay, The Use of Knowledge in Society (American Economic Review, vol. 35, no. 4), describes the price system as a kind of telecommunications system. People who don't know why something became scarce can still adjust, because they can watch a few "pointers" move. No central authority has to collect everyone's knowledge.

A marketplace is a small version of that problem. Thousands of sellers hold private knowledge about cost, quality and availability. Thousands of buyers hold private knowledge about need and budget. If the platform picks a price mechanism that hides this information, matches suffer. Sellers underprice or sit idle. Buyers overpay or leave. If the mechanism surfaces it cheaply, more trades clear.

Three questions sit underneath every pricing design:

1. Who proposes the price? The seller, the buyer, the platform, or an algorithm.
2. How many rounds does it take? One click, or a back-and-forth that can last days.
3. How much does the platform control the outcome? From none (a seller types a number) to all (the platform quotes it).

The five main mechanisms

Almost every marketplace uses one of these, or a mix.

1. Seller-set fixed prices (posted prices)

The seller lists a price. The buyer takes it or leaves it. This is how most online retail and most consumer marketplaces work today.

It's fast and simple. A buyer can compare listings in seconds, and the transaction needs no waiting. The cost is that the seller has to guess the right price, and a wrong guess goes unnoticed. Price too high and the item doesn't sell. Price too low and the seller leaves money on the table with no signal that they did.

Platforms often compensate with search ranking, filters and price comparison, which is why search and matching and price discovery are tightly connected. When listings are easy to compare, competition does the price discovery that an auction would otherwise do.

2. Platform-set prices

Here the platform decides. The rider doesn't negotiate with the driver, and the seller doesn't choose the fare. The platform computes a price from distance, time, demand and supply, and both sides accept it or don't participate.

Platform-set prices give the platform the most control and the buyer the most predictability. They suit services that are standardized enough to price from a formula, and markets where haggling would slow everything down. The cost is trust on the supply side: suppliers who feel the platform sets prices too low have few ways to push back, and some leave or pursue multi-homing. We'll come back to dynamic versions of this in a moment.

3. Auctions

In an auction, buyers compete and the price emerges from their bids. There are a few forms worth knowing:

• English (ascending) auction. Bids go up in public until nobody will outbid the leader. This is the classic image of an auctioneer.
• Sealed-bid auction. Each buyer submits one bid privately, and the highest wins. In the second-price variant, the winner pays the second-highest bid, which in theory encourages bidders to state what they'd truly pay.
• Proxy bidding. This is eBay's twist. A buyer enters a maximum, and the system bids in increments on their behalf, only up to that limit. The buyer pays the lowest price needed to win, not necessarily their maximum. In effect it turns a long English auction into something close to a second-price one, without anyone needing to watch the clock.

Auctions are good at discovering the price of things nobody can easily value: one-off items, collectibles, used goods. They're poor at speed. Someone has to wait for the auction to end, and the buyer who wants it now can't have it.

4. Negotiation and quotes (RFQ)

In many B2B and services marketplaces, there's no list price at all. The buyer describes the job, and suppliers respond with a quote. This is the request for quote, or RFQ. It's common wherever the product is customized: freight, construction, agency work, industrial parts, and home services where the scope depends on the house.

The mechanism fits because the price depends on details only the buyer can describe and only the supplier can estimate. The tradeoff is time and friction. Each quote takes effort to write, and a buyer may wait days. That friction is also where leakage shows up, since once a buyer and supplier have agreed on scope, they can take the deal elsewhere. Our B2B vs B2C marketplaces article covers why quote-based pricing is more common on the B2B side.

Some platforms reduce the friction by offering instant quotes for standard jobs and keeping manual RFQ for the complex ones. A managed marketplace goes further and has the operator produce the quote itself.

5. Dynamic and surge pricing

Dynamic pricing means the price moves with conditions. When demand outruns supply, the price rises. When supply is plentiful, it falls. Ride-hailing's surge pricing is the best known case, but hotels, flights, event tickets and some delivery services work similarly.

Dynamic pricing does two jobs at once. It rations scarce supply to the people who value it most, and it signals suppliers to show up where they're needed. That's Hayek's mechanism, run by an algorithm and updated every few minutes.

Researchers have been able to study it because the algorithm creates natural experiments. A 2016 NBER working paper by Peter Cohen, Robert Hahn, Jonathan Hall, Steven Levitt and Robert Metcalfe, Using Big Data to Estimate Consumer Surplus: The Case of Uber, used Uber's surge algorithm and a regression discontinuity design on almost 50 million observations. They estimated that UberX generated about $2.9 billion in consumer surplus in four US cities in 2015, or about $1.60 of surplus for every dollar spent. That estimate says something about demand, not about whether surge is fair, and the paper doesn't claim otherwise. But it shows that a platform-set, dynamic price can be studied like a market price, because the platform leaves a clean trail of how buyers responded as price changed.

The cost is perception. Riders who see a price double at the worst moment don't think about rationing. This is a trust issue, and platforms commonly respond by showing the price before confirmation or limiting how fast it can move, though practices vary by company.

What eBay's shift tells us

The best-documented real-world experiment in price discovery is eBay's move from auctions toward posted prices. Liran Einav, Chiara Farronato, Jonathan Levin and Neel Sundaresan studied it in Auctions versus Posted Prices in Online Markets, published in the Journal of Political Economy in 2018 (vol. 126, no. 1, pp. 178-215). Their data covers eBay listings from 2003, when auctions dominated, to 2009, when posted prices had overtaken them.

Their framing is the most useful single idea in this article. They model the choice as a tradeoff between competitive price discovery and convenience. Auctions find the price through buyer competition but cost time. Posted prices are convenient but depend on the seller guessing right. What they found on eBay fit that logic:

• Auctions were favored by less experienced sellers and for idiosyncratic products.
• Auction listings sold at a discount but with higher probability than comparable posted-price listings.
• The decline in auctions wasn't mainly because sellers and items changed. Seller incentives changed, through falling sale probabilities and a fall in the relative demand for auctions. Their estimates suggest the latter mattered more.

The paper also notes that proxy bidding helped eBay grow by letting buyers join a dynamic auction without paying constant attention. Convenience helped the auction era too, and eventually helped end it.

The takeaway for a marketplace designer isn't "auctions are dead." The same paper discusses why sellers might use a mix to price discriminate. It's that the right mechanism depends on how uncertain the item's value is, how much retail competition exists, and how much buyers value speed.

Comparing the mechanisms

Most mature marketplaces end up with more than one. They default to a fast mechanism for the common case and keep a slower one for the edge.

Pricing tools and suggestions

Between "seller sets it" and "platform sets it" there's a middle option: the seller decides, but the platform helps. Suggested prices, price ranges, "competitive price" badges, and automated repricing tools all fall here.

The logic is that sellers, especially new ones, are bad at pricing. Weak pricing hurts the whole marketplace, because overpriced listings don't convert and underpriced ones erode supplier earnings. A suggestion nudges listings toward a range where they sell, without removing seller choice. This matters most at onboarding, since a first listing is where a new seller is most likely to guess wrong.

There are tradeoffs to watch:

• Anchoring. A suggested price pulls sellers toward it, which can narrow the price range and reduce real competition.
• Herding. If many sellers follow the same tool, prices converge and the market stops revealing information.
• Transparency. Sellers need to know whether a suggestion reflects real transactions or something the platform wants them to do.

Which brings us to the legal edge.

The regulatory edge

Price-setting by a platform is a form of market power, and regulators have started to treat it that way.

Transparency rules in the EU. Regulation (EU) 2019/1150, known as the platform-to-business regulation, requires online intermediation services to set out in their terms and conditions the main parameters determining ranking and the reasons for their relative importance, including whether payment can influence ranking. It also requires plain, intelligible terms and at least 15 days' notice before changing them, with written reasons before restricting or suspending a business user. It doesn't set prices, but it governs the rules that shape which listings, at which prices, buyers see.

Shared pricing algorithms and antitrust. In the US, the Justice Department's January 2025 amended complaint in the RealPage case alleged that six large landlords used common pricing algorithms and each other's competitively sensitive information to set rents. RealPage isn't a marketplace, and the case is about landlords, not platforms. But the theory carries over to any tool that recommends prices to many competing sellers using pooled, nonpublic data. The more sellers follow one recommendation, the more it resembles coordination.

That doesn't mean pricing suggestions are unlawful. It means a platform that offers them should know where its data comes from, whether competitors' nonpublic information feeds the recommendation, and whether sellers truly stay free to decline it. Get legal advice for your jurisdiction. This article describes the issue and isn't legal guidance.

Choosing a mechanism: a practical sequence

These steps are reasoning from the above, not data.

1. Ask how standardized the product is. The more standard, the more a posted price works. The more unique, the more an auction or quote helps.
2. Ask how fast the buyer expects to transact. Speed pushes toward posted or platform-set prices. Patience allows auctions and quotes.
3. Ask who has the information. If sellers know value better than the platform, let them price. If the platform sees more of the market, let it set or suggest.
4. Start with the simplest mechanism that clears the market. Add auction, quote or surge logic only where the simple one demonstrably leaves value on the table.
5. Plan for leakage and trust. Slower, negotiated mechanisms create more chances to move off-platform. Algorithmic ones create more chances for perceived unfairness. See payments for how keeping the money flow on-platform changes the incentive to leave.

The mechanism also shapes conversion. A buyer who has to wait for a quote drops out more often than one who sees a price. That's covered in the marketplace conversion funnel, and the related choice of instant versus request-based booking in instant book vs request to book.

Key Facts: Marketplace Price Discovery

• Einav, Farronato, Levin and Sundaresan frame the auction versus posted-price choice as a tradeoff between competitive price discovery and convenience, using eBay data from 2003 to 2009 (published in the Journal of Political Economy, 2018).
• In that eBay data, auction listings sold at a discount but with higher probability than comparable posted-price listings.
• eBay's automatic bidding places bids on the buyer's behalf up to a maximum, and the winner pays the lowest price needed to win.
• Cohen, Hahn, Hall, Levitt and Metcalfe estimated that UberX generated about $2.9 billion in consumer surplus in four US cities in 2015, about $1.60 per dollar spent, using Uber's surge algorithm (NBER Working Paper 22627, 2016).
• Hayek's 1945 essay describes the price system as a system of telecommunications that lets people adjust without knowing the underlying cause of a change.
• EU Regulation 2019/1150 requires online intermediation services to disclose the main parameters determining ranking and whether payment can influence it.
• The US Justice Department's January 2025 amended complaint alleged that six landlords used common pricing algorithms and competitors' sensitive information to set rents.

Related Reading

• Marketplace Take Rate
• Marketplace Revenue Models
• B2B vs B2C Marketplaces
• Managed Marketplace
• Search and Matching in Marketplaces
• The Marketplace Conversion Funnel
• Instant Book vs Request to Book