Auction Reserve Prices: Maximize Expected Value, Not the Best Hammer Price

In this article

The direct answer: a sound auction reserve is not merely “the lowest price the seller likes.” It is the threshold that maximizes expected value after accounting for sale probability, net proceeds if sold, the outside option if unsold, and the cost of delay and relisting. A higher reserve can improve the price of successful sales while reducing sell-through, so average hammer price alone is a misleading KPI.

eBay’s guidance defines a reserve as the minimum amount the seller will accept and states that the item remains unsold when it is not met. That is a useful operational definition, but setting the threshold requires category data, platform rules and local legal review.

Separate the reserve from the opening bid

The opening bid shapes entry and the path of competition. The reserve protects the seller’s outside option. Combining the two creates predictable errors: a high opening price can discourage participation, while an excessive hidden reserve produces many no-sales even though the few successful auctions look excellent in a price report.

The correct unit of analysis is expected value per listing:

sale probability × net proceeds if sold + no-sale probability × net outside-option value − auction preparation cost.

The outside option may be a direct sale, relisting or continued internal use. Deduct storage, deterioration, administrative effort, delay and any expected price decline rather than using an optimistic appraisal.

Hypothetical example: used commercial equipment

Assumptions: these figures are not results from a real marketplace. Auction preparation costs QAR 2,000. If the lot does not sell, its outside option is worth QAR 74,000 net of delay costs.

• QAR 80,000 reserve: 85% sell-through probability and QAR 99,000 average net proceeds conditional on sale. Expected value = 0.85 × 99,000 + 0.15 × 74,000 − 2,000 = QAR 93,250.
• QAR 100,000 reserve: 60% sell-through probability and QAR 112,500 average net proceeds conditional on sale. Expected value = 0.60 × 112,500 + 0.40 × 74,000 − 2,000 = QAR 95,100.

The higher reserve wins in this example despite lower sell-through. If the outside option fell to QAR 50,000, however, no-sale risk would become more expensive and the conclusion could reverse. The reserve is therefore not a fixed percentage of appraisal; it is a decision based on the bid distribution and the strength of the alternative.

Build an auditable reserve policy

1. Segment the inventory. Separate liquid categories with abundant comparables from rare, seasonal or deteriorating assets.
2. Estimate the sell-through curve. For each reserve band, measure the probability of reaching it and net proceeds when it is reached. Exclude auctions affected by outages or exceptional promotion.
3. Value the alternative honestly. Use realizable net cash and timing, not a theoretical valuation.
4. Assign change authority. Document who may amend a reserve and when. Changing rules after attracting bidders can damage trust even if technically permitted.
5. Test within a controlled range. Compare similar groups of lots, not two isolated auctions, and monitor repeat participation as well as sale value.

Connect the reserve to the wider auction design

A reserve does not operate alone. Use bid increments that keep competition moving near the threshold, and set an auction bidder deposit that limits no-shows without suppressing liquidity. Procurement auctions have a different objective: use evaluated cost in reverse auctions rather than importing a seller-reserve rule.

Stress-test the number before approval

Every input is an estimate, not a fact. Recalculate with higher and lower sell-through, a 10% or 20% change in the outside option, and one extra month of storage. A reserve that remains best across most scenarios is robust. If a small assumption change reverses the result, do not give the seller a single confident number; show a range and explain the trade-off between speed and price protection.

Selection bias is another trap. The average price of sold auctions excludes failures, making a high reserve look better than it is. Capture every listing outcome: sold, unsold, withdrawn and relisted. Link qualified bidder identities across events so the platform can distinguish broad liquidity from repeated activity by a small group.

Metrics, limits and failure modes

Track expected value per listing, first-listing sell-through, qualified bidders, reserve attainment, net days in inventory, and the eventual result after relisting. Also measure a trust signal: how many active bidders stop returning after repeated “reserve not met” outcomes?

The model becomes unreliable when samples are small, inventory mix changes, sellers impose emotional floors or outside-option values are inflated. In those cases, use a reviewed range with expert override rather than an automatic point estimate.

Decision: choose one asset category, model three reserve scenarios and include the no-sale alternative. Adopt the threshold that improves net expected value across multiple listings—not the one that creates the highest screenshot-worthy hammer price.

اقرأ النسخة العربية