Platform matching is how a platform filters, ranks and suggests the right partners for each user at the right moment, so that a deal or connection becomes likely. Its aim is the right choice, not the most choice.
Table of Contents
Key factors
- Judged by results. Good matching shows in completed, timely matches, not in clicks or the number of listings shown.
- Enough data. Personal recommendations work only when there is enough activity to learn from. Small platforms rely more on search and filters.
- Hidden capacity. When sellers are busy but this is not visible, buyers waste effort on them. Letting sellers show availability improves matching.
- Crowded markets. When applying is cheap, people apply too widely and the other side wastes effort screening. Limits on applications can help both sides.
- What to show. Showing every quality score can make some users hold out for rare top partners, leaving more people unmatched.
Examples
Hinge. In 2018 the dating app Hinge launched Most Compatible, a daily suggestion based on the Gale-Shapley algorithm, which pairs people likely to like each other in return. Hinge said that in tests, users were eight times more likely to go on a date with a Most Compatible match than with other profiles. The algorithm narrows choice to improve results.
National Resident Matching Program. In the US, graduating doctors and hospitals rank each other, and the National Resident Matching Program runs an algorithm that assigns doctors to residency places. It is designed so applicants cannot gain by misstating their preferences. The 2024 Main Residency Match offered a record 41,503 positions and filled 93.8%.
Frequently asked questions
What is the difference between matching and liquidity?
Liquidity is having enough of the right supply and demand on each side. Matching is how the platform turns that volume into successful pairings.
Who developed the theory of matching markets?
Economists Lloyd Shapley and Alvin Roth. Shapley’s 1962 work with David Gale on stable matching, and Roth’s later market designs, won them the 2012 Nobel prize in economics.
Should a platform show users every option?
Rarely. Too much choice slows decisions and spreads attention thin. Most platforms rank or limit options so users see a few good ones first.
How do you measure matching quality?
Track the share of searches or requests that end in a completed deal, the time it takes, and whether both sides come back.