Ramesh Johari on Building Online Marketplaces

50.2K views
•
November 9, 2023
by
Lenny's Podcast
YouTube video player
Ramesh Johari on Building Online Marketplaces

TL;DR

A marketplace creates value by reducing the transaction costs that prevent supply and demand from finding and choosing each other. Because both sides depend on that service, operators must treat both as customers, use data science to improve matching, and recognize that reallocating attention or inventory often creates winners and losers.

Transcript

Marketplaces are a little bit like a  game of whac-a-mole. One example that   I came across with one of the companies I  worked with that I love is our new supply side was having a pretty bad experience. So what we decided to do is build some   custom bespoke features that were really going to  direct them to more experienced folks on the other sid... Read More

Key Insights

  • A marketplace sells the removal of transaction costs, rather than the underlying room, ride, job, or service. Suppliers provide the underlying offering, while the platform reduces the difficulty of discovering and connecting with a willing counterparty.
  • Both sides of a marketplace are customers of the platform. Riders and drivers depend on Uber to connect them, just as guests and hosts depend on Airbnb, so the platform must reduce friction and provide value for supply as well as demand.
  • Market failure can occur when willing buyers and sellers cannot find each other. A traveler may need a room or a rider may need transportation, yet the transaction cannot happen unless the participants can discover availability, willingness, timing, and suitable terms.
  • Data science is central to digital marketplaces because technology allows their structure to be changed continually. Unlike ancient physical markets built from stone, online platforms can repeatedly adjust how participants discover options, receive recommendations, and connect with one another.
  • Finding possible matches is a core marketplace problem. Guests need to locate hosts with available listings, hosts need to reach interested guests, and job posters need to identify applicants who might fit their work before any choice or transaction can occur.
  • Making the match is distinct from finding candidates. Once multiple options are available, a participant still needs help deciding whom to interview, hire, book, or otherwise select, making ranking and recommendation important parts of reducing marketplace friction.
  • Marketplace interventions often create winners and losers because attention and inventory are limited. Directing new suppliers toward experienced participants may improve onboarding metrics while reducing the opportunities or experience available to existing participants elsewhere in the market.
  • The value of a marketplace change depends on whether its winners matter more to the business than its losers. Operators should not judge an intervention from one improved metric alone, because benefits for one group can shift costs onto another group over time.

Install to Summarize YouTube Videos and Get Transcripts

Explore YouTube Video Summarizer or Get YouTube Transcript Extractor

Questions & Answers

Q: What does an online marketplace actually sell?

An online marketplace sells a reduction in transaction costs, meaning the friction that keeps willing participants from finding and choosing one another. Airbnb hosts sell listings, and Uber drivers sell rides. The platforms make those transactions easier by helping participants discover availability and suitable counterparties. Their central value is therefore not the underlying room or ride, but the removal of obstacles surrounding the exchange.

Q: Why are both buyers and sellers marketplace customers?

Both sides are customers because each depends on the platform to remove a different version of the same friction. A rider needs to find an available driver, while a driver needs riders in order to earn money. A guest needs suitable accommodation, while a host needs someone willing to book an available listing. The marketplace must therefore create value for supply and demand, not only for the purchasing side.

Q: How do marketplaces reduce market failure?

Marketplaces reduce market failure by making willing counterparties visible and reachable to one another. A person may need transportation or accommodation while another person is willing to provide it, but no exchange happens if they cannot discover each other, confirm availability, or assess suitability. The platform organizes the information and matching process that allows an otherwise difficult transaction to occur.

Q: Why is data science important for online marketplaces?

Data science helps online marketplaces find potential matches and decide which connections to prioritize. Digital platforms can continually change their architecture, unlike physical markets whose arrangements are constrained by buildings and booths. Data supports the systems that surface available listings, connect suppliers with interested customers, rank applicants, and guide participants among multiple possible choices, thereby reducing the friction that defines the marketplace opportunity.

Q: What is the difference between finding and making a match?

Finding a match means identifying the set of people or offerings that might satisfy a participant’s need. Making a match means choosing among those candidates. For example, a job poster may first receive several applicants and then must decide whom to interview or hire. A marketplace needs to support both stages because discovery alone does not resolve the difficulty of selecting the most suitable option.

Q: Why does marketplace management resemble whac-a-mole?

Marketplace management resembles whac-a-mole because improving one participant group’s experience can cause problems elsewhere. Directing new suppliers toward experienced people may improve the new group’s metrics, but it can divert attention away from existing participants. A later intervention may reverse or relocate that effect. Metrics consequently move as the platform reallocates limited attention and inventory among interconnected groups.

Q: How should marketplace teams evaluate winners and losers?

Marketplace teams should identify which participants benefit from a change, which participants experience a cost, and how important each group is to the business. An intervention should not be considered successful merely because one metric improves. Since consequential marketplace changes often redistribute attention or inventory, evaluation requires determining whether the newly created gains are more valuable than the losses imposed elsewhere in the system.

Q: What should founders focus on when defining marketplace value?

Founders should define their marketplace around the specific transaction friction it removes. Describing the company as a seller of rides, rooms, or services can obscure the platform’s actual role and lead its business model in the wrong direction. The founder should instead ask what prevents willing participants from transacting and how the platform can help both sides discover, evaluate, and connect with suitable counterparties.

Summary & Key Takeaways

  • Online marketplaces do not primarily sell rooms, rides, or other listed goods. Their suppliers provide those things, while the platform reduces the friction involved in discovering a suitable counterparty and completing a match. This distinction means that a marketplace must create meaningful value for participants on both its supply and demand sides.

  • Data and data science let operators continually redesign digital marketplaces. Johari identifies finding potential matches and deciding which match to make as central problems. Availability, preferences, applicant quality, and other marketplace information must be organized so each participant can identify and select a suitable person, service, listing, job, or opportunity.

  • Marketplace management resembles whac-a-mole because interventions redistribute limited attention and inventory. Helping new suppliers connect with experienced participants may improve one group’s results while worsening another group’s experience. Operators should therefore examine who gains, who loses, and whether the gains serve the business more meaningfully than the resulting costs elsewhere.


Read in Other Languages (beta)

Share This Summary 📚

Explore More Summaries from Lenny's Podcast 📚