Thread by @ljin18

Thread: We all know managed marketplaces (where the platform provides additional value-add in terms of intermediating the service delivery) can create a step-function improvement in user experience, but despite that, they can actually struggle to scale as businesses.

Examples of the functions that managed marketplaces often take on include vetting providers, handling logistics, or authenticating goods.

This translates into a much better user experience, but can also mean greater operational overhead, lower margins, and difficulty growing.

The complexity of intermediating the transaction is not trivial.

Consider Beepi--a managed marketplace for used cars--which created an amazing user experience (90+ NPS), but ultimately shut down due to challenges becoming unit economic profitable in many markets.

For Beepi, the costs associated with each transaction included rigorous inspections, free returns, delivery, and a guarantee to buy the car if it didn’t sell within a certain time frame.

In categories where services are viewed by users as commodities and/or users are especially price-sensitive, the open version of a marketplace often wins out over more managed versions.

For example, Craigslist and Thumbtack are much bigger businesses than marketplaces that have attempted to employ home services pros themselves and own the end-to-end experience.

But for more emotional categories that are more ongoing in nature or have variable outcomes, where each provider is meaningfully different, and the stakes or $ values are higher, the managed version often does better than the unmanaged version.

Here’s some examples where managed marketplaces have gained ground:
- Uber (more managed w/ centralized matching and pricing) vs. Sidecar (double opt in, no centralized pricing)
- Rover vs. finding a dog walker on CL/via WOM
- TheRealReal vs. unmanaged consignment marketplaces

What do these managed marketplace companies have in common?
1) Focus on a service category where users value the experience differential and are willing to pay a premium that offsets increased costs
2) Building software to improve efficiency
3) Focus on high AOV & high frequency

Another factor to consider is that labor/service marketplaces typically have more overhead than product marketplaces in terms of management.

That’s because every aspect of a service can be managed end-to-end, and there’s frequently a geo-specific aspect to the overhead, vs. a product marketplace can just manage specific aspects of the transaction (validation, logistics)--much of which can be outsourced / centralized.

With any unmanaged marketplace, there's going to be variance in quality. For categories where there’s less tolerance for uncertain quality/experience and desire for more consistency/certainty, a managed approach can win out.

For founders, the key is finding a way to provide a more delightful, managed experience where it matters, but doing so scalably, leveraging software.

For instance, designer goods marketplaces can reduce the # of items needing manual authentication by predicting which items are most likely to be fake.

In addition, validate that you can charge higher prices or increase take rate to cover the costs of managing the marketplace.

TL;DR: Managed marketplaces can create a step-function improvement is user experience, but they can be complicated to operate & challenging to scale. There's a number of factors to consider in whether to go unmanaged vs. managed, and to what degree.

For more on managed marketplaces, see our blog post: a16z.com/2018/11/27/ser…

and thanks @ai and @andrewchen for helping refine these thoughts!