tl;dr Marketplace = liquidity. Focus and fake as much as possible to get started.
Steve Schlaf has an excellent overview of the on-demand mobile services. Many, if not all, of these services represent modern evolutions of service marketplaces. This isn’t surprising, marketplaces from eBay to Kickstarter to Uber are a triumph of the power of the internet and now mobile computing. Goods and services that previously didn’t have sufficient demand in local markets can now market to the globe. Service marketplaces ticking Bill Gurley’s 10 Factors for a digital marketplace, large markets with high fragmentation, poor experiences, economic burden and network effects, can be rethought.
Schlaf’s map is partially representative of Andrew Parker’s earlier unbundling of Craiglist. The surprise isn’t that Craigslist is disaggregating, it’s that it is taking so long (note Andrew’s piece was written at the beginning of 2010). A marketplace that reaches liquidity is difficult to unseat even with as poor an experience and as diverse an offering as Craiglist.
The proliferation and funding of these service marketplaces reflects the value they have at liquidity. The lack of speed of this disaggregation reflects the incredible difficulty in reaching that liquidity. These two sided marketplaces have the classic chicken-and-egg cold-start problem. Below are some tips to drive the flywheel of marketplace liquidity.
As a quick background, I’ve spent most of my career focused on various marketplaces: search ads – Google, media – YouTube, local – PrimaTable, and most recently travel – Hotel Tonight.
Two Sided Marketplace
A two sided marketplace requires both buyers and sellers. Buyers are only interested in relatively valuable supply and vice versa.
To be successful, two sided marketplaces must build essentially two businesses. Uber must sell drivers on the value of their market relative to their other options through price, liquidity, aggregated demand and services like payments and insurance. At the same time, they also create a great customer experience and user acquisition through P&R, mobile user acquisition, and virality.
At scale, everything needs to be automated to facilitate liquidity. Until you get there manually faking the market can get the flywheel going.
Just as startups = growth, marketplace = liquidity. Let’s consider both sides of the market.
Supply and Demand
Faking the seller (supply) side of the market is often the easiest. For a seller, a new marketplace represents integrating a new channel for the possibility to make more money. In service marketplaces, this new channel is often exclusive and thus the risk associated requires a commensurate economic incentive in the form of relative value. Suppliers primarily care about liquidity – defined by wikipedia as “a market’s ability to facilitate an asset being sold quickly”.
Demand can be harder to build requires quality of supply (comprehensiveness and value) to enable trust in the market. This trust drives increased liquidity. Growing one side of a market offers increased value to the other due to network effects.
1. Single side utility
OpenTable built two different products with different value propositions. For top restaurants, not needing help generating demand, they built a CRM service to allow them to better services customers. The next tier of restaurants, aspiring to the top, joined valuing increased demand.
Google built a better search product, syndicated their service and aggregated user demand before offering a CPC ad model. For markets with consumer supply, this is similar to social networks covered in Andrew Chen’s fantastic post for solving social cold-start problems.
AirBnB famously both solicited existing Craiglist suppliers while simplifying syndication to a new Craigslist post. By leveraging other channel listings (often eBay or Craiglist), a marketplace can have comprehensive coverage of supply without a direct relationship with all suppliers.
Yelp focused on geography (SF) and vertical (restaurants), while actively recruiting power users, deriving value from participation, prior achieving utility to a broader set of users for local discovery. By limiting the market from all local search to primarily focusing on SF restaurants, Yelp was able to get to a critical mass and thus utility. They drove toward saturation and liquidity in their local market. Local discovery in particular is hard to get to ubiquity due to the sparsity of coverage as highlighted by Pat Kinsel’s recent post on the problem with local search. A network requires critical mass of coverage.
By explicitly purchasing inventory, marketplaces can unscalably generate initial supply. Rolling out Uber in a new market requires hiring drivers at an hourly rate to attract customers to the improved user experience. This inventory purchase is often subsidized but can represent an arbitrage opportunity. In a marketplace inventory purchase is not intended to last. This is different than one sided ecommerce businesses, like Amazon, where the model is to be the merchant of record and resell merchandise.
YouTube used MySpace embeds for both user acquisition and for broader media distribution. Early users derived value from uploading content to YouTube both on the YouTube platform and gaining initial distribution through embeds on existing channels.
A story about YouTube wouldn’t be complete without looking at their use of the DMCA. By adopting a relatively liberal (and ultimately vindicated) view of the DMCA, YouTube often had exclusive infringing content like rips of the Colbert Report and SNL. In markets where demand exceeds supply (e.g. collectibles, antiques, luxury products) – two sided marketplace problem can be solved by focusing on the in demand scarce inventory.
In lieu of ubiquity, curation can grease the wheels of a marketplace. Curation can increase efficiency from a user perspective due to the paradox of choice and restriction on mobile UI. It also limits the work necessary to sign up supply. Hotel Tonight didn’t launch a 1 to N search page to compete with mobile OTAs. By focusing on coverage of categories – luxe, charming, solid, and basic, instead of ubiquity, it benefits HT UX as well as simplifies the requisite supply sales.
This is liquidity brute force and is an extension of Paul Graham’s “Do things that don’t scale”. In time vs money delivery startups, like Postmates, offering broad local business coverage with explicit relationships can be prohibitively difficult. By manually fulfilling user demand, Postmates and others offer menus at merchants without agreements and send couriers to order like a normal customer. The network is able to offer better coverage than otherwise possible, driving like all tactics to ubiquity and liquidity.