Disruption vs. Enablement: Marketplaces
Like any good algebraist, the pilot is made to think sometimes in a forward fashion and sometimes in reverse; and so he learns when to concentrate mostly on what he wants to happen and also when to concentrate mostly on avoiding what he does not want to happen.
- Charlie Munger
This is part 2 of a 7 part series. If you haven’t done so already, you can read the first post and introduction here.
Marketplaces work best when creating new value by matching supply and demand more efficiently by making previously unavailable or difficult to attain or assess supply (and sometimes demand) easily accessible.
This is best suited for highly fragmented industries where supply is to a degree, a commodity. The most commonly used example is Uber whereas customers just want to get from point A to point B so when you hail a car on your phone, you choose Uber because it’s the fastest, cheapest or highest quality (depending on consumer preference) and as long as it shows up and takes you from point A to point B, you’re a happy customer and there is little nuance involved.
When supply is commoditized, frequency of purchase is extremely important. Due to the fact that supply switching costs are low in these markets, the take rate will need to be a small percentage of the transaction, otherwise you will attract competition that sees your margin as their opportunity (hat tip to Jeff Bezos for that one). This gives a marketplace less room to play around with acquisition costs and therefore when users are acquired, they must be kept and the vast majority must be profitable relatively quickly. Hence why frequency of purchase is an important characteristic for marketplaces with commoditized supply.
Marketplaces can also work in industries where supply isn’t a commodity but the suppliers don’t benefit from economies of scale due to certain industry dynamics (largely where supply is high touch services jobs/knowledge workers). Examples in our portfolio here include UpCounsel (marketplace for legal services) and Talkspace (marketplace for therapy). With these types of companies, startups build previously non-existent economies of scale by building technology to help route customers to service that best suites their needs as well as making the process friction-less on both sides. This helps the supply-side serve more customers while the demand-side enjoys better economics.
Because the tasks these companies provide are more complex than the previous example, there is usually a human involved in providing the service. This results in higher average transaction sizes and an opportunity for the startup to provide more value in making the service seamless for all parties. Therefore, purchase frequency is less important because it’s the marketplace’s job to maintain the liquidity of supply and therefore charge higher take rates.
It’s important to note that this doesn’t work for all complex tasks. If there is low fragmentation or a high degree of differentiation in quality of supply, it can be much harder to make the math work. To give an example, consider this quote from Y-Combinator partner and former founder and CEO of Tutorspree, Aaron Harris:
There is a varying degree of quality when it comes to tutors, and as a result, building a marketplace couldn’t alleviate parents’ hesitation to purchase services on Tutorspree. After realizing this, Tutorspree adapted their model to act more like a traditional agency, providing more hands-on services and using technology to match supply and demand. This model worked well for awhile, before earlier decisions on customer acquisition came back to bite the company which Aaron wrote about here.
Finding a market ripe for marketplace disruption is far rarer than it seems at first glance. Some markets have fragmented supply but aggregating the supply doesn’t provide for economies of scale that allows participants to benefit financially.
When looking at the cost structures and margins of incumbents, it’s important to discern their size and scale when determining their reaction to a new entrant. In industries that are so fragmented that nobody enjoys enough economics to compete with the startup, we’ve seen unions and lobbyists turn to regulation as their last ditch attempt to stop the impending shift. This is often a sign of desperation, as technology makes the company’s initial structure and reason for fragmentation obsolete.
On the flip side, if the industry is made up of large companies with high head count and big budgets, the reaction might be to refuse or cut out the middleman and sell their products directly to customers while throwing their brand and massive customer acquisition budgets behind the effort. This is especially the case in markets where agents and brokers are commonplace.
Case Studies:
Worked: Uber
Funding: $8.7B
Outcome: Active
- Is supply a commodity? Yes. Most riders just want to get from point A to point B and don’t care about the type of car that takes them there, the driver, or the company providing that service.
- Is the purchase frequent? Yes. People will always need to get from one place to another (at least until teleportation becomes a possibility :) ) and therefore there is high purchase frequency.
- Is the task complex? No. Almost anyone can get a driver’s license and drive a car for Uber.
- Is there a human component necessary? Yes, for now someone still needs to drive the car but it’s not specialized or knowledge labor and therefore outside of customer service, Uber doesn’t need a large workforce of humans interacting with customers to help them with the purchase or the service.
Takeaway: Uber might be marketplace that best fits within this criteria and is an incredible company as a result.
Didn’t Work: Homejoy
Funding: $64M
Outcome: Shut down operations in July 2015.
- Is supply a commodity? Yes and no. While home cleaning isn’t necessarily a complex task, different customers have different expectations and demands when it comes to home cleaning so it’s difficult for a marketplace to ensure quality. This can result in a highly differentiated quality of supply which makes it difficult to scale the marketplace in the same way that a company like Uber did.
- Is the purchase frequent? This depends on the customer’s personal preferences and willingness to pay. For some, a cleaning once a month might suffice but others may want their home cleaned once a week.
- Is the task complex? As mentioned above, it’s not necessarily a complex task but individual preferences can make it complex for the marketplace to commoditize supply and therefore the complexity doesn’t allow for this market to scale in the same way that others might.
- Is there a human component necessary? Yes, someone needs to clean the homes and a larger number of home cleaners doesn’t necessarily reduce the costs for the end user like it does with Uber. The supply side economy of scale isn’t nearly as powerful when the cleaner still needs to spend two hours at the customer’s home since the price is dependent on hourly wages and apartment size.
Takeaway: While initially Homejoy checks a lot of these boxes, individual preferences and fixed hourly rates make the commoditization of supply difficult. These supply side challenges as well as the lack of predictability on the demand side, proved to be difficult challenge for Homejoy to overcome.
