Some lessons for building online marketplaces…

When we were first building, I didn’t realize how generalizable some of what we learned about building online marketplaces would turn out to be. Now, after helping to build several more such marketplaces (such as the volunteer management system for the Obama Campaign), I’ve started to see some commonalities and when I mentor entrepreneurs who have marketplace as part of their plans, I find myself giving a lot of the same advice. This post is an attempt to boil some of that down.

What do you mean by marketplaces, anyway?

For purposes of this posting, I mean any online property where two unrelated parties come together to consummate a transaction. It could be online dating, a shoe store, home-sharing or a job finding site. In all of these cases, there are a few things that always happen:

  • Search. Parties need to find each other. Half of online marketplaces are really just “black book” providers: they don’t add a ton of value other than to connect the parties.
  • Establish Trust. The next level of value add is to establish trust so the parties have confidence doing business. With the possible exception of dating, the minimum here is ratings.
  • Community. Virtually all marketplaces combine trust-creating feedback mechanisms with best-practices, online support and other services from community members.

Why marketplaces are great businesses.

At the end of the day, what makes online marketplaces successful and what gives them staying power is their base of users. This has been called many things: “network effect,” a “liquid marketplace,” etc., but what it comes down to is that once you have a group of buyers and a group of sellers doing transactions, it’s really hard for the next guy to come in and compete with you. Even if they have a more streamlined, feature-rich and fancy product, people tend to stay where the action is. Don’t believe me? Then here are some of my favorite examples that are terrible web sites but still dominate their categories:

  • eBay. No significant UI improvements in a decade, it’s still a bear to use. But they still own the online auction category.
  • CraigsList. They are downright proud of the fact that they are nothing but a fancy listserv. But find any classifieds site with even a fraction of their market share if you can.
  • Alibaba. The $40bn web behemoth that very few have heard of – unless you do overseas sourcing in which case you live in its virtual dark alleys that are every bit as scary as anything in Hong Kong or Taipei.

That doesn’t even include reasonably well maintained services like which continue to be able to charge hefty monthly fees against free competitors. Life is good if you are a successful online marketplace.

Sounds great. How do I do it?

So this is where the magic is. Once you find an underserved niche (and there are many of them), you build your responsive site and mobile apps and up you go. Pretty soon you start marketing, and people will flock to you. Right?

No. Unlike straight up B2C, you can’t control fulfillment. You need a balance of buyers and sellers. Too many of one or too few of the other and the marketplace fails. People don’t give you a second chance. If they come to your beautiful site with no counter-parties present, it’s like walking into a shop with nothing on the shelves. It’s sometimes called the “empty house problem,” especially by social media types.

Filling the Empty House.

So how do you fill the empty house? There are a few strategies that I’ve found to be successful, but the most important thing to do is break down the problem.

You have a big idea. It has a massive addressable market. But that’s also going to limit its early success. It may have global reach, but two parties in the same country (or even city) are more likely to do business that two people spread around the world. And the more you tighten the initial message, the more like you are to have be able to get things going.

Think of some specific use cases, and develop, sell and market to those. They won’t limit you later: instead, they’ll create a nucleus to grown from. Some obvious ways to limit your initial scope would be by geography (Uber is growing city-by-city), by sub-category (remember when Amazon only sold books? Yes, I’m old.), or by deep value-add to one specific use case (Sittercity developed a four-step screening process for parents and babysitters to use.)

By breaking down the problem, you reduce the number of people you need to get the party started from tens of thousands to hundreds or even dozens, a much more manageable number.

Go Deep. And Ignore the Data.

When you are starting a marketplace, so much depends on the founders’ knowledge of the sector and intuition. With only a few hundred to a few thousand initial users, by definition you don’t have enough data to be statistically significant, so pretending you are doing data-driven decision-making is dangerously self deceiving. At the beginning, it’s much more important to go deep in the process: talk to users and make sure that the entire flow of the marketplace works as intended. Think in terms of unit economics, not market economics. Does the user get value out of the system? Am I charging reasonably based on the value that I generate (perhaps 20%)? And is the process fairly streamlined?

Most online marketing techniques are the reverse of general brand marketing: rather than achieving economies of scale, incremental users are more and more expensive as the early adopters get sucked away and the cheap keywords get bought up. Scaling a low or negative margin business does not bring you to profitability, it brings you to tears.

Get the marketplace working right and then scale the hell out of it.

My marketplace would be so great…if not for all the terrible users.

It’s the law of big numbers. We want everyone on the Internet to be nice, businesslike people. The customer is always right. Everyone should have their say. Right? Right?

No. Actually, not right. In the real world, people’s tendency to go pear-shaped or try to get away with crazy stuff is moderated by the fact that they actually have to show up and, sometimes, risk arrest in the process of screaming, bullying, pushing or carrying on. In the online world, the barrier to entry drops to zero. People feel justified in going to town on one other without knowledge and often just to self vindicate.

Once the number of users on your site gets large it will start to represent the community at large. You will have just as many felons, phishers and bullies as anyone else if you don’t take action. Reviews help in some cases (and usually they are the minimum price of entry in setting up an online marketplace), but they aren’t enough. Sometimes, review systems can even be used by bullies to extract concessions (remember when Yelp! merchants were reporting virtual racketeering?)

You also need other systems. “Report as inappropriate” links help. So do rudimentary content filters (look for profanity, hate speech, etc.). But you also need to know what to do with bad actors when you find them. Deleting accounts doesn’t work. They just come back twice as angry.

One pretty successful technique is the “echo chamber” or “hell ban”. You create the capability to put a user’s account into a status where they seem to be able to interact with your service, but their postings become invisible to the rest of the world. They can fulminate and excoriate other users, but it does them no good. They can think their  messages send or their purchases clear, but you quietly put them in the bit bucket. After awhile, they generally lose interest and leave for greener pastures.

Final thought…

This is hardly a comprehensive guide to building online marketplace (I may post more on this), but it does address some of the most common newbie mistakes. If you have seen other approaches to these problems that work better, I’d love to hear from you. I think some of the functions above (and other useful ones) could be done with specialized tools or business process outsourcing. That’ll be the subject of a follow-up positing. Hopefully someone will take advantage of the free business ideas.


Comments are closed.