The first 20 years of the Internet could be described as a race to build the infrastructure to support a global heterogeneous bazaar—a single “place” where everything was for sale, from entertainment to services to physical goods.
But just as such bazaars have largely disappeared from the modern world, I strongly believe that we are currently witnessing the undoing of the Internet bazaar and the companies that exist to support it. In another 20 years we will look back at the Internet bazaar not unlike the way we today look back at physical bazaars as a quaint memory of the past.
In their place will be a series of vertical applications like Postmates, Shyp and Opentable that dominate trade in different categories and don’t need the support of a broader marketplace to draw consumers.
Stepping back in time, a series of companies have built fortunes building up the digital bazaar.
Yahoo is credited with building the first map, and trying to organize a handful of early disparate vendors. It wanted to create a directory of the chaos, in some ways making an early attempt at central planning.
Google decided to go visit each shop to catalog what they offered and rank them. Their guide was good enough that they actually ended up achieving more along the lines of structured planning than Yahoo ever did by creating incentives for merchants of all types of content and goods to conform to the implicit incentives of their algorithms.
Companies like eBay created platforms and infrastructure that made it easy for people to start selling, while a host of companies ranging from Verisign to Recaptcha built out bits of infrastructure that made participating in the bazaar as a buyer and seller easier.
Even Amazon, which in some ways always positioned itself apart from the free form mele of the bazaar, was historically a mega-stall of the bazaar, with a huge percentage of its customers accessing the store via search.
Open Air Market
This whole ecosystem of companies effectively were built and designed to make the bazaar model of the Internet better. Where that generally left us is that the Internet operated a lot like an open air market. It sold everything, but actually buying things was a high friction and trying experience.
For operators, though, bazaars are an appealing business. That’s because in essence they’re heterogeneous double-sided marketplaces, where many buyers and many sellers interface and compete with each other. Having more buyers attracts more sellers, and having more sellers attracts more buyers. They’re heterogeneous because buyers are not sellers and sellers are generally not buyers (contrast this with a stock market where presumably everyone is a buyer and a seller).
Business people love this because it is supposed to represent the ultimate in theoretical lockin. If you have all the buyers, you get all the sellers. If you have all the sellers, you get all the buyers, and by sitting in the middle and owning the platform you can extract rents and no one can leave. It is beautiful, and it is why things which have this dynamic – credit cards, HMOs, communication networks – end up being such wonderful businesses.
There are two ways in which the bazaar comes under pressure and is ultimately undone, both of which we are now witnessing online.
The first method comes when a class of sellers gets to the point that they can afford to leave. Bazaars basically rely upon certain daily purchase use cases, like groceries, to drive the buyers to the agreed upon spot every day. That means that the sellers that have very high repeat rates theoretically have the ability to splinter off from the bazaar, and if they can provide meaningfully better quality or pricing without the support of the bazaar itself.
Vertical Apps
You see this happening now with highly vertical apps. I have Instacart for groceries, Postmates for restaurants, Uber for transportation, Zeel for massages, etc. Google, would prefer I go to the open bazaar for these services, but these services can survive on their own without the encasing of the market because they are high enough value and high enough recurrence to warrant their own attention without the wider bazaar, and therefore can leave and stop paying their rents.
This type of splintering usually starts out at the high end, where the people with the most loyal customers and highest margins leave, which ends up being all the more devastating for the Bazaar which is forced as a result further and further down market. As the best merchants leave, the bazaar gets dirtier and smellier.
The second method of splintering off from bazaars happens when markets are self-contained and can survive without the broader context of the bazaar. I think of these pockets as usually homogeneous markets rather than heterogeneous markets. In these cases, the buyers and the sellers are the same people.
Looking historically, stock markets are the best example of this. In a stock market basically everyone is always a buyer and a seller of the same thing, stock. This means that even a small number of people who are deeply engaged can create a lot of volume or activity on their own without the engagement of the broader market.
Homogeneous markets
On the Internet the best example of these types of homogeneous double sided markets are social or communication applications. These services don’t need to have everything and they don’t need to bother with the complexity of coordinating disparate parties. They just have to create a compelling loop where each participant is both a producer and consumer of content. Because the buyers and the sellers are the same people, they are easier to get to splinter off from the bazaar and end up reinforcing themselves in the splintered community. Companies like Facebook, Instagram, and Snapchat represent, at least on the media side, fully formed homogeneous markets.
Homogeneous markets are extremely hard to build; however, the nice part about them is that because a single population is trading with itself, they can scale massively quickly when they start to get momentum— much faster than their heterogeneous brethren.
When I think about startups I am only interested in companies that are following one of these two patterns and blowing up the bazaar. They are either going vertical and building a much higher quality experience of accomplishing something that is possible but high friction in the bazaar, or they are constructing some sort of loop, or alternative marketplace which has the ability to grow at warp speed because it is a homogeneous market rather than a heterogeneous market.
But I get very wary when I see companies that are trying to build new heterogeneous double-sided marketplaces, on an old model, or are building tools and applications around the assumption that the internet is one big dirty, messy, smelly, confusing bazaar.
The Internet bazaar isn’t going to go away, just as there still is a market in the old city in Jerusalem. But we are rapidly witnessing it become a tiny shade of its former glory, and all the value will accrue elsewhere. While not unfathomable, it will take a truly epic technological disruption for the bazaar to flicker back to glory in our lifetimes.