A Crowded Space

All GMV is not created equal

For marketplaces companies, GMV (gross merchandise value) is the number 1 metric that most focus on. It’s the most commonly quoted stat when talking with press and investors. People use it as a barometer to understand the health and growth of the business. 

However, just like revenue, all GMV is not created equal.  Some GMV builds enduring value, and some is very fleeting.  High-quality marketplaces are constantly building bigger moats and strengthening their network effects.  Weak marketplaces suffer from disintermediation, pricing pressure, and destructive competition.  

Let’s look at some important factors to think about when dissecting GMV. These factors tend to revolve around the type, size, quality, and frequency of matching that happens in a marketplace as well as the nature of the goods or services involved.  Each of these factors can contribute to stronger network effects and help a marketplace build enduring value.  

Take rate - Marketplace companies take anything from 0 - 60% take rate. Most are clustered in a range of about 15-30%. Sometimes it’s better to take a low take rate to expand fast and gain market dominance. Sometimes it’s better to take a higher take and gain net revenue faster. At oDesk, we started with a 50%+ take rate but quickly lowered that to 30% and then 10% as we became more self-service and also wanted to lock up the market and prevent any disintermediation. In general, higher take rates make the GMV more valuable, but that doesn’t always mean you should try to increase the rake. 

Transaction size - I saw a marketplace recently with 44M GMV. The only problem was that over 90% of that came from a single transaction. I invested in OfferUp when they had less than 1M GMV per month (now well over 1B per month!) - the key difference here was that the 1M in GMV was comprised of several thousand small transactions. All things being equal - it is much, much better to have lots of small transactions than 1 big one. Every single transaction builds data, reputation, and lock-in. And often users view the matching component of a marketplace as a relatively fixed value - therefore, the acceptable rake on high price transactions is fairly low. Eg, if value of the match is perceived to be $20 - a $1000 item can yield a 2% rake. Or, 10 x $100 items can yield 10 x $20 or $200 for a 20% rake and you also gain 10 pieces of transaction history and feedback instead of 1. 

Transaction frequency - Usually inversely correlated with transaction size is the frequency. I don’t buy homes very often. I buy food and transportation very frequently. The more frequent the purchase, the more value I get out of the matching component of the marketplace. 

Homogeneity - if I use a marketplace to buy the same thing over and over again, I don’t have as much lock-in as if I use it to buy lots of different things. This can apply across verticals or to different sellers within a vertical. Consider Thumbtack vs Homejoy (Thumbtack is better). Or consider OfferUp vs Chairish (OfferUp is better). This comes back to the value of the match. I get more value from a marketplace if it continually matches me with new and different sellers. Uber matches me with a different driver every time. If I just used it to book the same driver over and over again, it wouldn’t be as valuable to me. 

Disintermediation - GMV is stronger if customers are effectively locked in. Could the transactions happen off your platform? For Uber - there is almost no way to work around the system and cut Uber out of the loop. For oDesk (Upwork), companies can certainly decide to take the payments offline or hire freelancers full-time. If the GMV of the marketplace is heavily locked into the platform and cannot easily be disintermediated, that makes it more valuable. 

Value after the match - all marketplaces perform some sort of matching function between buyers and sellers. The best marketplaces also add value after the match. In a labor marketplace, this could be in the form of customized software to actually perform the work. If this software can improve the efficiency and productivity of workers, you will have much better lock-in and a stronger marketplace.  For example, at Rev, buyers place orders with Rev and we quickly match a worker to the job.  The worker then executes the entire work on the proprietary transcription editor which is much more efficient than editing in Word.

Share of wallet - if your marketplace becomes the sole or primary channel for your users, then you have a strong marketplace. If you are only getting a small share of a users business, it is not as strong. As an example - drivers around SF are very frequently on both Uber and Lyft and seamlessly switch back and forth. Sellers on Etsy likely post their goods for sale on eBay and OfferUp and other sites as well. The key to gaining share of wallet really boils down to money. If you can make your platform the easiest way for a seller to make the most money, you win. This is why productivity matters so much! Highest productivity = highest earnings for sellers and lowest cost for buyers. 

Reputation and Data - an absolutely critical component of marketplaces is to build a virtuous cycle. This means that the service should naturally get better with scale. Does the feedback system mean that every month the average quality goes up. Does the quantity of users on the platform mean that liquidity naturally go up every month. If a marketplace has these virtuous cycles, given enough time, it will be impossible for any competitor to catch up if they’re starting from scratch.  The reputation contributes to network effects as does the increased data about all users that enables better matching and higher liquidity.  

Reliability - Do workers on your platform get work anytime they want it? Do sellers get close to 100% of items sold?  The reliability of a marketplace can be a key determining factor in customer churn.  If a marketplace is highly reliable and always gets me what I want, I have very little incentive to look around at other options.  In contrast, if I list a bunch of used items for sale on one app, and only 25% of the time it sells, then I will likely look around for other options to accomplish my goal.  If your marketplace delivers very high reliability, it drastically increases user retention and limits churn.  

So, what’s the point of all this? Just remember, that not all GMV is created equal and should not be your sole quest as a marketplace founder. Look to the other factors and see if you can be building more enduring value that’s separate from just the GMV number. As investors, we look carefully at what comprises the GMV and what the future may hold. You should too.