Vertical or Horizontal

People in the startup ecosystem have been heralding the “unbundling of Craigslist” graphic for years now.  This is the original from Andrew Parker.  The conventional wisdom is to say that horizontal marketplaces will be replaced by vertical marketplaces.  And why not – if you are designing a vertical marketplace you can have the user experience custom tailored to that vertical and provide a better experience.  In theory, that should win in the long run.  However, the reality is that many verticals will be better served by horizontal marketplaces for a long time to come.  

There are two factors that matter most when looking at whether a horizontal or vertical marketplace makes the most sense:  the size of the transaction and the frequency of the purchase.  

image

These two factors combine to make a standard matrix.  The upper left is the high price vertical and the lower right is the high frequency vertical.  My contention is that the lower left square is really better served by horizontal marketplaces rather than by verticals for everything that falls in that quadrant.

The upper right is a square with few use cases.  (Maybe this is a flawed framework?) There are some possibilities in this square – some B2B marketplaces, perhaps healthcare on the consumer side. If anybody is pursuing verticals with high frequency and high price transactions, please come talk with me.

Let’s focus on the two different types of vertical marketplaces:  either high frequency or high price.

High frequency verticals – The high frequency vertical is dominated by things like taxis and meals, e.g. Uber, Lyft, Sprig, Munchery.  The successful marketplaces in these categories will involve daily or weekly usage patterns from consumers and very high repeat usage. What matters most to consumers in the high frequency vertical is high levels of convenience, low cost, and reliable quality.  The supplier-picks model works extremely well here.  

High price verticals – The high price vertical is dominated by marketplaces for cars and homes, e.g. Zillow, Redfin, Beepi.  These are infrequent purchases with a very high ticket price.  Consumers in these verticals care about price, selection, and service more than turnaround time and convenience.  Would you really want to be able to just push a $30,000 button and have a new car delivered to your door in 15 minutes?

Now, let’s look at a different vertical.  Let’s take locksmiths. Does an “Uber for Locksmiths” make sense? I’m sure it would be a wonderful buyer experience to be able to push a button and have a locksmith show up in 5 minutes to fix a lock for $75.  So, if somebody built this, would it be successful?  I believe the answer is a resounding “no.”  The economics of customer acquisition and usage patterns just don’t work because a consumer may only use a locksmith once every couple years and only pays a small amount of money.  The marketplace would end up paying too much money to acquire a customer that doesn’t spend enough money on the platform so CAC > LTV (customer acquisition cost is greater than lifetime value).  The infrequent usage also means that organic growth is slow relative to higher frequency use cases.

The way to address this problem is to get horizontal usage.  Let’s imagine that same locksmith marketplace also offers plumbers, gardeners, housecleaners, and carpet cleaners.  Now a user can sign up to get a locksmith, but also use the service for every other home service.  Suddenly, the LTV potential of each user has dramatically increased because of the cross-sell opportunities and the acquisition equation can work well because acquisition cost is unchanged.  LTV > CAC.  The marketplaces in these low frequency / low price use cases must look for opportunities to bundle services into a horizontal marketplace.  

In addition, we have to remember the constraints on individual attention span.  Most people don’t want to have user accounts at 234 different sites.  Much better to just have a single Amazon account where you can buy pretty much anything or a single Thumbtack account where you can find just about any service.

So, if you’re thinking about a marketplace business, plot out the expected frequency and price to see where your use case falls.  Figure out a way to test recurring usage patterns to understand frequency.  And don’t assume that you should copy existing vertical or horizontal approaches.  The “uber for everything” trend is dangerous, it just doesn’t work for everything.  

blog comments powered by Disqus
  1. jbreinlinger posted this