How to build a big marketplace

I’m not very good at keeping up with The Twitters, so I hopped into this conversation about building large marketplaces super late.

It’s an interesting one — when building a marketplace, which comes first?  The chicken or the egg — supply or demand?

chicken_or_egg

Image credit: Rushart’s blog

Obviously both are important for a successful marketplace!  But, if you want to build a really BIG marketplace, here are some observations from over the years.

1) Unlocking large amounts of supply matters A LOT!

This is a bit unintuitive.  Most people who set out to build a marketplace think that if you can get people to pay for something (the demand side), then you’re all set.  I’ll argue that it’s certainly important to test the demand side but it’s almost more important to test how hard it is to get supply.

From my experience in growing an ad network, it is possible to have lots of demand but not enough supply!  This is actually a very common phenomenon.  Case in point, most email newsletter companies have an easy time selling out their ad slots, but it’s incredibly hard to continue growing an email newsletter at a fast clip.  The way that a lot of email newsletter companies solve for this problem is by introducing new lists with the same audience.  E.g. you can receive the daily news digest and also the daily jobs email.  This gives them twice as much supply with the same audience.

2) If you are aggregating supply, the key is to unlock new supply

One of the big areas where I see marketplaces fail is by going after an existing market and trying to amass the same supply that already exists.  So for example, a marketplace for salons or a marketplace for wedding venues or a marketplace for co-working spaces.  These marketplaces are all amassing existing salons or existing wedding venues or existing co-working spaces.  These are existing places that consumers could ordinarily find themselves and pay for directly.  You are literally just moving supply around and not growing it.  The issue with doing this is that this existing supply already has certain expectations for payment, because they are already making money for this service or asset that they provide.  This then makes it hard to be a middle(wo)man and take a cut in between.  You are competing with a strong alternative — to be found directly.

The better way to aggregate supply is unlock new unique supply.  Airbnb is a great example of this.  People were not already using their extra bedroom as a hotel room or their couch as a bed.  They don’t have the same expectations around making a ton of money unlike the Holiday Inn.  Airbnb has effectively brought a ton of new “hotel room supply” to the market that didn’t exist before.  They were not try to resell existing rooms in existing hotels.  Uber and Lyft are equally good examples of doing this in the taxi market.  They brought into the taxi market new “cab supply” that didn’t exist before, and these drivers don’t have the same expectations for monetization as existing taxi drivers.

giphy-1

Image credit: Giphy

Ultimately, unlocking new supply drives demand.  If I can stay on someone’s couch next to the Moscone to attend a conference for 50% of what I’d pay for a room at the Holiday Inn, I’d do it.  You are reducing prices for the end user by unlocking new supply, and this drives demand.

So going back to the original examples of marketplaces for salons or wedding venues, etc, can you get clever / creative in creating new supply?  Can you turn new people who are not in the salon business into a salon owner?  In many cases, doing this might just be too high of a cost and not possible, but in some cases, this approach may be a good strategy.  A good example of this is Wonderschool.  Wonderschool is turning people into new daycare owners — they are not amassing a network of existing daycares but rather unlocking and creating new ones to add to their network.  So think about unlocking new supply rather than moving existing supply around.

3) The unit economics need to work in the long run at scale

Once you initially test both supply and demand, there’s going to be a constant tension between both sides.  Sometimes you’ll be supply constrained.  Sometimes demand constrained.  Often, it may not be clear if the unit economics will work out while you’re building this up.

In fact, ridesharing companies often get a lot of flack from the public, because they are not profitable yet.  But, the holy grail for them is autonomous cars.  Once these become mainstream, they will have access to infinite supply at a low cost.  So while the short-term numbers may be questionable, the long-term future of these companies seems very promising.

Similarly, you’ll need to think about what your long term control over supply will be.  Most marketplaces that are successful have a stronghold on at least a good portion of their supply to help with pricing pressures.  Successful ad networks are a great example of this — Google may run a large ad network across many properties they don’t own, but they also own a lot of their own properties including Google search and YouTube.  Likewise, although Airbnb doesn’t own properties today, it’s rumored they are going into real estate.  So once you get some footing on your marketplace, the next question is how can you think about controlling your future by having access to or creating at least a good portion of your supply?

4) What should I look for in amassing unique supply? 

If I were to build a large marketplace today by amassing supply, I would start by looking around at what is currently wasted (space / time / assets).  Then, I’d think about how this wasted stuff might be cleverly transformed into something else that consumers and businesses currently spend a lot of money for.

Summarizing all of this, to make a marketplace fly, you need to cleverly come up with a LOT of unique supply (obv there has to be demand).  1) Turn something else into supply where people don’t have high monetization expectations (Airbnb, Uber, Lyft).  And/or 2) eventually you own it or part of it (e.g. scooters / Google search).

An analysis of going from $0 to $1m revenue in 1 month

At Hustle Fund, we talk a lot about speed of execution. Part of this is grit and founder hustle. But what most people don’t realize is that speed is also baked into the business model itself.

For example, one of the things that people don’t realize is how much sales cycles and payback periods matter.  As a concrete example, fast growth direct-to-consumer companies are now able to go from $0 to $1m revenue within a month or two with just < $25k in ad spend. This is incredibly fast — unprecedented and very low cost.

How is this possible? All because of fast sales cycles. Let’s dive in:

Let’s start with a simple example. Suppose the following:

  • We start with $25k in ad spend on day 1
  • As soon as people click on our ad, they buy immediately and we get that revenue immediately
  • For every $1 we spend on our ads, we make an average $1.15 in revenue
  • We pour everything back into ads the next day
  • You can see that within a month, we are at an almost $2m revenue generated with just initial ad spend of $25k! This is crazy. You can raise a small pre-seed tranche and get to Series B benchmarks within a month.

Screen Shot 2019-02-19 at 3.05.37 PM

Now, in practice, the scenario I’ve outlined is near impossible. Namely, most of the time, your ad spend is not this effective on day 1 and often requires iterations and testing. Moreover, it often takes some time to clear with your payments provider even if you generate revenue right after someone clicks your ad.

But this scenario is actually not that far off from what many of the direct-to-consumer incubators are achieving.

Ok, let’s add some delay to make this a bit more realistic. Let’s suppose:

  • We start with $25k in ad spend on day 1
  • As soon as people click on our ad, we do get our initial ad spend back immediately but we don’t make any profit until the even numbered days
  • For every $1 we spend on our ads, we average $1.15 in revenue (on the even days)
  • We pour everything back into ads each day

Screen Shot 2019-02-19 at 3.07.06 PM

This is a contrived example (profit only on even numbered days??  weird.), but you can see in this example, we went from a near $2m revenue generated at the end of the month to just a $200k revenue generated, even though the delay in our profit is not that long. Not only that, we are making about half the profit as before, but our revenue generated has now dropped by nearly 10x. Why? Because time is a compounding factor.

So when investors are concerned about sales cycles and payback times, this is what they mean. Regardless of whether you are using ads to get customers, if you calculate this out, you can see how just small affects in sales cycles / payback periods hurt you big time. The other thing is that as an entrepreneur, you’re working just as hard in the first scenario as in this one, but your sales are affected 10x.

Both of these examples are still a bit contrived, because as I mentioned above, you usually don’t know if your ads — or any of your customer acquisition channels – are working (they probably aren’t) on day 1.

So now let’s assume that you don’t start with $25k in ad spend.  Let’s start with $5k. Suppose the following:

  • We start with $5k in ad spend
  • As soon as people click on our ad, they buy immediately and we get that revenue immediately
  • For every $1 we spend on our ads, we make $1.15 in revenue
  • We pour everything back into ads the next day
  • We add an additional $5k in ad spend every 5th day

Screen Shot 2019-02-19 at 3.08.12 PM

You can see that with this example, we can get to near $1m revenue generated by the end of the month ($800k+). In this case, we end up spending a bit more on ads, but this is a more gradual (and realistic) approach to ramping up ad spend. Again, in this example, because we make the $1.15 for every $1 immediately, you can see that this helps us achieve greater revenue than in the prior example.

Here is the spreadsheet with these 3 examples that you can copy and play with.  I’d encourage you to fiddle with the levers of this model in the green boxes. The amount you make for every dollar of spend. The payback periods. Ramping up ad spend. Etc.

In your own business, how can you decrease your sales cycle + payback period by even just 10%? How can you increase your margins by even just 5%? These all add up significantly as you can see over short periods of time.

The tl;dr is never underestimate the value of time in making money.

Featured image credit: Freeimages.com