ArticlesBlog

Financial Projections: 2 Sided Marketplace

Financial Projections: 2 Sided Marketplace


Hey, Brandon from Poindexter here with
another exhilarating episode of how to create financial projections, this time
for a two sided marketplace. This example is a little different than the others
we’ve covered so far in that the marketplace model has more than one
target user base, which makes our task all the more challenging, so we’ll just
dive right in to illustrate how it works we’ll start by naming our project, which
we’ll call Airbnb for Uber, and it’ll be a marketplace where people can find
unused ubers to sleep. In in this example one side of the market is known as the
buyers, and these are the people looking for a place to sleep for the evening. On
the other side of the market we have our target sellers, and these are the owners
of the uber vehicles that are willing to rent them out to people that are looking
for a place to sleep. We’ll start by forecasting our sellers market which
will provide the inventory of Ubers available for users to book. One of the
first ways I’ll acquire sellers is by telling my driver about the app each
time I ride in an Uber. I’ll take a series of short trips just
to spread the word, and based on a budget of $100 a month, and an average
price of $5 per ride we can expect to generate 20 leads on average. Next we’ll
set the conversion rate for drivers that actually sign up for the app. I’m sure
most drivers are going to jump at the chance to let random strangers sleep in
their car, so we’ll set the conversion rate to 75 percent. The second way we’ll attract sellers will
be through paid advertising on Billboard’s along popular Uber routes,
where we’ll spend about ten thousand dollars a month and generate 50,000 leads. We expect 50% of these leads to follow up and visit our website, and then 60% of
them will continue onward and sign up for our app. We’ll be optimizing our
landing page for signups, so we’ll also assume our conversion rate will increase
by 5% per month over the next twelve months. Now that we have our sellers we,
can take a look at how we’ll attract the buyers. We’ve been building relationships
with travel bloggers over the past few months so we’re primed and ready to do
some guest blog posts targeted at people looking for place to stay in the LA area.
Out of a total of one guest post per month, over the next year we can expect
20,000 leads per article. About 10% or so of these leads should click through to
our site, and 50% of them will go on to download the app and become our
customers. Okay so we’re now ready to go on and set up our revenue models. The key thing to think about here is how either side of our two-sided market drives
revenue. For instance, if we’re charging a transaction fee, it’ll make sense to
forecast revenue based on the users that are actually making transactions.
Alternatively, if we’re charging users each time they list a new product we’ll
want to make sure we apply that revenue to the appropriate side of our
market. This is where being organized with our customer acquisition channels
will come in handy. Another consideration with the marketplace model is that
typically one side of the market is subsidized, so we can amass a large
number of either buyer’s or seller’s, whichever we need in greater quantity.
For us, we’ll be subsidizing the buyers, so it will be free for them to sign up
and use the app, but once they book a car to sleep in, we’ll take a cut of the
purchase price. The first thing we’ll do is give this
revenue stream a name, and check out our customer sources to make sure we don’t
include any sellers in this calculation. Next, there are two ways we can set up this revenue model. The first is charging a flat fee every time a user books a car.
In this case all we need to do is set the flat fee in the price field. The
second method is by charging a percentage of the overall transaction,
which is the method we’ll use here. Assuming an average rate of $100 a night, and
based on a transaction fee of 3%, we can expect three dollars per booking. If we
expect users to book multiple nights per month, because for instance we help them realize how cool it is sleeping in cars, we can then set that expectation in the
number of purchases field to account for that increased level of usage, so we’ll
set it at an average of three nights, or transactions per month. Lastly, since
users are no doubt going to love this experience, we’ll say that an average of 10% of our user base will return to the app each
month. This revenue stream is now ready to go, but if you expect different tiers
in purchase sizes, we’d want to set up an additional revenue stream for each level
of purchase that we expect users to make. For instance, the $100 per night rental
rate might be good to book a small sedan, but the rate for a large SUV could be
something like $200 per night, which would mean an average fee of six dollars
per transaction based on our three percent transaction fee, so it’s
important to consider the purchase size as a factor. Another method to forecast
revenue from transaction fees is by calculating the Gross Merchandise Volume (GMV). This method is more commonly used in e-commerce, but I want to cover it
because many marketplaces include the sale of products.
We’ll set up the revenue model the same way we just did, but the only difference
here is that instead of inputting our transaction fee in the price field, we’ll
input the entire purchase amount. After we click “Save,” the next thing we’ll need
to do is go over to the costs page. We’ll then select the percentage-of-sales
model, and input the sellers cut as a percentage of the revenue. Since we’re
collecting 3% of each transaction, that means the seller gets the other 97
percent, which is what we’ll want to input in the “expense as a percentage of
sales” field. Just make sure we’re applying the expense to the right revenue stream
and categorize this cost as a Cost of Revenue, and we’re all set. Now we’ll move on to the sellers. We’ll be charging five dollars per month to list their vehicles
in our system, so we’ll use the subscription model forecast this revenue. We’ll start by making sure the customer
channels are set up correctly, and that no buyers are included in this payment
option. And assume that 75% of potential sellers
that signed up will actually list their vehicle available for rent. Now we can just
set the price to $5 per month, and include some churn, because, and I know this is going to be very hard to believe, but the app won’t be ideal for everyone, so
we’ll set it to 10%. Then we just click Save and we’ve now successfully forecast
our revenue for both sides of the market. From here we can go on to include our expected costs to finish out our projections. Poindexter has seven
pre-built cost models to allow for maximum flexibility in setting up your
cost structure. We’re not going to go into the details of each cost model
right now, since it’s going to be different for each business, but whether
you want to set up your costs on a per transaction basis, as a percentage of
sales, per user, or even simple recurring or fixed costs, we have you covered. We’ll
create another video of the future that covers how to think about using each of
these cost models, so you can apply it to your own situation. I hope you enjoyed
this tutorial on forecasting revenue for marketplace business models. If you have
any suggestions for more videos, or questions on the one you just watched,
be sure to leave a comment below, and we’ll get back to you as quickly as we
can. That’ll do it for now, and as always, good luck planning the success of your
business.

Comments (1)

  1. Thanks for sharing these information but I wish you got more in depth in the topic than showing how fancy is your app!

Comment here