How to Calculate Unit Economics for Platform Businesses
A platform connects individuals or businesses looking to provide a product or service (supply) with consumers or businesses looking to buy such a product or service (demand) and usually takes a percentage commission (take rate or “rake”) on the gross transaction value (GTV). There are various well-known examples of two-sided digital platforms matching supply with demand and enabling strangers to transact successfully, including: Airbnb in the housing market, Uber and Lyft in the mobility market and Shiftgig in the temporary staffing market.
Getting traction on both sides of the platform and scaling a platform is not easy. But if a platform can be scaled successfully, the underlying asset-light business model promises significant long-term profits, a fact that attracts entrepreneurs and venture capital investors alike. When pitching to investors, entrepreneurs need to be able to demonstrate various KPIs in a compelling venture capital pitch deck. KPIs include GTV, GTV growth rate, take rate, gross margin, contribution margin, monthly cohorts, engagement as well as (signs of) viral growth and strong network effects.
While there seems to be a common understanding on how to calculate most KPIs, we continuously see some inconsistent approaches as to calculating unit economics. We have seen pitch decks in which founders present their platform unit economics separately for both the supply and the demand side. Another approach adds (a percentage of) the costs incurred for acquiring the non-paying side of the platform into the costs incurred for the paying side of the platform and essentially calculates a two-sided CAC. Others include the costs incurred for acquiring the non-paying side of the platform into the contribution margin calculation component of the CLV calculation for the paying side of the platform. This heterogeneity as regards calculating customer acquisition costs (CAC) and customer lifetime value (CLV) leads to confusion and an unnecessary amount of time spent discussing what the correct approach is.
With this article, we want to share with founders some of our insights we gathered over the years and explain how we believe unit economics for a platform business need to be calculated.
Founders may find this article helpful also from a business performance improvement perspective. Improving the variables underlying the unit economics calculation will also improve the viability of the business model and the long-term success chances.
Root cause for inconsistent approaches as to calculating unit economics
The root cause for the inconsistencies may be that platforms often collect revenues from only one side of the platform but incur acquisition costs on both the demand and the supply side of the platform. For example, temporary staffing platforms like Shiftgig may incur costs for acquiring workers (supply) and businesses booking such workers (demand) but only take a commission on a shift done by a worker. Catering platforms like EZCater may incur costs to acquire caterers (supply) and customers (demand), but only charge a commission on the catering volume delivered. Founders sometimes seem to struggle to correctly allocate the costs associated with acquiring the non-paying side of the platform.
Calculating unit economics for both sides of the platform
We have seen pitch decks in which founders present their platform unit economics separately for both the supply and the demand side. This would for example mean that temp staff platforms on the one side show the costs incurred for acquiring a worker (worker acquisition costs) in relation to the value generated by an average worker (worker lifetime value). On the other side, they would show the costs incurred to acquire businesses booking workers (business acquisition costs) in relation to the value generated by an average business (business lifetime value). Here is an example of how CAC and CLV are sometimes presented for Shiftgig-like staffing platforms.
The charts show positive unit economics and indicate a viable business model. The values generated by workers and businesses by far exceed the costs incurred for acquiring workers and businesses, respectively. This approach however neglects the fact that the revenues generated by the worker side are the same revenues that are generated by the business side of the platform (since the platform charges only one fee for a shift worked). Hence, this approach double-counts the commissions charged by the platform and results in a misleading picture of the platform’s unit economics. In fact, this approach can lead to an extreme situation in which the unit economics look positive, while the contribution margin on a company level is negative and hence indicates that the platform loses money on each transaction.
Calculating unit economics with a two-sided CAC
Another approach adds (a percentage of) the costs incurred for acquiring the non-paying side of the platform into the costs incurred for the paying side of the platform and essentially calculates a two-sided CAC, which then can be put in relation to the CLV of the paying-side of the platform. The two-sided CAC thus takes into account the costs for acquiring both supply and demand. To account accurately for the relation between supply and demand, this approach considers the number of sellers needed per buyer.
The key formulas:
Seller to Buyer relation (SBR) = # of sellers / # of Buyers
Two-sided CAC = Buyer CAC + SBR * Seller CAC
While this approach avoids double-counting revenues, it has not only a terminological drawback. Unit economics are meant to show the relation between the costs of acquiring an average customer and the lifetime value of an average customer. But costs incurred to acquire non-paying supply has in fact not much to do with costs incurred for acquiring paying demand. To return to the temporary staff platform example: The costs incurred for acquiring a worker are not costs incurred for acquiring the paying business that books the worker. Rather, the costs incurred to acquire a worker relate to the “product” transacted on the platform. Without acquiring workers there would not be any workers that could be booked by paying businesses. We therefore consider these costs direct costs that need to be considered when calculating the platform’s contribution margin, which, in turn, goes into the CLV calculation for the paying-side of the platform. As also Andrei Brasoveanu, venture partner at Accel, put it: “Don’t forget to include seller acquisition costs in your contribution margin! You need both sides of the marketplace to scale.”
The CLV formula
As the costs for acquiring the non-paying side of the platform may rather be considered product-related and hence direct costs, we approach the unit economics calculation by including the costs incurred for acquiring the non-paying side of the platform into the contribution margin (CM) component of the CLV calculation for the paying side of the platform. We can then put this CLV in relation to the traditionally calculated paying-side CAC to arrive at a meaningful CLV/CAC ratio.
We calculate CLV by analysing the value generated by an average customer in the first active month (first month value (FMV)) and adding the value generated by an average customer after the first month (terminal value (TV)). Similar calculations could be done on a transaction basis (rather than on a monthly basis). This leads to the following formulas:
CLV = First month average customer value (FMV) + Terminal average customer value (TV)
The first month average customer value can be derived by multiplying the first month average booking value (FMABV) (or the gross transaction value) with the take rate (TR) (in order to get to net revenues or commission) and the contribution margin (CM) (in order to get to the value being generated after direct costs).
FMV = First month average booking value (FMABV) * Take Rate (TR) * Contribution Margin (CM)
The terminal average customer value (TV) reflects how long an average customer returns after month 0 in months (n) and how much value an average customer generates during this retention time. The terminal value (TV) can be calculated by multiplying the first month average customer value (FMV) with the average booking value retention rate (AR) and the retention in months (n).
TV = FMV * Average booking value retention rate (AR) * retention in months (n)
This leads to the following CLV formula:
CLV = FMABV * TR * CM + FMABV * TR * CM * AR * n
Calculating first month average customer booking value (FMABV), average booking value retention rate (AR) and retention period in months (n)
We suggest calculating first month average customer booking value (FMABV), average booking value retention rate (AR) and retention period (n) on the basis of a respective cohort analysis. Cohorts demonstrate the monthly customer activity in a specific period of time. The typical metrics to run cohorts on are gross booking value, gross margin, contribution margin, or number of transactions. A cohort analysis will reveal the average first month booking value of an average customer (FMABV) and the value an average customer generates in the following months (AR). At the same time, especially when looking at early stage companies, the cohorts often do not clearly show the full period of time an average customer remains active or returns, respectively, after the first month (or his first transaction) (n). In this case, respective assumptions need to be made.
We assume that founders of platform businesses are — and should be — familiar with cohort analysis, so that we refrain from getting into more details. However, an introduction to cohort analysis can be found here.
Calculating take rate (TR)
The take rate is the average commission taken by the platform for a transaction, usually between 10% and 30%. The actual take rate can be calculated by dividing the commission effectively paid to the platform for an average transaction by the gross booking value of an average transaction or by dividing net revenues by gross booking value (post-cancelations & returns).
Calculating contribution margin (CM)
The contribution margin can be calculated as follows: The costs of goods sold (CoGS) need to be subtracting from net revenues in order to get to gross profit. Now, all other direct costs — including the costs incurred for the non-paying side of the platform — need to be subtracted from gross profit in order to get to the relevant contribution margin. The following costs are sometimes disregarded, but constitute CoGS (and in any case direct costs):
- costs related to keeping the platform environment functional (e.g. hosting costs)
- costs related to matchmaking, i.e. connecting supply and demand and hence enabling a transaction (often called operations)
- costs related to onboarding and retaining supply and demand (onboarding costs, activation, customer support), but excluding acquisition, up- and cross-selling costs (which would go into Sales & Marketing and hence CAC).
CLV Calculation Example
If we assume that a fictional temp staff platform shows a first month average customer booking value (FMABV) of 1,000€, a 20% take rate (TR), a 40% contribution margin (CM), an average booking value retention rate (AR) of 40% as well as a retention period (n) of 47 months (months after month 0), the variables can be plugged into the formula for calculating CLV:
CLV = FMABV * TR * CM + FMABV * TR * CM * AR * n
CLV = 1,000€ * 20% * 40% + 1,000€ * 20% * 40% * 40% *47
CLV = 1,584€
In order to calculate the customer acquisition costs, the sum of all marketing and sales costs — but without the costs incurred for acquiring the non-paying side of the platform — need to be divided by the number of new paying customers acquired (fully-loaded CAC). Assuming that the fictional temp staff platform spends 40,000€ per month on acquiring on average 100 paying businesses, the fully-loaded customer acquisition costs (CAC) amount to:
CAC = 40,000€/100 = 400€
In our fictional temp staff platform example, the unit economics would show a 3.96 CLV over CAC ratio.
CLV/CAC = 1,584 / 400 = 3.96
Our CLV formula is pretty detailed and similar results may sometimes be derived by using less detailed formulas, especially when analysing platforms with high retention rates (e.g. formulas that do not differentiate between first month value and terminal value but work only with lifetime averages: Average Booking Value * Take Rate * Contribution Margin * Average Retention Rate * Lifetime in months). But being precise has never been a disadvantage.
Further, especially at early stage companies with little automation and little traction but with many bodies at work, the contribution margin may be negative so that the CLV will be negative too.
It is therefore vital to think carefully about unit economics on a case-by-case basis and only as a small component of a more comprehensive business assessment. All KPIs should eventually demonstrate that the business model is viable or will be viable in the future. If the contribution margin is still negative, founders should be able to demonstrate that the contribution margin is negative for good reasons and will turn positive.
Founders may also find it helpful to look at the above unit economics calculation and the underlying variables from a business performance improvement perspective. Certainly, often easier said than done but the platform business performance can generally be improved by:
- Increasing the first month average customer booking value (FMABV)
- Increasing the take rate (TR)
- Improving the contribution margin (CM) by reducing CoGS and other direct costs (often by reducing touchpoints per transaction) and reducing costs incurred for acquiring the non-paying side of the platform
- Improving the average booking value retention rate (AR)
- Extending the retention period (n)
- Reducing costs to acquire paying customers.
If you have established a digital platform business in the hospitality, retail or food industries and if your business shows significant traction and growth, please reach out to us.
We are always happy to engage and help!
- Fabrice Grinda of FJ Labs on the importance of marketplace unit economics
- Former Airbnb product manager Jonathan Golden on questions every marketplace start-up should be able to answer
- Andrei Brasoveanu, venture partner at Accel on 10 marketplace KPIs
- Bill Gurley, partner at Benchmark, on the 10 factors he considers when evaluating digital marketplaces
- VersionOnes comprehensive Guide to Marketplaces
- The Speedinvest Marketplace Scorecard
- For an in-depth analysis of some of unit based KPIs of the Uber and Lyft see this interesting piece by VC investor Benjamin Tseng
- A16z on network effects
- Andrew Chen: Required reading for marketplace startups: The 20 best essays
- Mike Lloyd: Unit Economics: The Vital Signs of Online Marketplaces
- Samaipata’s 3-step/5' cohort analysis (Part 1): First steps to a cohort analysis
- Samaipata’s 3-step/5' cohort analysis (Part 2): Run a full stack cohort analysis from scratch in just 5 mins and leave your tech team to focus on the hard stuff
If you are generally interested in how to achieve high growth and predictable revenues, please subscribe to my blog. If you are a founder, I would be delighted if you joined my LinkedIn group Smartscaling.