Published in

--

## Calculating LTV/CAC for a two-sided marketplace

Some of the most iconic companies of recent history are two-sided marketplaces. Companies like Airbnb and Uber have reshaped how we live our daily lives, created entirely new categories and have given birth to the sharing and gig economies.

When assessing the strength and scalability of a two-sided marketplace, either as an investor or as an entrepreneur, there are a series of important KPIs to consider. One widely used, but often miscalculated metric, is the ratio of the lifetime value (LTV) of a customer to the cost of bringing that customer into the ecosystem in the first place, or customer acquisition cost (CAC). It answers the important question ‘how many dollars does this company generate for each dollar it spends acquiring customers?’ It’s also relevant when evaluating SaaS companies and ecommerce businesses, where it is relatively straightforward to calculate. There some nuances with two-sided marketplaces though which create additional complexity.

Unlike ecommerce businesses where the platform is also the supplier, two-sided marketplaces are intermediaries who need to source both buyers and sellers in order to complete a transaction, resulting in a separate cost to source each. For Airbnb, a seller is a homeowner, and a buyer is a renter. Sellers and buyers often have different customer lifespans on a marketplace, different frequencies of transacting, and usually there is a ‘many to one’ relationship where a single seller can supply multiple buyers.

Before a worked example, lets dig into the terminology to make sure we are all speaking the same language:

Lifetime Value (LTV) is the total value (in \$) an average user of the service generates for the company, over their active life on the platform. It can be calculated on either the seller or the buyer but shouldn’t be done for both — since the value generated on the platform is based on completed transactions with both a buyer and a seller. Because of the many to one relationship I prefer to calculate LTV for the seller as that is normally the unit of value (e.g. one car, one apartment) — but you can run with either. A caveat here is that this rule applies if the marketplace doesn’t earn non-transactional revenue from either party (e.g. car lease fees or insurance for an Uber driver) as these revenues sit outside of the traditional two-sided transaction.

When you’re calculating LTV, its best to start with raw cohort data. Cohorts are groups of buyers or sellers who joined the platform in a specific period, usually a given month. In any cohort you’ll find both buyers and sellers who transact once, some who transact occasionally and the super users who transact consistently. The important thing to get to grips with is what weighted average user behaviour looks like — from the perspective of average order value (AOV), transactive frequency, and the point at which users churn. Using all three metrics you can calculate the average lifetime GMV (Gross Monetary Value = value of transaction(s) on the platform, e.g. value of rentals on Airbnb) for either a buyer or a seller.

Once you have a lifetime GMV figure you can move to LTV. LTV should be calculated on a contribution margin basis net of variable costs. GMV is first multiplied by the marketplace’s take rate % to reach Net Revenue. Any COGS (incl. payment costs, insurance etc) should then be deducted, as well as any costs associated with maintaining customers (e.g. customer success agents, tech support). Those costs should be spread out across the full company transaction base and allocated proportionally.

Customer Acquisition Costs (CAC) are the cost to acquire and convert someone into a paying customer, both on the seller and the buyer side of the transaction (remember, there are two CACs for a marketplace). Two-sided marketplaces typically employ digital-first customer acquisition strategies using a combination of paid search / targeted social media marketing alongside brand awareness campaigns and more conventional offline marketing techniques.

Digital marketing cost can be calculated on a CPC (cost per click) basis divided by a conversion rate. Conversion rates reflect the number of people who start a conversion funnel process (by clicking on a link or engaging with a piece of content) vs those who go all the way through the process and become a paying customer. Ideally, the fully weighted CAC for both seller and buyer should also include a proportional allocation of all other variable S&M costs associated with converting new customers like staff, brand awareness campaigns and media (TV, Billboard etc).

Once you have an idea of both LTV and a CAC for your seller and your buyer you can move on to calculating your ratio. The final piece to consider is what time-period to use. If you have good historic cohort data, then run your numbers on the actual calculated seller (or buyer) lifetime. Many investors and entrepreneurs will also want to see what the one and three-year ratios look like to compare against other market-places on a like for like basis and as a proxy where the customer lifetime (especially seller lifetime) is uncertain. These proxies are useful when evaluating companies that either have very rapid growth and/or a limited operating history. As a rule of thumb, anything between 1–1.5x for a one-year ratio and >3x for a three-year ratio is considered positive.

Having obtained all the requisite data, the actual calculation is:

LTV:CAC = LTVs / (CACs + CACb * Marketplace Ratio)

Where the Marketplace Ratio is the number of buyers supplied by a single seller. Note that one buyer might purchase from multiple sellers — that does not alter the equation here. If you were to calculate LTV for the buyer, then you would divide your seller CAC by the marketplace ratio instead.

LTV:CAC = LTVb / (CACs / Marketplace Ratio + CACb)

A worked example is often the easiest way to bring together a concept like this. In this case, lets imagine a hypothetical business, MCFLY — a two-sided marketplace for hoverboards. Board owners (seller) make their hoverboards available for rent by the day to users (buyers) of the platform.

On average, boards are rented for one day at a time, for an average price (or AOV = average order value) of \$10. Over the course of a year, each hoverboard is rented 60 times, or five times a month. That accounts for boards rented just once, those which are rented consistently almost every day, and others somewhere in between — note, this includes all hoverboards on the platform, regardless of whether they are being rented or not. This is an important test for market liquidity and is crucial to factor in for an accurate view on LTV/CAC. The take-rate of MCFLY on each rental is 50% and the contribution margin (incl. all variable costs) is 65% of Net Revenue. Each hoverboard owner has a three-year expected life on the platform based on historic churn rates.

On the user side, the average user rents a hoverboard four times a year with a 1.5 year expected life on the platform, based on historic cohort activity. Our example is simplistic because we don’t assume any cohort decline. In the real world, a given cohort would rent proportionally more early on in their time on the platform — so we would expect to see higher monthly rental rates for a group of buyers in their first six months than if we looked across their entire 1.5 year expected life on the platform. That complicates the LTV/CAC calculation but is a subject for another article.

Back to our example. Users are acquired primarily through paid search and targeted social media marketing. It costs \$0.75 to acquire a lead and 20% of leads convert to paying customers, resulting in a digital acquisition cost of \$3.75 (\$0.75/20%). Sellers are generally more expensive to acquire because of significantly more targeted digital acquisition strategies. It costs \$2.00 to acquire a seller lead and they have much lower conversion rates of 5%, resulting in a digital acquisition cost of \$40. There is also a substantial marketing overhead from staff, TV commercials and offline flyering. These have been proportionally allocated at \$1 per converted buyer and \$15 per converted seller. The difference between the two is due to the marketplace ratio, you’ll see the calculation below.

Let’s work things through step by step, starting with a 1-Year LTV/CAC calc. First up, calculating the seller LTV:

AOV * 1 Year Avg. # Rents * Take Rate * Contribution Mg = Seller 1 Year LTV

\$10 * 60 * 50% * 65% = \$195

Next Seller CAC…

\$2 / 5% + \$15 = \$55

\$0.75 / 20% + \$1 = \$4.75

Our marketplace ratio here can be obtained with

Avg. # of Annual Rents per Seller / Annual Frequency of Rents per Buyer = Marketplace Ratio

60 Rents / 4 Rents = 15

So, bringing that all together using the LTV/CAC formula defined above gives us a one-year ratio of:

\$195 / (\$55 + \$4.75 * 15) = 1.54

You might find some examples of CAC calculations which proportionally allocate CAC based on the expected lifespan of buyer/seller on the platform. So, in this example, because our expected seller lifetime is three years, we would only include a seller CAC of \$18.33 (\$55 * 33.3%) and a buyer CAC of \$3.17 (\$4.75 * 66.7%). Although I see the logic, I disagree. LTV/CAC’s primary purpose is to reflect the actual unit economic profitability of a company’s operations over a given time period — it might be punitive on a one-year view in some cases, but it helps inform the cash requirements of the business you are assessing and gives a clearer picture of the businesses’ operating leverage.

To finish off, lets look at a three-year LTV/CAC ratio. Our seller lifetime is three years so in our simplified example we simply multiply the one-year LTV by three = \$585. Our buyer lifetime though is 1.5 years, so we need to account for the need to acquire a second cohort of buyers, which we do by adjusting the marketplace ratio. So, our LTV/CAC calculation becomes:

\$585 / (\$55 + \$4.75 * 15 * 3 Years / 1.5 Years) = 2.96

Some final thoughts to take-away in case you skipped to the end:

• Calculating LTV/CAC on a two-sided marketplace needs two CACs but only one LTV (which can be calculated on either the buyer or the seller),
• LTV should be calculated on a contribution margin basis
• CAC should include marketing overheads, as well as the digital acquisition cost
• CAC shouldn’t be allocated proportionally if you’re looking at shorter time-periods than your customer lifetime — you either pay it or you don’t
• Remember that lifespans are different for buyer and seller,
• ..and that sellers can be idle or unutilised — but should still be included!

Hope you enjoyed the post. Any comments welcome and remember to clap if you like and share if you love 😊

With thanks to Brian, Matt and Disi for your feedback and edits

--

--

Editor for

VC @ D2 Fund. Investing in the next generation of equity efficient founders