*ELI5:LTV A (super simple) explanation of how to calculate customer lifetime value*

*ELI5:LTV A (super simple) explanation of how to calculate customer lifetime value*

June 20, 2017

As a VC who comes from an investing (rather than technical) background, I have been lucky to have Reddit's "Explain Like I'm Five" (ELI5) subreddit as a resource. On ELI5, redditors answer questions in simple, accessible language that requires as little prior knowledge as possible. It is perfect for anyone in the early stages of breaking down a complex topic they know little about.

I've realized in interactions with founders (and reading posts from fellow VCs around the internet) that there's a need for a similar resource for financial topics aimed at technical founders who have the same response to "DCF" that I did to "RISC." So I'm going to do my best to create one, starting with three little letters that are both important and remarkably complex to fully understand: LTV, or lifetime value.

Why do we care? Companies generally produce and sell products. A key determining factor in how big they can grow is how effectively they can sell their product. In other words, do they get more back from customers than it costs them to sell to those customers?

Imagine we're founders of a widget-distributor which buys widgets for 80 cents and sells them for $1 (generating 20 cents of gross profit). If an enterprising salesperson offers to help us sell widgets and charges 30 cents for every widget she sells, we’re unlikely to take her up on it. Yes, we will sell some widgets, but the process of selling will cost us more than what we make on each widget. No one stays in business long acting like that. If a salesperson could sell our widgets at a 10 cent cost per sale, that would be great: we'd make $1.00 - 80 cents - 10 cents = 10 cents/widget. Even at a 19 cent cost/widget sold, we would make more money paying the salesperson than we could without her (assuming we have no cheaper way to reach the same customers).

In the example above, the lifetime value of a customer is 20 cents: that is the marginal profitability from selling a widget. This is important information, as we've just seen, because it helps us figure out how much we can pay to sell a widget and still make money. If we can sell widgets for 5 cents reliably, we have a pretty big business on our hands. If it costs 25 cents to sell them, we're in trouble. Without knowing that 20 cent figure, we're flying blind and have no way to gauge which ways of selling the product are helpful and which are harmful, or how much to spend on the sales in the first place. The same basic lesson applies to every company, and that is why LTV is important.

Calculating LTV when you don't sell widgets: The above example is easy- we assumed that each customer buys once, and that each sale has a directly associated marketing cost. The real world is complex. Consider a subscription business model like Netflix, where users pay $100/year for access to a central pool of content. What is the LTV of any given customer? That depends on three main variables:

1) The gross margin on a new customer => As we discussed in the widget-example, the more a company makes per unit of revenue, the more it can afford to pay to buy a dollar of revenue. In this case, Netflix's gross margin is very high, (we’ll use 90%). This is because Netflix buys a piece of content and can then then let any user watch it. The associated costs are primarily storage space/servers.

2) The churn rate of customers => How long will a typical customer keep using Netflix and paying their subscription? Or, put another way, how many customers will leave (churn) each year? For simplicity's sake, let's say Netflix's churn is 10% per year => 10% of subscribers on January 1st will have canceled their subscriptions by December 31st. We will also assume that this is a constant rate: in any given year, any given customer has a 10% chance of churning.

3) The discount rate we use => This could be a whole different ELI5 post, but the key concept is that a dollar in a few years is worth less than a dollar today. How much less is a matter of intense study (and human psychology) but it relates to the length of time, other available investment opportunities and the risk inherent in investing. In start-up land I generally use 10-15%, depending on the level of risk in the business. Here, we'll say 10%.

With the first two items, we can break a customer down into a stream of expected gross profit (or, put another way, incremental cash flows) over time. Then we can use the discount rate to figure out what we would be willing to pay for each of those cash flows today. So using the example above, a cash flow 1 year away is worth 10% less to us than a cash flow today. A cash flow 10 years away is worth about 65% less, as we compound the discount rate just as interest compounds in a bank. Once we've figured out what each future cash flow from the customer is worth (also known as it the present value of that cash flow) we can sum up those present values to get a total, which is (drumroll...) the lifetime value of that customer. Here's what the first six years of that table look like:

One way to calculate the total present value would be to build the above table out for fifty years and sum them all up. Thankfully there is a simpler way: some math I won't replicate shows us that the sum of all of the present values is equal to year one's gross profit (63) divided by the [discount rate (10%) + the rate of annual churn (10%)], + the initial year’s gross profit (70). This gets us the right answer (385) without a huge table to sum. Here’s how that looks in equation notation:

You can download an excel sheet demonstrating his equivalence here (it can also be used to calculate LTV on your own). One note: this all assumes that cash is collected up front. That is why we don’t discount the initial year’s gross profit and just add the whole amount, because that is cash in the door today. Payment terms can make a meaningful impact on LTV!

One final wrinkle: Many companies want to include upsells in their LTV calculations, with the notion that after "landing" with an initial product, they will be able to "expand" within that customer at a lower cost of sales/marketing than before (in some cases with volume based pricing there may be no additional sales costs). This varies tremendously company to company, but the same basic logic we've used so far applies to upsells. My only recommendation is to calculate the value of potential upsells separately: it can help make very clear how important upsells are to the model and is easier to wrap one's head around than assuming a "growth rate" in customer spend which is harder to support factually.