top of page
  • Writer's pictureRyan

Unit Economics

One of my favorite memos by Howard Marks is called Economic Reality. As the title suggests, he talks about how economics dictate reality, whether we like it or not.

He uses a story to illustrate this point…

“In 1977, responding to the difficult energy outlook brought on by the Arab Oil Embargo, President Jimmy Carter created the position of Secretary of Energy and chose James Schlesinger as America’s first “energy czar.”
…In his early days he taught economics at the University of Virginia…[And] as the story went, Schlesinger was such a convincing evangelist for capitalism that two students in his economics class decided to go into business after graduation. Their plan was to borrow money from a bank, buy a truck, and use it to pick up firewood purchased in the Virginia countryside, which they would then sell to the grandees in Georgetown. Schlesinger wholeheartedly endorsed their entrepreneurial leanings, and they proceeded with great enthusiasm. From the start of their venture, the former students could barely keep up with the demand. Thus it came as quite a shock when their banker called to tell them the balance in their account had reached zero and the truck was about to be repossessed.
They contacted Schlesinger, and he listened attentively as they recounted their experience: they had, in fact, been able to acquire vast amounts of wood for $50 a cord, and they’d been able to sell all they had for $40 a cord. How could they be broke? Where had they gone wrong? Schlesinger puffed on his ever-present pipe and said, “The answer’s obvious: you need a bigger truck.”

This is a great story because it reveals how we can’t get around the laws of economics. Buying something for $50 and selling it for $40 will never work. You can’t make up losses with volume. Reality doesn’t work that way.

In this vein, let’s talk about unit economics.


So what are: unit economics?

Well, unsurprisingly, it’s economics at the unit level (obviously I never learned that lesson where you can’t define a word with the word itself).

For instance, let’s say you're an executive at Chipotle.

The unit is the restaurant. And the economics dictate how long you will stay in business.

In short, how much does it cost you to build and run each restaurant and how much cash do you make from them?

This is unit economics.

Or take software as an example.

The unit is the customer. And the key economic questions are: how much does it cost to acquire a customer and how much value is the customer worth to the business?

[Side note: The industry terminology for software is CAC (customer acquisition cost) and LTV (life-time value).]

A lot of people shy away from software companies because it looks like they are losing money. But the real question is: how are the unit economics trending?

Here is a visual representation of what happens:

Breakeven Point

Initially, software companies spend to acquire customers. But then as these customers spend more money over time, the company becomes profitable.

However, if the software company isn’t diligent on the unit economics, it will, as in the wood cutting example, go out of business.

As a reminder, CAC stands for customer acquisition cost and LTV represents lifetime value.

Let’s make these acronyms come to life shall we?

In a recurring revenue business like software, these metrics are simpler to understand so we’ll stick with that for now.

First, we need to figure out the CAC, which is derived from sales and marketing expenses. While most people make this formula seem like an exact science, it isn’t.

There are many questions to address. For instance:

- Should we include salaries? (Fully loaded vs. non-loaded).

- Should we include rent costs allocated to the number of sales and marketing employees?

- Should we include software tools used?

- Should we include support if it helps win customers?

There are all sorts of questions that can vastly change the calculus. But the bottom line to calculate CAC is:

The total cost typically consists of sales and marketing expenses. But as we said, sometimes deciding what goes into that column can be tricky.

But let’s say it costs you $10,000 in one year to acquire 50 new paying customers.

Therefore, your CAC is $200/year. This is the baseline.

Now we need to figure out how valuable this customer is via LTV, which is made up of 4 parts.

1. Annual recurring revenue per customer

2. Gross margins

3. Churn

4. Discount rate

We first want to see how much gross profit each customer makes for us annually.

[Gross profit is just sales with the cost of goods sold taken out. For example, if we were making shoes, the materials of the shoes would be the cost of goods sold.]

In our LTV calculation, we use gross profit because it is a good measure of how efficiently a business can serve a customer. Put another way, you wouldn’t value a customer that makes you $90 per $100 of sales as you would a customer that brings in $50 for the same transaction.

Then we have churn. The best way to calculate this is in terms of gross profit as well. Some people only calculate the customer churn rate but we think that isn't the best idea because different customers can have vastly different values. Dollar values matter more than customer count.

Now let’s dive into the fourth thing you need to know before putting some numbers around this idea to give it some context: the discount rate.

This discounts the lifetime value of the customer to the present so we can compare it against the customer acquisition cost which is money that has already been spent. Not all LTV/CAC calculations take this into account but it is important to always measure against an opportunity cost.

Just think of a discount rate as the possible return you could get, had you not used the money for this business.

For illustrative purposes, we’ll use 6% because that is a reasonable cost of capital in today's low-interest-rate environment.

So let’s put it all together.

Phew, that’s a mouthful and probably still confusing.

So we'll use some numbers for context:

So we can see, for the year, the CAC was $200 per customer. But the LTV gets a little more complicated.

The goal is to figure out how valuable a customer is right?

First, we know the value is on a per-customer basis because this is unit economics after all.

Second, we established that gross profit makes more sense because it gives a more holistic picture.

Third, we need to know churn because customers come and go. We must have an idea about how long a customer stays. To figure this out, you can divide 1 by the churn rate.

Fourth, the discount rate helps us keep in mind our opportunity costs.

So this is what we're left with:

This means that the lifetime value for a customer is 3.6x the cost of acquiring one. That’s awesome!

You want that number to be above 1 but not too high or else that means you’d be losing out on potential revenue.

For instance, if it was below 1, that would mean the cost of acquiring is higher than the lifetime value. Obviously that’s not good, although some companies need to get a network effect going (Ex. OpenTable used to pay restaurants to use it before gaining mass market appeal).

On the other hand, if the LTV:CAC ratio was above, say 5, your customers are so valuable that maybe you want to spend more so you can acquire even more of them to grow the business.

An ideal LTV:CAC ratio is somewhere between 3 and 5. Lower than 3 and you probably have too much churn. Higher than 5 and you probably aren't spending enough to acquire customers.


While CAC and LTV are not an exact science, they are crucial for software companies. More broadly, these concepts are crucial to understanding unit economics.

Business operators monitor these metrics very closely but as investors, we typically view them as too hard to understand or not important enough. While it's true that most companies don't always provide transparent metrics, taking the time to understand the unit economics of a business is crucial.


Want more? Subscribe to our free weekly newsletter. It goes out every Friday morning with a ton of free investing and business-focused content.

Recent Posts

See All

Discontinuous Disruption

Discontinuous Disruption The year was 2000, the beginning of the tech bubble descent. Still groggy from waking up at 4 am, three men boarded a private plane at the Santa Barbara Airport. Little did th


bottom of page