Guide to Best in Class Unit Economics
Venture capital backed technology businesses lean heavily on unit economics to report on the financial profile of the company. They are a powerful storytelling device used by founders to convert unprofitable income statements into long term financially viable business models.
Unit economics are a projection of the free cash flow generation of cohorts of customers. Backwards looking profit and loss statements are inherently slow to demonstrate value accruing to shareholders.
The primary components of unit economics are lifetime value of a customer (LTV) and customer acquisition cost (CAC). The definitions of unit economics are broadly understood (albeit by a niche audience within Finance and Tech) and documented by thought leaders like David Skok; whose site For Entrepreneurs blog I highly recommend.
A CFO once emphasised to me - unit economics are miles away from accounting standards. While the logic of unit economics requires a bit more work outside of the general ledger, there are better (and worse) ways of preparing unit economic reporting. Here I cover what I consider to be best in class unit economics, so you know where you sit on the conservative / aggressive spectrum.
Topics Covered
LTV/CAC theory
Cohort selection; CAC and LTV
Gross margin analysis
Cost to serve v. CAC
Arithmetic of fractions
Time value of money
LTV / CAC Theory
A quick recap on definitions, including an exploration of what financials should be included.
Customer Acquisition Cost
CAC is a measure for the cost associated with acquiring a cohort of customers. If you spend $1M acquiring 25,000 individuals in a given month, your CAC will be $40 per user.
CAC I break up into three distinct components, in line with the customer journey. Marketing (awareness, lead generation), Sales (conversion of prospects) and Onboarding (getting the customer using your product).
If your go to market strategy is primarily sales force led, the vast majority of your CAC should be Sales, paying healthy on-target salaries to successful sales staff, with Marketing and Onboarding positioned to support Sales.
If you have a direct to consumer, or online checkout-journey your CAC should focus on Marketing, as Sales and Onboarding functions should be automated.
Marketing CAC (what’s included)
Advertising costs (online and offline; everything!)
Salaries for your marketing team
Overheads associated with your Marketing team (office overheads; you can use a proxy such as average cost and just multiply by headcount for your reporting period)
All brand activities – the long-term value accretion of companies building a brand does not justify any deferral. Coca Cola could stop all marketing in a year and sales would not drop, but if they did that for ten years sales would dive.
Any other costs your Marketing team need to do their job (e.g. Salesforce Marketing Cloud)
Sales CAC
All compensation paid to your sales force
Systems used by sales (e.g. CRM software suite)
Onboarding CAC
All compensation paid to your onboarding / customer success / retention teams
Systems used here might overlap with Sales and Marketing (make allocations as you see fit)
Onboarding is a useful category for anything where you have a real tangible widget associated with your product that the customer needs. Onboarding costs are likely harder to influence and may be reliant on partners (for instance a fintech shipping out a physical card to customers). They are likely less significant and so by parsing out, you can focus on the components of your CAC that matter.
Lifetime Value of Customer
LTV measures the marginal profitability of a customer over their (expected) lifetime. Marginal profitability = the gross profit contribution of a customer. Gross profit is the revenue collected from a customer less all the expenses associated with servicing said customer. Apple makes a 30% gross profit margin (link) – they charge $1,000 for a shiny iPhone and after manufacturing, shipping, provisions for faulty devices they make $300.
Lifetime value of customer =
Average Gross Profit per Customer x Average Customer Tenure
Where:
Average Customer Tenure =
Conventional methodology takes the reciprocal of customer churn.
Customer churn is the % of customers that left in one reporting period, mostly cited on an annual basis. If 20% of customers leave each year, then customer tenure is 1 / 20% = 5 years. You can be more precise with customer lifetime if churn is not stable; most consumer products (particularly freemium models) are not stable and don’t have stable churn rates – whereas B2B software is relatively stable.
Average gross profit per customer:
Average revenue per user (ARPU) * gross margin.
Average revenue per user can be segmented into cohorts with dimensional overlays where relevant. Growing ARPU is impressive, hard to achieve and indicates customer stickiness.
Putting the components together:
Customer churn = 20%
Customer tenure = 5 years
$250 ARPU
60% gross margin
Lifetime value of customer: 1 / 20% * $250 * 60% = $750
LTV:CAC and Payback Period
LTV : CAC is the reciprocal of LTV over CAC. If you are below 1 you are losing money each time you acquire a customer. If you are above 3 you probably have some kind of viable business model. If you are above 5 you have a great business model. Any higher than 5 and you probably should be investing more into growth.
The other perspective is payback period. How long it takes you to repay your CAC.
VCs are impatient, so the shorter the better. Anything less than 12 months is great, however between 12 and 24 months can still be a compelling story if the LTV:CAC is there.
Cohort Selection (LTV; Cost to Serve)
In forecasting customer LTV you can select cohorts in two ways, that should be guided by your go to market strategy; the options I’ll phrase as the pool method and the sign up method.
Signup Method
I wanted to call this the cohort method, but I thought that might be too confusing. By signup method, you can break up the LTV based on when you sign customers up and make projections on LTV for each cohort.
Pool Method
You pool all existing customers into a group to determine the average LTV.
Under both signup and pool method you are then dividing against CAC.
Signup (cohort) v. Pool (all customers)
The difference between the two methods is whether you want LTV to reflect the value of your new customers or all customers.
Here are some directional clues on which works best:
Longer sales cycle —> pool
High variability in marketing activities —> signup / cohort
B2B —> pool
B2C —> signup / cohort
High variability in ARPU —> signup / cohort
80% or able net-dollar retention —> pool
You have a non-recurring product —> signup/cohort
Churn is low —> pool
Churn is high —> signup / cohort
Cohort Selection (CAC)
CAC is relatively easier in determining your denominator, new customers. Monthly is conventional and will cover most business models. There is an argument to conduct trailing quarterly new customer figures. The longer the sales cycle the more persuasive is the argument to spread new users over a longer period of time (trailing three months). Anything longer than three months approaches capitalisation of expenses – which is the antithesis of unit economics.
An alternative approach when the sales cycle is very long is to report an estimated new user figure. A use case I’ve seen for this is to forecast new users based on historical sales rep productivity. This was done because of a long sales cycle coupled with a long sales rep onboarding cycle. It was an internal management metric and was useful to cut through the math of spending more on a sales force that is on-target, but the customers haven’t gone live yet.
What Goes Into Gross Margin
Gross margin should reflect the cost to serve customers and nothing else.
Typical constituent parts:
Payment processing fees (e.g. Stripe)
Provision for bad debts
Hosting costs
Customer success / retention teams (including overhead allocations)
Account managers
Software used directly to service the customer
Allocations between CAC and Cost of Sales
There will be some allocations your financial reporting team will likely conduct between costs that become cost of sales or a component of CAC.
For simplicity staff who have blended roles, between account management and sales, if the mix of one (versus the other) is less than one quarter, it is likely not worthwhile to conduct an allocation. Unless the expense on absolute $ terms is significant.
Allocations in general; the less the more. It makes your reporting cleaner and avoids any notion you are trying to manage your figures.
Hosting Costs
Hosting costs often justify allocations between cost of sales and R&D. You might encounter a difference in reporting as you as a legal entity record as many costs as possible in the R&D bucket for tax purposes but then from a reporting perspective present a more balanced view.
It will be quickest to start from reviewing R&D spend that is purely R&D related and leave the reminder as cost of sales. This could be a periodic review, so you have a set % allocation policy for the year. You should avoid having significant monthly variability particularly if it is not constructive to decision making.
CAC or Cost to Serve
Making allocation decisions between whether an expense is classified as cost to serve (lowering LTV) or (increasing) CAC can be meaningful for the picture of the business you are painting.
Let’s consider the arithmetic first and fundamentals later.
On a 5-year basis the SaaSy company in scenario A makes $10,000 revenue per year, with a gross margin of 80%. They have an extremely healthy LTV:CAC over 5x.
Consider if they were able to justify reporting an expense as a cost to serve rather than CAC. Presented as Scenario B the LTV:CAC becomes a whopping 7.5x.
In deciding what sits in CAC or cost to serve, it boils down to periodicity. If the expense is a one-off it should sit in CAC, if it is recurring or is linked to customer revenues then it is cost to serve.
A debatable example includes include discounts, promo codes and referral monies granted to customers. If it is a recurring discount, such as a discount for the life of a contract then it probably should sit in as a cost to serve. Referral codes are interesting, and they can be an instrumental channel for businesses to scale (particularly in the crypto space). I would argue the payment to the referrer is CAC and the payment to the new user goes to cost to serve.
If there is a free trial period granted to customers, I would recommend you don’t record the revenue for period x until you receive cash payments, rather than double enter the numbers as revenue and then CAC (giveaway). You’ll be inflating CAC and this methodology wouldn’t align with accounting standards on revenue recognition.
Fractions
The illustrative example above showed the power of financial storytelling and where numbers sit. In an example of LTV over 5x the differences are directionally less important. However if the LTV:CAC is hovering around 3x, then differences is more meaningful to prospective investors.
LTV/CAC is a fractional representation of the long-term profitability of your go to market strategy. It becomes more and less useful at certain ranges.
Take the LTV:CAC of the low margin business. If they improve their gross margin from 20% to 25%, their LTV improves by 25%. For the high margin business, improving the absolute cost to serve by the same $50 only garners a ~6% increase in LTV:CAC. A useful reminder of the reality of absolute versus relative differences.
A way to avoid becoming swayed through numerators and denominators is to focus on LTV less CAC. LTV less CAC actually is what equals your long-term profit per customer, presented in dollars not reciprocals.
Under the following Low Margin Business, Scenario X and Y, they either +/- cost to serve by $50 or +/- CAC by $50 respectively. The LTV:CAC ratio changes but the LTV less CAC is constant. Unit economics can make you spellbound at times, but it’s simply a way to present the profitability of customers, focusing on the $ value again often puts things back in perspective.
Time Value of Money
Unit economics makes long-term estimates of the revenue generation from customers without a nod or a wink to conventional finance’s greatest mechanism, discounting of cash flows. Applying a discount rate, even a modest discount rate is not convention in the world of unit-economics, and I doubt would gain traction.
There are three reasons why discount rates don’t matter.
Payback period should guide your decisions. Any initiative with a payback period over 3 years should be greatly questioned. If the payback period is short then the incremental specificity added by discounting your cashflows is negated.
Low interest rates. Unless interest rates materially rise from today, discounting cash flows as a piece of analysis is less relevant than in a high interest rate environment. VCs investing in your startup have been given a long-period to invest by their original investors, discounting cash flows inherently places a preference on shorter term investments rather than strategically
No one does it today. There’s a lot of benefits to sticking with what the crowd does when it comes to financial reporting.
Wrapping Up
Unit economics are a powerful tool to demonstrate the viability of a business model. Highly economically profitable go to market functions can be cash flow negative and lead to income statements in reporting periods resulting in large losses.
For leaders in a business, placing unit economics at the heart of your business by making them the key result your team is measured against is a way to enforce strategic and commercial alignment. For would-be investors unit economics are a way to separate the wheat from the chaff. How well a company uses unit economics is a major step in determining the former or latter.