#108 - Business Model Triage

Customer-Based Corporate Valuation

Jan 23, 2022

Hi 👋 - Customer-based corporate valuation (say that three times fast) is a tool for modeling customer behavior and triaging unprofitable business models. Today, a look at CBCV’s key insights. Thanks for reading.

If you’re finding this content valuable, consider sharing it with friends or coworkers. ❤️

For more like this once a week, consider subscribing. ❤️

Going For Broke

Losing money has never been more fashionable1. Many recent IPOs like All Birds, Rent the Runway, and Warby Parker are growing quickly, but losing money. Customer-based corporate valuation (CBCV) is a tool for triaging unprofitable business models and assessing paths to profitability. It does this bottoms-up by analyzing customer behavior and unit economics.

If a cohort analysis and a DCF had a baby, it would be CBCV. The model’s champion is Daniel McCarthy, an entrepreneur and Assistant Professor of Marketing at Emory’s Goizueta School of Business. Below is a non-technical primer (here’s a technical one). Even if you never touch financial models, a basic understanding of CBCV offers insights for business analysis.

CBCV 101

CBCV starts from first principles: all revenue comes from customers. Using customers as the unit of measurement, it distills revenue into its constituent components, modeling each separately2. The goal is predicting customer behavior. In particular, the number of customers acquired and the average revenue per customer. To do this, CBCV answers five questions3:

Customer Count: How many customers are acquired? This is a function of customer acquisition, customer retention, and market size.
CAC: Customers can be new or returning. Acquiring new customers costs money. If the answer to “do I need to spend this every time I bring in a customer?” is yes, then that expense is part of CAC4. For e-commerce and DTC businesses, a big slug of this will be Google and Facebook. Other examples include Toast's point-of-sale hardware (in addition to sales and marketing expenses) and Warby Parker’s home try-on costs (in addition to media spend). While a rising CAC often spells danger, what’s important is the cost effectiveness of acquisition spend (e.g. ROI).
Retention: How long do customers stay? A related concept is churn, which measures what percentage of customers stop buying in a given timeframe. High retention (low churn) is good.
Order Frequency: How often do customers transact? How many orders are placed during a customer's lifetime? Once again, higher is better. Repeat customers tend to be higher quality and typically don’t need to be reacquired making them more profitable.
Average Order Value (AOV): How large is each order?

CBCV combines these components to forecast revenue:

20/20

Enough theory. Warby Parker’s S1 illustrates CBCV components in the wild:

CAC5: Warby Parker’s CAC increased to $40 in 2020 from $26 in 2018 as it deliberately spent more to grow brand awareness, promoted new products like contacts, exams, and visions tests, and had higher home try-on expenses due to a Covid-related bump in e-commerce orders6. Those explanations are anodyne enough, but investors should monitor this closely. If CAC continues to rise, that’s a red flag unless it’s offset by a proportional increase in revenue.

Source: Warby Parker S1, August 24, 2021.

Average Revenue per Customer (ARPC): ARPC grew to $218 in 2020 from $188 in 2018, a 16% increase. This metric combines average order volume and annual purchase frequency. Over the same period, CAC increased 54% to $40 from $26, reducing Warby Parker’s contribution profit and contribution margin7. Revenue growth in isolation isn’t valuable. You can sell $1 for $0.80 until you go bust. More important is the investment required to generate that revenue (i.e., what’s the ROI?). Warby Parker’s unit economics are healthy, but they’re less healthy in 2020 versus 2018. That’s another area for investors to monitor.

Retention: Sales Retention Rate measures the revenue from each cohort as a percentage of the initial cohort revenue. For example, if the 2015 cohort spent $100 in 2015, it spent another $24 in 2016, $24 in 2017, $25 in 2018, and $24 in 2019. Between 2015 and 2019, it spent roughly $200, with $100 in 2015 and the remainder spread over the following years. Stability across cohorts is a good sign, suggesting durable revenue growth. However, while aggregate cohorts are stable, individual customers are heterogeneous. For example, customers that purchase contacts and receive an eye exam have higher retention rates8. That’s one of the reasons why the company is spending to build awareness of these services.

Limitations

All models are wrong, but some models are useful9. CBCV has it limitations. At a business level, a prerequisite for utilizing CBCV is knowing who your customers are and their transactional data. If you don’t know how many customers you’ve acquired, it’s impossible to calculate CAC. The growth of subscription, e-commerce, DTC business models, and advanced point-of-sales hardware has reduced these limitations. However, brick-and-mortar stores without a CRM or lots of cash transactions still might struggle.

The largest limitation for investors and analysts is data access. Understandably, companies limit what they disclose in SEC filings. The examples from Warby Parker above are about as good as it gets. Additionally, comparing companies is difficult due to a lack of standardized definitions. For example, Warby Parker includes store operating expenses in its definition of contribution profit while All Birds doesn’t. Three different companies will have three different definitions for the same metric. Analysts need to be on guard for community adjusted EBITDA.

Unit Economic Land Mines

The ability to effectively acquire and retain customers determines a company’s profitability (or lack thereof). Rising CAC is a canary in the coal mine. McCarthy’s research found that when CAC goes up, customer quality tends to go down10. This creates a double jeopardy situation: a company works harder and spends more to acquire a lower quality customer (e.g., higher churn, lower purchase frequency). This is the cause of death of many once high flying growth companies. A corollary is that finding good customers usually gets harder over time. As Tren Griffin points out, a rising CAC can indicate that a business is hitting its TAM ceiling:

Tren Griffin @trengriffin

5/ Simba: "Is TAM best thought of as a probability distribution with a range of possible outcomes? Mufasa: "Exactly. But there are clues. For example, a rising CAC is a sign you are starting to reach a TAM ceiling. Everything is connected in the Great Circle of Unit Economics."

The Lion King Simba GIF by Walt Disney Studios

Even without access to transaction logs, the CBCV framework provides useful sniff tests:

Purchase Frequency & First Order Profitability: If the first order isn’t profitable, the business needs multiple orders to recoup its CAC and even more orders to produce a profit. When the first order is profitable, additional orders run up the scoreboard. This is why repeat rate matters.
Repeat Rate: To McCarthy, repeat rate is a crucial metric in assessing the health of a business model11. If customers don’t come back, a business becomes reliant on acquisition, which gets expensive quickly. Strong repeat behavior gives companies confidence to invest in acquisition.
Decreasing CAC: Unless you’re a network effects business (and you’re probably not), CAC likely increases over time. Forecasts for decreasing CAC deserve extra scrutiny.
Decay Curves & Depletion: McCarthy likens customers to oil wells with depletion curves12. At some point, you run out of plum spots to sink wells. While decay curves within cohorts might look similar, decay curves between cohorts could differ as customers become more marginal over time. Assumptions about improving customer quality are another area deserving additional scrutiny.
Customer Heterogeneity: Another consideration is how customer behavior evolves over time. Past performance is not necessarily indicative of future behavior. Not all customers are the same. Not all leads are the same. Not all traffic is the same. There is no average customer. Tracking quarterly or annual customer cohorts helps to show what is happening on the margin.

These aren’t laws of physics, but they still could help you avoid a WeWork.