Customer cohort analysis & KPIs
An acquirer’s lens to understanding revenue quality
One subject I've been eager to explore is cohort analysis and its relevant KPIs. While I may have been contemplating a bit too long, as many other writers have posted about the topic, today I hope to share a different perspective that will prioritize evaluating a business model from the perspective of an acquisition entrepreneur. The truth is that obtaining clean data sets for extensive analysis is often a pipe dream when you're in the initial stages of evaluating a company. It’s not like you are in the operating seat and have better access to near-perfect information, so you must cut some corners and make a judgment call based on the information at hand.
With today’s post, let’s assume that you’ve had a preliminary conversation with a business owner and, through your qualifying questions, have determined that there may be a potential partnership between you and the business owner. For next steps, you’ve requested some financial information from the owner and a granular sales by customer by period by product data set for the past four years. And because we’re really reaching for assumptions here, I will go on a limb and say that the owner sends this information over promptly.
From here, I’ll discuss conducting the cohort analysis and some key metrics you can easily calculate with this data in mind, although this exercise has many caveats. I'll try to highlight as many nuances as possible, but some will inevitably slip through the cracks. To cover all my bases, I will share other helpful resources you can leverage as you determine what makes the most sense for you. Depending on the level of interest, I am more than happy to elaborate on these topics in future posts.
One last thing before diving in. While cohort analysis and its related KPIs are well-understood in venture capital and enterprise SaaS, this analysis can be broadly applied. It's relevant for any business with a reasonably stable customer base and predictable revenue streams, especially those relying on subscriptions or long-term contracts.
Here’s the agenda:
Nature of revenue streams
What is cohort analysis?
Setting up the analysis
Key metrics
Some nuances
Additional resources worth exploring
Wrapping up
Nature of revenue streams
Building upon the article by A.J. Wasserstein, Mark Agnew, and Brian O'Connor of On the Nature of Revenue, revenue can be classified into five major categories:
Contractual Recurring Revenue: Customers are contractually obligated to use the service or products over multiple periods.
Non-contractual Recurring Revenue: Customers voluntarily subscribe without a contract, but the usage pattern remains regular and predictable.
Repeat Revenue: Customers make recurring purchases without a contractual obligation.
Actuarial Revenue: A predictable revenue form based on similar infrastructure and consistent customer cohorts.
Transactional Revenue: These are one-off, unpredictable revenue events.
For the pragmatic searcher looking to acquire a business, the emphasis should likely be on the first two categories, with occasional situations where a repeat revenue business may be highly attractive. Here’s why:
Strong preference for contractual relationships
Contractual relationships serve as an economic moat for businesses, allowing first-time CEOs to focus on mastering the core competencies of the operation rather than waking up each day to a revenue stream that might be as unpredictable as a game of Russian roulette.
From an investor's standpoint, the predictability of these cash flows reduces the element of surprise, increasing its value, all else equal. The less time you spend chasing volatile revenue, the more time you can invest in strategic growth and operational efficiency.
Disclaimer: While contracts provide security, be cautious of the “ironclad contract” myth. Legal documents are only as good as the relationship behind them.
Don't dismiss non-contractual revenue
Many businesses successfully operate with non-contractual recurring revenue, and metrics like Customer Lifetime Value (CLV) can sometimes indicate a better retention rate than contractual models.
Consider multi-revenue stream businesses
It’s common to encounter businesses with more than one type of revenue stream. Understand the interplay between them. Do they complement each other or pose a risk?
Review revenue concentration
Scrutinize the number of recurring customers and the percentage of revenue contributed by the top 10 customers. If it's too high, even a 'secure' contract can pose a risk.
Seasonality & market cycles
Understand the effect of external factors on the predictability of the revenue streams. Are certain revenue streams more resilient?
Impact of customer behavior
Certain 'repeat revenue' businesses can show remarkable resilience due to high customer satisfaction or unique value propositions, sometimes outlasting contractual models.
Switching costs
Consider any contractual or non-contractual barriers that might make it costly or cumbersome for customers to leave, enhancing revenue predictability.
Engagement level
For non-contractual and repeat revenue businesses, customer engagement levels can indirectly measure revenue quality. High engagement often means higher predictability.
Conclusively, while there's a strong preference for contractual relationships, taking a nuanced view is imperative. The real world often serves up a mixed bag of revenue types, and blindly adhering to a single 'preferred' type can lead to missed opportunities or even mistakes in valuation.
What is cohort analysis?
In analytics, especially when evaluating potential acquisitions for a search fund, cohort analysis can be an invaluable tool for due diligence. Appcue defines cohort analysis as “a type of behavioral analytics in which you take a group of users, and analyze their usage patterns based on their shared traits to better track and understand their actions. A cohort is simply a group of people with shared characteristics.” In this case, our users are paying customers, and we may look at them based on the time in which they signed up for a product or service (most common), or some behavior that you may want to track. For the sake of simplicity, we’ll only discuss the first case: acquisition cohorts.
Why does it matter?
Customer retention
Knowing how many customers continue to use the service or product after an initial period can significantly inform the valuation of a prospective business. Higher retention rates often correlate with a higher Customer Lifetime Value (CLV), making the asset more attractive for acquisition.
Revenue predictability
Cohort analysis can provide a granular understanding of a business's revenue patterns, enabling you to discern whether the revenue stream is sustainable and predictable—a critical factor in valuation and risk assessment.
Product-market fit
Revenue increases within specific cohorts over time can serve as a strong indicator of product-market fit, pointing to the business's growth potential.
Cost management
Cohort data can inform resource allocation decisions. For example, if a vertical B2B SaaS company finds customers from its enterprise segment are 30% more profitable than those from SMB, marketing and sales efforts can be tailored more towards larger prospective customers, optimizing resource spend.
Caveats and limitations
Data quality
Your analysis is as good as your data. If your data is incomplete or inaccurate, your findings will be too. One gut check early in the process is to see how the granular sales data compares to recurring revenues identified on the income statement if you have access to both.
Operational feasibility
Cohort analysis is data-intensive and requires a sophisticated operational setup to manage this effectively. Sometimes, you may find that the owner has not looked at the business this way before, so the data is not readily available when you request it.
Scalability
For small companies or those with less operational history, the results from cohort analysis may offer limited actionable insights due to smaller cohort sizes. This may be especially true for businesses with less than $3 million in ARR.
In the context of a search fund transaction, cohort analysis is not merely an academic exercise but a lens through which you can more accurately gauge the quality of the asset you're attempting to acquire. It can guide you in deciding whether to disqualify the lead, how to value the company, prioritize due diligence questions, and optimize post-acquisition operational strategies in the event of a successful transaction.
Setting up the analysis
Before diving into the intricacies of cohort analysis, it's paramount to set the stage with the right data. As we explored in a previous post, Cracking the code, a granular dataset covering sales by customer and by product over time, will serve as the bedrock of your analysis. It's akin to laying a strong foundation before constructing a skyscraper: get it wrong, and the entire building could be at risk. A meticulous due diligence process isn't just a mere formality in this context; it's the financial equivalent of a medical check-up before running a marathon. Fail to scrutinize your dataset adequately, and you're essentially flying blind through a storm—a risk no seasoned investor or operator should willingly shoulder.
After you've solidified that robust data foundation, performing calculations becomes a relatively straightforward endeavor. There are several key metrics to focus on, each providing a unique lens through which to understand customer behavior. These are:
Retained Revenue: This represents the customers acquired in a particular previous period and still active. Retained customers contribute to both recurring revenue and offer a strong indication of customer satisfaction.
Expansion Revenue: Customers who not only stayed but also increased their consumption of the company’s services. These often contribute to a higher MRR or ARR and are indicators of upselling and cross-selling opportunities.
New Customer Revenue: Freshly acquired customers within the cohort period. This offers a snapshot of the company’s customer acquisition capabilities.
Downgrade Revenue: Customers who have reduced their consumption of the company’s services. While they are still clients, their reduced engagement can be a warning sign.
Lost/Churned Revenue: Customers who have entirely stopped using the company’s services. Losing these customers impacts the periodic churn rate and provides insights into possible problem areas.
Resurrected Revenue: While less common, these are customers who had churned but have come back. Understanding why they returned can offer interesting insights, and for the sake of simplicity, we will ignore resurrected customers for this post.
Data structuring
To set up this analysis in a tool like Excel, one way to approach it is by using logic-based functions such as IF and AND statements to identify what type of revenue a particular customer should be identified as in a specific year or month. Because we mainly care about where there has been a change in revenue period-over-period, either up or down, it helps to build four tables that focus on a particular component (expansion, downgrade, new, and churned) that is at the customer level and broken down into period spend to the company.
I would share more specifics on how to go about building the structure step by step, but I want to be mindful of the templates I have received from multiple investors and respect their privacy. I suggest digging on Google and through your professional network to get a template to get the basics down and then modify it for your needs. Here’s a good starting point from Lifetimely.io.
Once you have structured the data properly, you can then begin to analyze sales by cohorts and the behavior of a particular cohort over time in grids such as the one below. What I commonly see is the cohort will be placed in individual rows (up and down) while the number of months since the first order will be in columns (left to right). Looking vertically on the grid will help you understand retention behavior and spending after a particular period, which allows you to spot any potential product wins or issues that may be developing if there is a trend. Looking horizontally on the grid will help you understand the retention and spending behavior of a particular customer cohort over time. You can even break this information down further to highlight specific elements that make up the cohort: retained revenue, expansion, downgrade, etc.
Key metrics
Monthly Recurring Revenue (MRR)
The sum of all recurring revenue generated within a specific month.
MRR = Sum of monthly subscriptions
MRR is akin to the monthly 'salary' your business earns from subscriptions, offering a snapshot of immediate financial health. This is less relevant for businesses that sell only annual or multi-annual contracts with their customers.
Annual Recurring Revenue (ARR)
Run-rate of the business’s recurring revenue at a given point in time.
ARR = MRR × 12
ARR provides a longer-term view of your revenue, often scrutinized by investors to gauge the company's growth potential. Another way to look at ARR is simply the sum of the annual value of the company’s active contracts if it does not sell monthly subscriptions.
Gross Revenue Retention (GRR)
Measures the retained revenue from existing customers, excluding upsells and cross-sells.
GRR= (Beginning ARR - Downgrade - Churn) / Beginning ARR
GRR provides a conservative perspective on your revenue stability, excluding any revenue expansion from upsells and cross-sells. Some people may look at GRR by only considering churned revenue, but this understates the magnitude of downgrades. I prefer to capture both in the calculation.
If you are interested in looking strictly at churn and there is any difference between the two, you can calculate Revenue Churn separately by taking the following calculation:
Revenue Churn = Churn / Beginning ARR
Net Revenue Retention (NRR)
Measures the change in recurring revenue from existing customers, accounting for upsells, cross-sells, downgrades, and churn.
NRR = (Beginning ARR - Downgrade - Churn + Expansion) / Beginning ARR
NRR tells you how well you monetize existing customers. An NRR over 100% means you're growing revenue from your current client base. Note that this does not include any new customer revenue that was generated in the period you are measuring.
Logo Retention
The percentage of customer accounts retained over a specified time frame.
Logo Retention = (Ending Logos - New Logos) / Beginning Logos
This reveals how many of your customer accounts (logos) have remained with you over a specified period, indicating overall customer satisfaction.
Average Order Value (AOV)
The average amount a customer spends in a single transaction.
AOV = Total Revenue / Total Number of Orders
AOV shows the average expenditure per transaction, indicating product-market fit or the effectiveness of upselling strategies.
Quick Ratio
A financial metric that quantifies the growth efficiency of a subscription business, considering both new and lost revenues.
Quick Ratio = (New MRR + Expansion MRR) / (Churned MRR + Downgrade MRR)
The Quick Ratio evaluates how well a company can grow recurring revenue despite customer churn. A Quick Ratio greater than 1 is usually a sign of a healthy and growing business, as it indicates that the company is adding revenue faster than it's losing.
Expansion MRR Rate
The rate at which Monthly Recurring Revenue (MRR) from existing customers grows due to upsells, cross-sells, or additional purchases.
Expansion MRR Rate = Expansion MRR / Beginning MRR x 100
This metric helps understand how well the business is maximizing the value of its existing customer base. A higher Expansion MRR Rate can offset churn and contribute to overall business growth.
Some nuances
While metrics like MRR, ARR, and others offer a structured approach to evaluating a business, they don’t capture the full complexity of real-world operations. Here are some considerations:
Monthly subscriptions in annual analysis
If the business you're analyzing mainly offers monthly subscriptions, conducting an annual cohort analysis will inherently introduce some margin of error. While less than ideal, this level of granularity may have to suffice for preliminary evaluations.
Variability in contract types
In their quest for product-market fit, small businesses often end up with a mix of annual, multi-year, and monthly customer contracts as they battle to win new business. Such a blend can obscure true customer satisfaction and revenue predictability. You cannot assume that every customer is priced at $X per user per month or that they perfectly follow a three-tiered good, better, best pricing strategy.
Unreliable renewal rates
If you compare two businesses with identical operations but different contract durations, one with monthly subscriptions may be more attractive. Why? Because the renewal rate of a multi-year contract is harder to predict and therefore introduces more risk. If you do not have information about contract renewals when doing this analysis, recognize that you may be overly optimistic about a business’s revenue quality based on past performance.
Customer concentration
I mentioned earlier how much customer concentration can influence the quality of a business’s revenue streams. Pay close attention to how top customers affect key metrics. For instance, if your largest customer expands their contract significantly, your NRR and GRR could become skewed, portraying a misleadingly rosy picture. This can lead to misinformed decisions if one doesn’t dig into the details.
Additional resources worth exploring
While this post aims to provide a comprehensive look at evaluating revenue quality through cohort analysis and key performance indicators, it is by no means exhaustive. As you refine your approach and gain access to additional data on a company’s financials, more nuanced metrics can come into play, further enriching your understanding of a business's true potential. If you're asking yourself, "What does 'good' look like in practice?" invaluable resources are available in the search fund and SaaS ecosystems. My top four references for deepening your understanding include:
Wrapping up
This post has explored the intricate process of assessing a business's revenue quality, focusing on the types of revenue streams and how to perform a detailed cohort analysis. We delved into various KPIs that can provide an insightful lens into the health of a business, all while acknowledging the nuances that often complicate these metrics.
If you found this analysis helpful, I encourage you to share it within your network. Furthermore, for those who appreciate this level of detail and wish to delve deeper into search funds, SaaS metrics, and more, consider subscribing to Maverick’s paid subscription. Your subscription will grant you access to in-depth reports, exclusive insights, and actionable advice to help you lead a more effective search.