2.1.4 Customer Growth Models

Sample Full Marketplace Growth Model - Growthzilla Book

This post is part of the Growthzilla Book series, which is an online draft of the print edition that will be available in 2018.

In the previous section we covered common revenue growth models to understand how companies make money on different types of products. In those models, revenue was driven by essentially the same three factors:

  • the number of customers making purchases,
  • the number of purchases that those customers make, and
  • the size of those purchases.

The number of customers that a company has depends on how well it acquires new customers and retains existing ones. On the other hand, how often those customers buy your products and how much they spend per transaction is based on how well the product and the entire customer experience is able to engage them. The models in this section focus specifically on increasing the customer base and how much customers buy.

Lifecycle in Revenue Growth Model


Even though revenue depends both on how many customers a business has and how much they purchase, many business leaders focus on growing only the number of customers. Moreover, they tend to focus on growing the customer base by acquiring new customers without giving much thought to retaining existing ones. This trend is unfortunate because increasing how often customers buy and the size of those purchases can also be an extremely effective way to grow revenue.


Basic Customer Growth Model

The concept that your customer base will grow bigger if you improve how well you acquire new customers is a simple, universal belief. Perhaps this is why many businesses start with trying to improve acquisition in order to grow the customer base and revenue. However, the rate at which the customer base grows is not just a function of acquisition but also depends on how well the business is able to retain existing customers. Neglecting existing customers can really drag down your growth because you are essentially fighting an uphill battle. Every time you get more customers, you lose some number of them, and the resources that went into acquiring them are squandered. Consider the major cellular providers such as AT&T and Verizon. They know that retention is critical to the profitability of their business, so they bake it into their business model with long-term contracts. Moreover, they give great deals to existing customers when their contracts run up to make sure that existing customers stay loyal to them. Of course, no business can retain all of their customers, but the ones that do a better job hanging onto theirs have a leg up on the competition.

Basic Customer Growth Model

Let’s take a simple example to see how both acquisition and retention factor into the growth of your customer base. Imagine that you are able to acquire a hundred customer per day. At the same time, you lose one out of four of those customers the very same day. Perhaps they sign up for your product, but then realize that it does not really suit their needs and cancel their account right away. Therefore, the effective customer growth rate is seventy-five customers per day. In reality, the dynamic between acquisition and retention is a little more complicated.

First, the acquisition rate is rarely linear over time, so the business might be acquiring a hundred customers per day one month and three hundred customers per day the following month. Often times, seasonality plays a factor in acquisition rate. However, these nuances only matter when you are trying to develop more rigorous models for predicting growth. Just understanding the concept that acquisition and retention play a part in the net growth rate is enough for a high-level model of your business that will feed into your growth strategy.

Second, the retention rate is also not typically a simple function like described above. Instead, retention is almost always a function of time where the more time passes, the more customers abandon the business. For example, let’s say that your business acquired a hundred new customers today. It might be that ninety-eight of them will still be active customers tomorrow, but only half will be your customers in a year from now. Therefore, it is important to compare acquisition and retention rates over the same time periods to get an accurate sense of overall growth rate.

Third, the retention rate is often calculated in terms of the overall customer base rather than by cohort. This is a critically important distinction because how you calculate it will often give you vastly different results. Let’s consider the cohort example first. Imagine that you acquire 1,000 customers in a month, and the cohort retention rate is 75% for the second month. That means that of the thousand new customers that you acquired in the first month, you will have 750 in the second month.

Cohort vs Absolute Retention

Now let’s consider an absolute retention rate where your business loses 25% for your entire customer base every month, so you retain 75%. After a few years in business, you have 10,000 existing customers and your acquisition rate is a thousand new customers per month. That means that you business will be adding 1,000 new customers but losing 2,500 existing customers resulting in a net loss of 1,500. As you can see, the result is very different in this case than in the cohort analysis, and this is important to keep in mind when more rigorous calculations are needed.

As you can see, we are continually zooming in. We started with the big picture and revenue equations. Then we focused on how acquisition and retention work together to drive customer growth. In some cases, however, it’s important to focus solely on the acquisition part of the customer growth equation. The next three models provide insight into factors that drive acquisition rate.


User Generated Content Growth Model

Many websites and apps rely on user generated content to provide value to their customers. Products such as Twitter, Wikipedia, and YouTube all rely on some of their users creating new content for others. The most successful of these are able to build thriving businesses by creating a sustainable ecosystem of content creation and consumption.

The way user generated sites work is quite simple. The majority of users will just consume the content that is being created. However, a small portion of the overall user base will create content. As that new content is created, it will draw in even more new users, and some of those new users will also create new content repeating the cycle, which is modeled by the equation below.


User Generated Content UGC Growth Model

For example, let’s say that we’re starting out with 1,000 users in the first period. Most of those users will just view content, but on average, one post will be created for every ten users. Those posts might get indexed by Google, and each of the new posts will then attract fifteen new users when they search for related topics. That means that in the second period, we should have 1,500 new users or a total of 2,500 users, and the cycle repeats agains. The important things to note from the above model is that the rate at which new users are acquired is driven by the rate at which new content is created as well as the rate at which that new content attracts new users.


Viral Growth Model

Some products such as social networks rely heavily on existing customers inviting others to grow their customer base. This acquisition model is often called the “viral model” although to be truly viral, each user has to successfully get more than one new user to sign up, on average. All the same, viral growth can be modelled similarly to user generated growth, where the number of new users is a product of the existing users, the average number of invitations per user, and how many of the invitations are accepted.

Viral Growth Model


For example, if we have a thousand users in the first period, and each of those users invite an average of two people in that period, they will send out 2,000 invitations. Let’s now suppose that only one in four of their invitations get accepted by their acquaintances. That results in 500 new users in period 2. Therefore, you will have a total of 1,500 users in the second period, and the cycle repeats.

One thing to note is that the invitation and acceptance rates need to be over the appropriate periods of time. For example, we could figure out the average number of invitations that existing customers send over a day, week, month or a year. We would need the same period length for comparing the acceptance rate. In other words, if we are calculating the invitation rate over a week, we need to use the acceptance rate over a week as well.

Another point to note is that the acceptances have to happen after the invitations, so in order to be truly accurate we have to remember to take into account a time period lag. For example, if current users send the invitations in period 1, the acceptances will happen in period 2. The above model captures this in a simple way, but you might have to be more explicit in your equation to be more rigorous.


Paid Acquisition Growth Model

There are a number of products that cannot greatly rely on growing their customer base through viral features or user generated content spurring search engine traffic. These products are often in the software as a service (SaaS) space or other commercial product space. In these cases, those companies have to rely on paid acquisition such as paying a partner company money for getting their customers to try your product. This strategy works great when existing customers pay for the acquisition of new customers through the revenue they generate for the company, which the company reinvests in paid acquisitions.

The cycle transpires as the following. First, existing customers generate revenue for the company, which translates to profit after we subtract the expenses of serving them. The company then decides how much of that profit to reinvest into paid acquisition — the investment rate. Together, this determines the total amount that the company is going to spend on acquiring new customers in a given time period. However, that’s not the end of the story since a dollar does not necessarily buy you one customer. Each new customer will cost you a certain amount of paid acquisition spend, which we call the acquisition rate per spend. Now, putting all these parts together will give you how many new customers you can get in the second period if you start with a given number of customers in the first period.


Paid Acquisition Growth Model

Let’s imagine that we have 1,000 customers in period 1, and each one generates an average of $10 of profit for the company for a total profit of $10,000. The company decides to reinvest one out of every five dollars for paid acquisitions, costing $10 for every new customer. In period 2, the company can expect 200 new customers or a total of 1,200 customers.

The great thing is that the above model can apply to any paid marketing activity that is subsidised by profit earned from existing customers. Simply generalize paid acquisition to paid marketing, and the model still holds true. The acquisition rate will take into account the probability of acquiring a new customer if the conversion rate is not one hundred percent.


Lifetime Value (LTV) Model

The last model that we’ll cover in this section is not really a customer growth model nor is it a model for the total revenue. Instead, the customer lifetime value model shows us how the average revenue that a company earns from an individual throughout the time that he is an active customer is the product of the average revenue that a customer brings in a given time period and the average lifetime of the customer. The customer lifetime value model helps us see that we can grow revenue by either increasing how much customers spend in a given period or how long they stay active customers. For example, imagine that the average amount that your customers spend per week is $15, and they tend to remain customers for ten weeks. (In other words, your company loses one customer every ten weeks — the churn rate.) That means that the average lifetime value (LTV) is $150 per customer.


Lifetime Value LTV Model - Growthzilla Book

It’s worth noting that the above model is a simplified version of a more rigorous model that takes into account interest rate for a discount rate and a gross margin. Nonetheless, this simplified model provides great insight into how retention rate affects revenue.


Combining Customer Growth and Revenue Models

As we saw with the revenue models in the previous section, the number of customers that you have greatly determines how much money your business can make. While the revenue models provide a broader picture of the different ways to influence revenue growth, a key driver is often simply how many customers your business has. The customer growth models in this section can be plugged into the revenue models in the previous section to give you a deep understanding of factors that contribute to increasing the number of customers, which in turn influences your company’s revenue.


Sample Full Marketplace Growth Model - Growthzilla Book

For example, we discussed how the total revenue that the company makes from the marketplace is a product of the number of items available for sale from sellers, the probability of a sale, and the average transaction fee. If you wanted to see how a feature allowing users to invite their friends might play into that model, you can simply plug in the viral growth model into the marketplace revenue model to see how virality can affect the number of sellers, which drives the number of products for sale and ultimately revenue. While this is one combination, you can combine any of the customer growth models with the revenue models to gain deeper insights into factors that you might optimize as part of your comprehensive growth strategy.

Be sure to check back tomorrow to learn about strategizing and prioritizing experiments. New sections of Growthzilla are published every weekday.

2.1.3 Common Revenue Growth Models

SaaS Revenue Growth Model

This post is part of the Growthzilla Book series, which is an online draft of the print edition that will be available in 2018.

When modeling a business, it’s best to begin with its core: how it generates revenue. By starting at the most fundamental level, you can progressively drill deeper into the variables that affect revenue to understand what factors are ultimately responsible for the success of your company. Furthermore, you can map key stages of the customer lifecycle to the factors that drive revenue. For example, one common realization that business leaders make when they really evaluate their revenue model is that increasing how often existing customers make purchases can be just as effective as garnering new customers. You could also just focus on modeling pure customer growth, which we will cover in the next section, but starting with how your business makes money will help you to tie everything together. In this section, we’ll start with the simplest revenue model and explore how it can be applied to different types of businesses.

Simplest Revenue Model

How much money a company makes quite simply depends on how much stuff it sells and how much revenue it makes per unit. This is the most basic revenue equation known to mankind, and even though it’s trivial, it’s the perfect jumping off point for us to delve deeper into more sophisticated models. We can start by breaking down what factors drive the number of transactions (or sales). The more active customers, the higher the number of transactions. Moreover, if each customer makes more purchases, that will also increase the total number of sales. Therefore, the number of transactions that a company gets is a product of the number of customers that it has and how many purchases each of those customers makes.

Simple revenue growth model

The amazing thing is that even by digging one level deeper, we can start to see insights! Since the number of transactions is a function of the number of customers and their average purchase volume, we can ask ourselves how can we get more customers or how can we increase how many purchases each customer makes. We can start to form hypotheses around increasing each of those factors independently. For example, we can assume that we can get more customers by increasing marketing spend or by introducing a feature where customers can invite their friends for cash credits.

The underlying factors that drive overall revenue vary with different business types, which we will explore next, but I will also note that there is no right way to break out the revenue model for any business. The examples below are simply meant to give you a starting point, and you’ll need to contextualize them to your business. Another important note is that some of the models below do not take time into account, but you can simply add a temporal component by applying the equation to any period of time be it a month, quarter, or year.

E-Commerce Revenue Model

In order to model an e-commerce business or any retail business you should explore what drives the volume of transactions as well as average revenue per transaction. There are many ways to think about this, but one common perspective is to consider the transaction or sales volume to be a function of the total number of products that are being sold in the online store as well as how likely they are to be sold. The probability with which a sale is likely to be made could also be influenced by how many products the average customer views as well as how likely they are to purchase each product. You could dive even deeper and assume that the number of product views is a function of how many visitors you get and how many products each one views, on average.

Ecommerce revenue growth model

Let’s not forget that we could also increase revenue by increasing how much an average customer spends. What drives the revenue per transaction? That question might be a little more tenuous. It could be the nature of the goods you’re selling (budget vs. luxury), how easy it is to find products, the checkout conversion rate, or the amount of cross-selling that you do on each product page. In reality, probably all of these things have an effect on the average revenue per transaction, and the best way to figure out on what to focus is to go back to your trusty customer journey map and identify how your customers perceived each of those experiences.

SaaS Revenue Model

The following model has a time component since software as a service businesses tend to worry about revenue on a monthly basis given the typical billing cycle for this class of products. At the simplest level, monthly revenue for SaaS companies is a product of the total monthly subscribers and the average monthly subscription cost. One could break down the number of monthly subscribers as being the product of the total number of customers that are paying for subscriptions and the average number of accounts or seats that each buys. Delving a little deeper into the model, one could assume that the number of customers is a function of marketing effectiveness or the size of your sales team. Furthermore, the number of subscriptions that each customer buys might be greatly affected by how easy it is to invite colleagues.

SaaS Revenue Growth Model

What is a bit unique in the SaaS model is that it’s often more difficult to affect the monthly subscription cost if the business is selling just one product, in which case the pricing is often set by market conditions. In other cases, when the business sells multiple products or tiered plans for access to advanced features, it is possible to affect the monthly subscription cost. Moreover, the cost of enterprise subscriptions is often tied to the size of the client company, so larger companies often pay higher subscription fees for bigger teams. For SaaS companies that have tiered pricing, as most do, it’s wise to create a model for each tier when performing more rigorous analysis or creating predictive models.

Marketplace Revenue Model

Marketplaces such as eBay or Uber bring buyers and sellers together and make money by charging a transaction fee. The total revenue that the company makes from the marketplace is a product of the number of items available for sale from sellers, the probability of a sale, and the average transaction fee. Each of these factors can be further broken down into underlying factors that drive them. For example, we can assume that the number of items for sale in the marketplace is driven by the number of sellers on the platform as well as how much each seller posts, on average.

Marketplace Revenue Growth Model



We can also think of the probability that a sale will be made as a function of how many times buyers view the item as well as the conversion rate for purchases (e.g. how many purchases are made for every thousand views of products). The number of views of product also depends on how many buyers are on the marketplace and how engaging the buying experience is. For example, if the product images are really unattractive, buyers probably won’t spend a ton of time browsing products.

Finally, the average transaction fee can also be affected by a number of factors. The two most obvious ones are the fee that the company is charging as well as the average sale price since many marketplaces charge a percentage commission. The average sale price could also be determined by the type or quality of goods that are being listed on the marketplace. It’s also worth noting that, in some cases, the marketplace charges flat fees. For example, many job sites charge a flat fee for every job posting that the employer publishes.

Community Revenue Model

There are a large number of online and mobile products that enable people to connect and share content with each other. The class of products include social networks and user-generated content sites such as Facebook, Twitter, Instagram and Quora. One thing they have in common is that they make money by displaying pay-per-click advertisements on their pages and in their apps. The revenue that these kinds of companies make is a product of the total number of clicks on ads and the average cost per click (or revenue per click). This model, with slight variation, can also apply to any business driven by cost-per-click advertising.

Community Growth Model

The easier of the two variables to optimize is usually the number of clicks on ads, which we can think of as a function of the average ads per page, the number of times that users view those pages (page views), and the rate at which those users click on the ads. Diving a level deeper, we might assume that the number of ads on the site or in the app is determined primarily by the number of advertisers and the average number of ads that those advertisers buy. The number of pages that are viewed by users is probably a function of the number of users and how engaging the content on those pages is.

The average cost per click is a little more difficult to model, but it is usually driven by two key inputs: the number of advertisers bidding on ad placements and the categories that are being advertised. For example, costs of ads for lawyers are usually a lot more than for common commodities such as grape jelly. Another thing to note is that the cost per click of ads almost certainly impacts how many of those ads or impressions advertisers buy.

Key Takeaways for the Sample Revenue Models

The main point to understand is that the above models are just simple examples, and although we hope they are very relevant for your business, you should use them as starting points to develop your own model. As you are creating a revenue model for your business, you should lean heavily on the customer journey map that you developed since it will provide you with great insights about what variables are important for growth and how they relate to each other. Finally, it’s important to remember that other growth models such as those for viral growth usually plug into the left hand side of the above models and drive the number of customers that a business has. We’ll cover customer growth models and other useful models in the next section.

Be sure to check back tomorrow to learn about different kinds of growth models that you might use. New sections of Growthzilla are published every weekday.