NRR doesn't matter

What does this mean for those of us managing or trying to understand fast-growing companies?
October 7, 2022

I've recently been spending a lot of time studying the dynamics of Fivetran's revenue growth. As part of this, I've been playing with toy models of businesses with different levels of growth and net revenue retention (NRR), and I've observed something very strange. Consider a business that is doubling its number of customers every year, and its existing customers grow 50 percent each year. In the first year, the company starts at a $1 million annual revenue run rate, adds $1 million from new customers, and $0.5 million from expansion, for a total growth rate of 150 percent. Nice! Over the course of 10 years, this company grows to $2 billion in revenue. But something strange has happened: The growth rate has declined to just over 100 percent:

Year Starting annual
revenue run rate
New
customers
Growth from new
customers
Growth from
expansion
Growth
rate
1 $1,000,000 100 $1,000,000 $500,000 150%
2 $2,500,000 200 $2,000,000 $1,250,000 130%
... ... ... ... ... ...
9 $486,371,094 25,600 $256,000,000 $243,185,547 103%
10 $985,556,641 51,200 $512,000,000 $492,778,320 102%

(Source)

If you run the simulation forever, the growth rate will reach exactly 100 percent. And if you try this many times with many different initial conditions, as long as the growth rate due to expansion is less than the growth rate due to new customers, the long term growth rate will always converge on the customer growth rate. Why is this? Consider the following equation:

[Growth] / [Revenue] = [New] / [Revenue] + [Expansion] / [Revenue]

If overall revenue is growing faster than new customers, the term [New] / [Revenue] will get smaller over time.[1] Thus, the overall revenue growth rate will always converge to the customer growth rate, given enough time.[2]

We can get another perspective on this same mathematical observation by looking at the growth of revenue in terms of cohorts. This is our company with a 100 percent customer growth rate and a 50 percent expansion rate:

The x-axis is the year. The contribution of each cohort to revenue is shown as a different color. Now consider another hypothetical company, where the customers are twice as valuable to begin with but they don't expand at all:

The growth rate and overall revenue is nearly identical, and will continue to be forever. Positive NRR is equivalent to a higher customer value. This may seem strange, because NRR compounds and customer value doesn't, but as long as the customer growth rate exceeds the expansion rate, the expansion of the earlier cohorts will get washed out by increasing size of the newer cohorts. 

So far we've been simulating hypothetical businesses where customer growth continues forever and expansion compounds forever. In reality, customer growth eventually slows as we approach the total addressable market, and expansion stops compounding as individual customers fully adopt your product. But neither of these realities changes the core observation that your revenue growth rate converges to your customer growth rate over time.

What does this mean for those of us managing or trying to understand fast-growing companies? Despite my clickbait title, positive NRR is a good thing! But it's fundamentally similar to higher average customer value, not to a higher growth rate. A company with positive NRR is more valuable and will have more revenue, but the rate at which it grows is still fundamentally governed by its number of customers.

[1] In this example, the initial ACVs of new customers are flat. If the initial ACVs of new customers get x% larger each year, then the long-term growth rate of the company will be 100% + x%. Either way, expansion has no impact on the equilibrium growth rate.
[2] If the expansion rate exceeds the new customer growth rate, then the revenue growth rate will eventually converge to the expansion rate. 

Start for free

Join the thousands of companies using Fivetran to centralize and transform their data.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Data insights
Data insights

NRR doesn't matter

NRR doesn't matter

October 7, 2022
October 7, 2022
NRR doesn't matter
What does this mean for those of us managing or trying to understand fast-growing companies?

I've recently been spending a lot of time studying the dynamics of Fivetran's revenue growth. As part of this, I've been playing with toy models of businesses with different levels of growth and net revenue retention (NRR), and I've observed something very strange. Consider a business that is doubling its number of customers every year, and its existing customers grow 50 percent each year. In the first year, the company starts at a $1 million annual revenue run rate, adds $1 million from new customers, and $0.5 million from expansion, for a total growth rate of 150 percent. Nice! Over the course of 10 years, this company grows to $2 billion in revenue. But something strange has happened: The growth rate has declined to just over 100 percent:

Year Starting annual
revenue run rate
New
customers
Growth from new
customers
Growth from
expansion
Growth
rate
1 $1,000,000 100 $1,000,000 $500,000 150%
2 $2,500,000 200 $2,000,000 $1,250,000 130%
... ... ... ... ... ...
9 $486,371,094 25,600 $256,000,000 $243,185,547 103%
10 $985,556,641 51,200 $512,000,000 $492,778,320 102%

(Source)

If you run the simulation forever, the growth rate will reach exactly 100 percent. And if you try this many times with many different initial conditions, as long as the growth rate due to expansion is less than the growth rate due to new customers, the long term growth rate will always converge on the customer growth rate. Why is this? Consider the following equation:

[Growth] / [Revenue] = [New] / [Revenue] + [Expansion] / [Revenue]

If overall revenue is growing faster than new customers, the term [New] / [Revenue] will get smaller over time.[1] Thus, the overall revenue growth rate will always converge to the customer growth rate, given enough time.[2]

We can get another perspective on this same mathematical observation by looking at the growth of revenue in terms of cohorts. This is our company with a 100 percent customer growth rate and a 50 percent expansion rate:

The x-axis is the year. The contribution of each cohort to revenue is shown as a different color. Now consider another hypothetical company, where the customers are twice as valuable to begin with but they don't expand at all:

The growth rate and overall revenue is nearly identical, and will continue to be forever. Positive NRR is equivalent to a higher customer value. This may seem strange, because NRR compounds and customer value doesn't, but as long as the customer growth rate exceeds the expansion rate, the expansion of the earlier cohorts will get washed out by increasing size of the newer cohorts. 

So far we've been simulating hypothetical businesses where customer growth continues forever and expansion compounds forever. In reality, customer growth eventually slows as we approach the total addressable market, and expansion stops compounding as individual customers fully adopt your product. But neither of these realities changes the core observation that your revenue growth rate converges to your customer growth rate over time.

What does this mean for those of us managing or trying to understand fast-growing companies? Despite my clickbait title, positive NRR is a good thing! But it's fundamentally similar to higher average customer value, not to a higher growth rate. A company with positive NRR is more valuable and will have more revenue, but the rate at which it grows is still fundamentally governed by its number of customers.

[1] In this example, the initial ACVs of new customers are flat. If the initial ACVs of new customers get x% larger each year, then the long-term growth rate of the company will be 100% + x%. Either way, expansion has no impact on the equilibrium growth rate.
[2] If the expansion rate exceeds the new customer growth rate, then the revenue growth rate will eventually converge to the expansion rate. 
Topics
No items found.
Share

Related blog posts

No items found.
No items found.
No items found.

Start for free

Join the thousands of companies using Fivetran to centralize and transform their data.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.