NRR Research: Churn vs. Expansion Revenue

Steven Forth is a Managing Partner at Ibbaka. Connect on LinkedIn

The standard advice on improving NRR is to ensure that you are near your natural level of churn (assumed to be low) before you work on optimizing other things. The logic here is that one has to keep customers before one can expand revenue from customers. It is hard to grow in a shrinking pond.

Ibbaka still stands by that guidance, in principle, but the generative AI disruption is making us willing to consider an exception for new and disruptive segments.

On Thursday Oct. 24 Ibbaka will release the Peakspan Ibbaka Net Revenue Retention Report for 2024. We are finalizing our data analysis for this and we are seeing a very interesting trend among many companies that are bringing in generative AI applications.

  • High churn

  • High expansion revenue

  • High Net Revenue Retention

Here is an example.

From the Peakspan Ibbaka NRR Net Revenue Retention Report for 2024

This is a generative AI company with churn of 20% and expansion revenue of 110%. The result is Net Revenue Retention of 190%.

It has a community led growth motion (like Hugging Face but this is not Hugging Face), sells multiple independent modules to B2B and B2G and uses per user, per model and per site pricing metrics.

I think most of us would agree that 20% churn is high, most companies with churn this high would have red lights flashing. We would also agree that Net Revenue Retention of 190% is excellent, rivaling Snowflake in its glory days. This is driven by expansion revenue of 110%. The customers being retained are growing use and the amount they are paying very rapidly.

What is happening here?

Based on our conversations with companies bringing generative AI to market this is what we are hearing.

This pattern is typical of new sectors that are growing rapidly but where use cases are still fluid and both users and vendors are looking for how to get the most value. Customers that cannot find enough value churn out quickly while those customers that hit on differentiated and valuable applications expand their use. If there are usage or consumption pricing models in play this can lead to rapid growth in expansion revenue. This is true at the company level, some companies find the valuable use cases and grow quickly, and at the customer level, some customers get value and expand use while others churn out.

What is the relationship between NRR performance and churn?

The Peakspan Ibbaka survey gathered data from 503 companies across a wide range of SaaS verticals.

We asked “what is the relationship between growth and churn?” We graphed this to explore this question a bit further. Which is more closely correlated with NRR performance, low churn or high expansion revenue?

For all 503 companies the Pearson correlation coefficients were as follows

  • NRR Performance and Churn 0.069

  • NRR Performance and Expansion Revenue 0.971

Now there is some confounding going on here. If one has high churn then one has fewer customers with whom you can expand revenue. Managing churn, keeping customers, is still a priority. But controlling churn is not enough. You have to be able to grow inside existing customers.

Here is a graph of Rev. Growth by Churn (NRR = Current Revenue - Churned Revenue + Expansion Revenue). The company at the top of the chart has NRR of 120% (100-10+30), the company at the far right has NRR of 95% (100-40+35).

From the Peakspan Ibbaka NRR Net Revenue Retention Report for 2024

Lessons learned about churn and expansion revenue

  1. The importance of churn depends on the maturity of the market
    Mature Market > Minimize ChurnEarly
    Market > Optimize for revenue growth

  2. Design to enable in account growth
    Include one or both of the following types of pricing metric
    Usage or Consumption Metric > The greater the use the higher the payment
    Success or Outcome Metric (Preferred) > The customer only pays for successful outcomes (however they are achieved)

  3. Adapt pricing to enable emergent use cases
    Do not assume that you know all of the use cases or even the main use case in an emergent market, make sure your packaging and pricing allows for exploration

  4. Avoid per user metrics
    The more people that use the application the more likely they are to discover ways to create value

  5. Keep your best customers let others churn
    It is not worth keeping every customer especially in early stage markets like generative AI applications; your pricing and packaging should let customers who are not getting value churn out while expanding revenues from those who are getting value - this requires careful pricing and packaging design

Ibbaka can show you how to package and price innovative applications to optimize long-term revenue growth and profitability.

Previous
Previous

Value - The Foundation for Customer Engagement

Next
Next

Generative pricing for AI is a blend of dynamic pricing and value based pricing