
Code Generation That Actually Works: Building Bulletproof Quality Gates with LLM-as-a-Judge

Leonardo Steffen

Gross Revenue Retention (GRR) measures the percentage of recurring revenue you keep from existing customers, excluding any upsells or expansion: how much of your revenue base stays intact without leaks from churn or downgrades. It's your revenue retention floor and can never exceed 100%.
GRR shows what percentage of your recurring revenue from existing customers you retained over a period. You calculate it by starting with your revenue at period start, subtracting churned and downgraded revenue, then dividing by starting revenue. The formula is straightforward: GRR = (Starting MRR - Churned MRR - Contraction MRR) ÷ Starting MRR × 100%. Note that it can never exceed 100% because it excludes expansion revenue by definition.
Why it matters in practice
GRR reveals your core business health by showing how much revenue stays without upsells. For founders, it's critical because investors use it to assess business stability and predictability. A SaaS business with 95% GRR loses only 5% of revenue annually from existing customers through churn and downgrades, while 85% GRR means 15% revenue leakage, a significant difference in unit economics.
GRR has an exponential impact on your Customer Lifetime Value through the formula LTV = (ARPU × Gross Margin) ÷ (1 - GRR). Research shows improving GRR from 90% to 95% literally doubles your LTV because you're halving the denominator. That 5 percentage point improvement can transform unprofitable unit economics into profitable ones.
GRR vs NRR: The key distinction
While GRR focuses purely on retention (what you kept), Net Revenue Retention (NRR) adds expansion revenue (upsells, cross-sells) on top. NRR can exceed 100%, GRR cannot. Here's a simple example: if you start with $100K, lose $10K to churn, but gain $15K in upsells, your GRR is 90% but NRR is 105%.
Both matter, but GRR gives you the honest picture of customer satisfaction without growth masking problems. The gap between your GRR and NRR directly measures expansion effectiveness. According to SaaS Capital's portfolio data, the average gap is approximately 12 percentage points. A narrow gap (under 5%) signals weak upsell and cross-sell capabilities despite good base retention. A wide gap (over 20%) could indicate strong land-and-expand, or dangerous churn being masked by expansion revenue.
What affects your GRR
For technical teams, several factors directly impact revenue retention:
Technical debt and system instability directly impact your revenue retention. Poor product experiences render all other retention efforts ineffective.
Session replay tools, feature analytics, and monitoring user frustration signals (rage clicks, error rates) become essential business intelligence for measuring product quality. Product capabilities must exist before customer success efforts can meaningfully impact retention. Research shows product quality and user experience are foundational determinants. You can't customer-success your way out of a broken product experience.
Benchmarks and what "good" looks like
90%+ is generally considered healthy for B2B SaaS, with enterprise products (high ACV) often hitting 95%+. However, benchmarks vary significantly by business model:
Compare yourself to similar companies, not all SaaS businesses. An SMB-focused founder targeting 95% GRR might be setting an unrealistic goal. An enterprise founder accepting 85% GRR has a fundamental problem.
GRR is not the same as customer retention rate. One customer leaving might be 1% of customers but 10% of revenue if they were a large account. Revenue-weighted metrics like GRR reveal problems that logo retention completely masks.
"Good" GRR doesn't mean you can ignore NRR. You need both retention and expansion for healthy growth. A company with 95% GRR but only 97% NRR has weak expansion capabilities despite strong base retention.
Free trials and freemium users aren't included in GRR calculations until they become paying customers. Many founders accidentally include trial conversions as "retention," which inflates the metric and obscures actual revenue durability.
The biggest misconception? Including expansion revenue in GRR calculations. Some founders believe GRR can exceed 100%, conflating it with NRR. This creates a false sense of security: you might think retention is healthy when you're actually experiencing revenue losses from downgrades and churn that are only being masked by new upsells and expansion revenue.
Your SaaS started January with $500K MRR. During the year, you lost $40K to churn and $10K to downgrades. GRR = ($500K - $50K) ÷ $500K = 90%. This means you retained 90% of your revenue base, or viewed another way, you're losing 10% of revenue annually just from existing customer attrition.
At this rate, you need to acquire new customers generating at least $50K annually just to stay flat. Without expansion revenue from your existing base, you're running up a down escalator: constantly replacing lost revenue before you can grow.
But here's where it gets interesting: that 90% GRR, while decent, represents a significant missed opportunity. If you could improve retention to 95% GRR, you'd only lose $12,500 annually instead of $50,000. That extra $37,500 in retained revenue compounds over time and dramatically improves your unit economics. Every percentage point of GRR improvement flows directly to your bottom line and customer lifetime value.
The choice, as always, is yours to make.

Leonardo Steffen