Yet, many teams still reduce churn rate to a single percentage that doesn’t capture the different types of churn. This makes it difficult to identify what really needs fixing.<\/p>\n
That’s what this article aims to solve. I’ll show you how to calculate churn correctly, benchmark it against 2026 standards, and diagnose the specific churn patterns<\/a> hiding behind the headline number. The standard customer churn rate formula: divide the number of customers lost during a period by the number of customers you had at the start, then multiply by 100. For example, if you began the month with 500 customers and lost 15, your monthly churn rate (15\/500 x 100) is 3%.<\/p>\n When<\/strong>? Track this monthly and annually. Monthly figures help you spot seasonal trends and the impact of product or pricing changes<\/em>, while annual figures provide a clearer view of long-term retention performance<\/em>.<\/p>\n<\/div>\n However, customer churn only tells you how many accounts left, not how much revenue left with them. That’s where revenue churn becomes useful. If you lose ten SMB customers paying $200 per month but retain a single enterprise customer paying $20,000 per month, customer churn looks alarming, while revenue churn looks healthy. Revenue churn measures the percentage of recurring revenue lost from existing customers during a period<\/em>. This way, you understand the true business impact of churn across different segments.<\/p>\n And if your company is growing quickly, there’s another adjustment worth making. The standard formula assumes your customer base remains relatively stable throughout the period. But when you’re adding large numbers of new customers, it can overstate churn. For example, if you start the month with 300 customers (W), add 200 more (closing at 500; X), and lose 30 (Y), the basic formula reports a 10% churn rate. Using the average customer count during the period instead produces a more accurate figure of 7.5%.<\/p>\n Formula: <\/strong>Divide churned customers (Y) by the average of your opening (W) and closing (X) customer counts, then multiply by 100. Written out: (Y \/ [(W+X)\/2]) x 100. When?<\/strong> As a rule of thumb, consider switching to the adjusted formula once new customer additions regularly exceed 10% of your opening customer count.<\/p>\n<\/div>\n According to Culta and Recurly:<\/p>\n Median monthly churn for B2B SaaS is at 3.5%: 2.6% represents voluntary cancellations. And 0.8\u20130.9% is for involuntary churn (failed payments and billing issues).<\/p>\n<\/li>\n Contract type creates variation: Annual prepaid contracts are often around 0.5\u20132% monthly churn. Month-to-month products are 3\u20138%.<\/p>\n<\/li>\n The weight of the contracts also contributes: Enterprise SaaS companies with large contracts have < 1% monthly churn. For startups with < $1M ARR, it is 5\u20137% monthly churn.<\/p>\n<\/li>\n Industry is also a factor: At 1.8% annual churn, infrastructure SaaS has the lowest churn. But EdTech is higher, with up to 10% annual churn.<\/p>\n<\/li>\n<\/ul>\n Unlike typical SaaS, AI-native products use a different retention curve. ChartMogul research reveals that AI-native companies reported 40% gross revenue retention (GRR) and 48% net revenue retention (NRR). That is in comparison to the B2B SaaS median NRR of roughly 82%.<\/p>\n Kyle Poyar, Analyst-in-Residence at ChartMogul, calls it the “AI tourist effect.” It happens when users sign up out of curiosity and, like tourists, leave before activating because there is no recurring use case.<\/p>\n The same report reveals that AI-native median GRR increased from 27% in January 2025 to 40% by September 2025. This means committed customers are replacing the initial wave of AI tourists gradually.<\/p>\n My take?<\/strong> Use the churn benchmarks based on your context. This is what I mean: a 3% monthly churn rate for an enterprise vendor selling annual contracts is low. At the same time, it is good for a self-serve SaaS product. So, don’t ask, “Is my churn good?”, ask, “How does my churn compare to products like mine?”<\/p>\n I am now seeing three distinct patterns that produce the same churn rate number but require completely different responses. Treating them as one problem is why most retention playbooks fail to move the aggregate number.<\/p>\n AI tourists are users who signed up only to test your product. So, they have no job to be done, and hence no reason to stay. And because of that, they are often found in the first 30-60 days (especially in low-price or free-tier cohorts). They also account for 23% GRR of AI tools’ budgets.<\/p>\n The standard churn formula captures these AI tourists as churned customers. However, the problem here is about the acquisition cohort (qualification before free trial). And even worse, better onboarding will not fix the problem. The product was never the issue.<\/p>\n This happens when accounts with AI agents access your product through MCP integrations. They look retained, but the human in the loop has not even logged in since Q3. And I think these are the worst.<\/p>\n I have experienced this firsthand with one of our customers at Userpilot. The customer had lots of logins with decent activity, but didn’t renew. And when I looked closely, there was indeed activity but no renewal, despite zero tickets. Thankfully, I got the customer back on track by contacting the executive stakeholder.<\/p>\n When I discussed this with Yazan Sehwail, our CEO. He calls it an infrastructure problem in the agentic area.<\/p>\n As teams start deploying their own AI agents, those agents are gonna tap on our existing infrastructure that will be powering all of the usage and all the product data, and that’s extremely powerful.<\/p>\n<\/blockquote>\n Jason Lemkin, founder of SaaStr, experienced this with a breakout AI company at over $100M ARR. Their customers were happy. The product was working. Every deal, including the large ones, came back as a one-year commitment: “We’ll sign. For one year. To start.”<\/p>\n So, he argues<\/a> that AI agent switching costs are structurally lower than typical SaaS because prompts are portable. As such, smart buyers are not committing to three years because they don’t want to explore.<\/p>\n And when I look at the math, it compounds. A company at $100M ARR dropping from 92% to 82% GRR faces $10M in additional annual bookings needed just to stay even. No retention campaign fixes this.<\/p>\n I use these signals to identify and fix the dominant churn type in a given cohort.<\/p>\n AI tourist churn has a consistent fingerprint: high first-30-day churn, shallow feature adoption, and a strong correlation with low-friction acquisition channels such as self-serve trials.<\/p>\n The key distinction is whether users ever reached a meaningful activation milestone. A user who activated a core feature and later disengaged is an onboarding or adoption problem. A user who signed up, explored, and left without reaching value is usually an acquisition problem.<\/p>\n The fix starts at the top of the funnel. Tighten targeting for high-churn acquisition channels and use welcome surveys to route users into onboarding paths based on their actual use case. Track churn by acquisition source alongside overall churn so you can identify which channels are attracting curiosity rather than genuine demand.<\/p>\n Ghost accounts look healthy at first glance. Product events continue firing, APIs remain active, and usage appears stable. The warning sign is that human engagement gradually disappears.<\/p>\n To identify these accounts, compare product activity against human session counts. If workflows are still running but the people behind the account haven’t logged in for 60 days, you’re likely looking at a pre-churn account rather than a healthy one.<\/p>\n The intervention window is typically 60 to 90 days before renewal. At that stage, re-engagement campaigns, targeted in-app prompts, behavioral emails, and executive stakeholder outreach can still change the outcome. Seven days before renewal, you’re running a save attempt. Ninety days before renewal, you’re still influencing adoption.<\/p>\n One lesson I’ve repeatedly seen at Userpilot is that successful saves often come from conversations with executive sponsors rather than day-to-day users. The person using the product isn’t always the person approving the renewal.<\/p>\n Structural churn happens when customers receive value but can leave easily because little ties them to your platform. This challenge is especially relevant for AI-native SaaS, where switching costs are often low, and competitors can replicate features quickly. As Jason Lemkin argued, your AI model is rarely the moat. Integrations, workflows, proprietary data, and customer relationships are.<\/p>\n The fix is to build stickiness beyond the product’s core feature set. Deep integrations, embedded workflows, outcome reporting, and customer success relationships all increase the cost of switching. Some companies also experiment with outcome-based pricing, tying renewal decisions directly to measurable results rather than seats or usage.<\/p>\n Most retention conversations focus on voluntary cancellations: customers who actively decide to leave. That’s understandable since voluntary churn is the larger number (at 2.6%) and usually signals a problem with activation, adoption, pricing, or product value.<\/p>\n But involuntary churn deserves attention for a different reason: it’s often the easiest churn to fix. While only 0.8\u20130.9%, it represents customers who often wanted to stay but were removed because the payment process failed.<\/p>\n The effort is also minimal. Cutting voluntary churn usually means improving onboarding, increasing adoption, refining pricing, or strengthening customer success processes. Recovering involuntary churn often requires better payment recovery workflows, proactive card-expiry reminders, and smarter retry logic.<\/p>\n That’s why involuntary churn is one of the fastest retention wins available. A customer who left because their card expired doesn’t need to be convinced of your product’s value; they already saw enough value to remain a customer.<\/p>\n Modern dunning systems take advantage of this. AI-driven payment recovery tools optimize when and how payment retries occur, while proactive billing reminders help prevent failures before they happen. Together, these systems can recover revenue that would otherwise be counted as churn.<\/p>\n To understand why customers leave, how much damage the churn caused, and whether you could have spotted it earlier, I recommend these supporting metrics:<\/p>\n The first is net revenue retention (NRR):<\/strong> While churn rate measures customer loss, NRR shows whether expansion revenue from existing customers is offsetting those losses. A company with 10% gross churn and 120% NRR is in a healthier position than one with 5% gross churn and 95% NRR because existing customers are growing faster than others are leaving. As a rule of thumb, focus on gross churn during the early stages of growth, then shift attention to NRR once expansion revenue becomes a meaningful part of your business.<\/p>\n<\/li>\n Next is cohort retention analysis<\/a>: <\/strong>This acts as an early warning system. Aggregate churn is a lagging metric. On the other hand, cohort analysis tracks retention by signup month, helping you identify whether newer cohorts are retaining better or worse than previous ones. If month-three retention is declining for recent cohorts, you have a retention problem long before it appears in your overall churn rate.<\/p>\n<\/li>\n Finally, track human engagement alongside product activity: <\/strong>This has become particularly important as AI agents and automated workflows generate more product activity. An account may appear healthy because API events are still firing, while the humans behind the account have gradually stopped using the product. Monitoring human sessions separately from automated activity helps surface these “ghost accounts” months before they show up as churn.<\/p>\n<\/li>\n<\/ul>\n Combined with churn rate, these metrics provide a much more complete view of customer health.<\/p>\n A 4% monthly churn rate in 2026 could mean three things. First, you are losing AI tourists at the top of your acquisition funnel. Second, humans disengage from accounts, but they are still technically active. Third, rational buyers are not committing to multi-year terms in a market moving this fast.<\/p>\n So, a single churn rate can’t help. The teams I have seen get results in 2026 now look at churn by cohort, acquisition channel, human session, and contract type. And they run tests before deploying any strategy.<\/p>\n This is where Userpilot can help. You get behavioral analytics and cohort views. Even better, Lia (Userpilot AI) lets you run diagnostics and act on the problems without switching tools.<\/p>\n Book a demo<\/a>, and I’ll walk you through how it works! Customer churn is the rate at which customers discontinue their subscription or stop using a particular product or service. This article shows you various ways to calculate churn, make sense of the data, and create proactive solutions to retain more customers.<\/p>\n","protected":false},"author":105,"featured_media":641066,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"content-type":"","inline_featured_image":false,"footnotes":""},"categories":[7561],"tags":[5088,533,505,834,1035,285,86,245],"class_list":["post-115725","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-churn-retention","tag-business-growth","tag-churn-rate","tag-churn-surveys","tag-customer-churn-rate","tag-customer-loyalty","tag-customer-retention","tag-customer-satisfaction","tag-user-retention"],"acf":[],"yoast_head":"\n![\"demo](\"https:\/\/blog-static.userpilot.com\/blog\/wp-content\/uploads\/2026\/06\/CTA-blog-banner-1-1.png\")
<\/a><\/p>\nWhat does the customer churn rate formula actually measure, and which version should you use?<\/h2>\n
What are the updated SaaS churn benchmarks for 2026?<\/h2>\n
![\"customer](\"https:\/\/blog-static.userpilot.com\/blog\/wp-content\/uploads\/2026\/06\/churn-benchmarks-list.png\")
\n
Why is a single churn number increasingly misleading in 2026?<\/h2>\n
AI tourist churn: The wrong users arrived and left on schedule<\/h3>\n
Ghost-account churn: Humans disengaged while the account stayed active<\/h3>\n
\n
Structural contract churn: Rational buyers refusing multi-year commitments<\/h3>\n
How to figure out which type of churn you’re dealing with (and fix it)<\/h2>\n
![\"Identifying](\"https:\/\/blog-static.userpilot.com\/blog\/wp-content\/uploads\/2026\/06\/churn-taxonomy-infographic-1.png\")
<\/p>\nAI tourist churn: Fix acquisition, not onboarding<\/h3>\n
Ghost-account churn: Fix disengagement before renewal<\/h3>\n
Structural churn: Fix switching costs and product stickiness<\/h3>\n
Why do most SaaS teams underestimate involuntary churn, and what is the quick win?<\/h2>\n
Which metrics should you track alongside churn rate to get the full picture?<\/h2>\n
\n
![\"Lia's](\"https:\/\/blog-static.userpilot.com\/blog\/wp-content\/uploads\/2026\/05\/userpilot-agent-lia-proactive-analytics-view.png\")
Stop treating churn rate as one number<\/h2>\n
<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"