Source: https://www.revvly.ai/blog/renewal-forecasting-for-saas Retention / NRR 9 min read Jul 19, 2026 # Renewal forecasting for SaaS: a framework you can actually *trust*. For a sales-driven business, the renewal question gets answered by a pipeline. For a subscription business, it usually gets answered by a shrug, or a *gut feeling*. SA Samuel Adeyemo Revenue Systems, Revvly Every subscription business eventually asks the same question in a board meeting or a cash planning session: how much of next quarter's recurring revenue is actually going to renew? For a sales-driven business, that question gets answered by a pipeline. For a subscription business, it usually gets answered by a shrug, a spreadsheet someone updates the night before the meeting, or a customer success lead's gut feeling about a handful of accounts they happen to remember. None of those are a forecast. They're a guess with a deadline. ## Quick answer > Renewal forecasting is the practice of predicting, before the fact, how much of your existing recurring revenue base will renew, expand, or churn in a given period. A credible forecast combines four inputs: billing and payment health (failed charges, downgrades, plan mismatches), product usage trends, contract and term data, and a churn-risk score built from those signals together. Businesses that rely on CRM opinion or a single spreadsheet instead of this combined signal set tend to find out their renewal number was wrong only after the quarter closes. ## Why renewal forecasting breaks in subscription businesses Traditional sales forecasting assumes a discrete event: a deal moves through stages, someone owns it, and it closes or it doesn't. Subscription renewal isn't a deal. In the default case, nothing happens at all. The card gets charged, the subscription continues, and no one on your team ever "wins" the renewal in the way a sales rep wins a new logo. That default-continuation model is exactly what makes it dangerous to forecast the way sales pipelines are forecast. Two problems follow from this. First, cancellation is often quiet. A customer who has stopped using the product doesn't call to complain, they just let the subscription lapse or cancel with two clicks at the end of the term. There's no "at-risk" flag unless someone is actively watching engagement data, not just billing status. Second, the revenue outcome for any given account isn't binary. A renewed account can renew flat, expand into a higher tier, contract into a lower one, or renew for a different quantity of seats or usage than last term. A forecast that only asks "will this renew, yes or no" throws away the information that actually determines whether recurring revenue grows or shrinks. This is also why forecasting gets harder, not easier, as a subscription business scales. ChartMogul's analysis of over 2,500 SaaS businesses found that only 6% of companies with more than 12,000 subscribers reach 100% or higher net revenue retention, down from a much larger share among smaller subscriber bases. The same report found that companies with net revenue retention at or above 100% get more than half their revenue from expansion, while companies in the lowest NRR range get roughly 70% of their growth from new business and only around 15% from expansion. Those are structurally different businesses with structurally different renewal math, and a single flat renewal-rate assumption applied across the whole customer base will be wrong for most of them. ## The four inputs a credible renewal forecast actually needs ### Billing and payment health The most concrete, least subjective signal available is what's already happening in the billing system: failed payment attempts, expired cards, downgrades initiated mid-term, and gaps between contracted seats or usage and what's actually being billed. A subscription in active payment trouble should never be forecast as a clean renewal, but in most stacks this signal lives in the billing tool and never makes it into whatever spreadsheet or CRM field the renewal forecast is actually built from. ### Product usage trends Login frequency, feature adoption, and usage-versus-plan-limit trends are the earliest window into whether a customer is getting value, well before the renewal date arrives. A usage-based or tiered account that has been running well under its plan limit for two consecutive billing cycles is a different forecast risk than one running at or over its limit. This is also where the case for expansion, not just retention, gets made: an account trending toward its usage ceiling is a renewal-plus-expansion opportunity, not just a renewal. ### Contract and term data Term length, auto-renewal status, notice periods, and multi-year versus month-to-month structure all change the shape of risk. A month-to-month account and a two-year contract account behave completely differently even if every other signal looks identical, and a forecast that treats them the same will systematically misjudge timing. ### A churn-risk score built from the above None of the first three inputs is a forecast on its own. A credible renewal forecast weighs them together into a per-account risk score, typically using AI-powered churn prediction that weighs billing health, usage trend, and account context rather than a single rule like "usage dropped, therefore flag it." This is the layer that turns three separate data feeds into one number a founder or finance lead can actually build a quarterly forecast around, and it's also what separates a forecast from a scorecard: a scorecard tells you an account looks risky, a forecast tells you what percentage of next quarter's ARR that risk represents in dollar terms. ## Building the model: three levels of maturity **Level one: cohort-based estimate.** Take your trailing gross renewal rate by cohort (by plan, by signup quarter, by segment) and apply it forward. This is better than nothing and is where most early-stage subscription businesses start, but it assumes every account in a cohort behaves like the average, which becomes less true the more your customer base diversifies. **Level two: weighted risk scoring.** Layer a per-account risk score, built from billing health and usage trend, onto the cohort baseline. Accounts flagged high-risk get pulled out of the "assume it renews" bucket and forecast separately, usually at a lower probability. This is the point where a forecast starts reflecting the actual state of individual accounts instead of a population average. ### Why the gross number and the net number diverge A forecast that stops at "will this account renew" still misses half the picture. Even a business with strong logo retention can post a flat or shrinking revenue number if the accounts that do renew are contracting, and a business with mediocre logo retention can still grow if the accounts that stay are expanding fast enough to cover the ones that leave. 2x Approximate median churn rate of companies below 100% net revenue retention versus companies at or above it, per ChartMogul's analysis of over 2,500 SaaS businesses. ChartMogul's research found that companies below 100% net revenue retention run a median churn rate of roughly double that of companies at or above 100% NRR, but the reverse relationship, how much expansion offsets that churn, is just as decisive for the final number. This is the same reason a bookings number and a recognized-revenue number tell different stories about the same quarter, and it's worth forecasting expansion and contraction as their own line items rather than netting everything into a single renewal percentage. **Level three: real-time, unified forecast.** Billing health, usage trend, and contract data update continuously, in one system, feeding a rolling forecast rather than a point-in-time spreadsheet exercise. This is the level most subscription businesses describe wanting and few actually have, mostly because the underlying data lives in three or four different tools that were never built to talk to each other. ## Why fragmented tools make this harder than it should be Ask most RevOps or finance leads why their renewal forecast keeps missing, and the honest answer usually isn't "we don't know how to build a model." It's that the inputs to the model live in different systems that don't reconcile automatically: billing status in Stripe or a billing tool, usage data in a product analytics tool, contract terms in a CRM or a shared drive full of PDFs, and the actual forecast itself in a spreadsheet someone manually updates before board meetings. Every handoff between those systems is a place where the number can quietly go stale. By the time someone notices the spreadsheet disagrees with what billing shows, the quarter is usually already closed. This is also why a forecast built on CRM opinion alone tends to run optimistic. A CRM records what a customer success manager believes about an account relationship. It doesn't automatically know that a customer's card failed twice last week or that usage dropped 40% after a champion left the company, unless someone manually notices and updates a field. A forecast that's only as good as its slowest, most manual data source will always lag the accounts that are actually moving. ## What good looks like A renewal forecast worth trusting shows, for the current period and at least one period ahead: expected gross renewal dollars by cohort, expansion and contraction dollars forecast separately (not netted into one number), accounts flagged by risk score with the specific signal driving the flag, and a rolling accuracy check comparing last period's forecast to what actually happened. If your current process can't answer "which of last quarter's flagged accounts actually churned, and which of the ones we assumed would renew didn't," the forecast isn't being used as a forecast, it's being used as a snapshot. Questions worth asking before your next board meeting or planning cycle: - Does our forecast separate gross renewal, expansion, and contraction, or is everything netted into one number? - Is the risk score behind the forecast built from billing and usage data, or from a CS rep's memory of the account? - When was the underlying data last reconciled across billing, usage, and contract systems? - Could we explain, account by account, why a specific customer was flagged as at-risk? - Did last quarter's forecast match what actually happened, and do we know by how much it missed? ## The throughline: forecasting is a data problem before it's a modeling problem None of the scoring logic above matters if the underlying data is scattered and stale. Most subscription businesses don't have a forecasting-accuracy problem so much as a data-unification problem: billing lives in one tool, usage in another, contracts in a third, and the "forecast" is really just whoever last reconciled all three by hand. A [real-time view of subscription health](https://www.revvly.ai/blog/mrr-gap) that ties billing, usage, and churn-risk signals into one system is what turns renewal forecasting from a quarterly fire drill into a number people actually check before it's due. That's the same underlying gap that makes [net revenue retention hard to compound](https://www.revvly.ai/blog/nrr-compounds) at scale and the same reason a [fragmented revenue stack](https://www.revvly.ai/blog/revenue-stack) keeps showing up as the root cause behind late surprises, whether the symptom is a missed forecast, an unexplained churn spike, or a [churn signal](https://www.revvly.ai/blog/churn-product-signals) nobody noticed until the account was already gone. ## Managing renewal risk without a unified view of the data The businesses that get renewal forecasting right aren't the ones with the most sophisticated statistical model. They're the ones where billing status, usage trend, and contract data live somewhere close enough together that a risk score can actually be built from all three, updated continuously, instead of reconstructed by hand once a quarter. Getting there is less about buying a forecasting tool and more about closing the gap between the systems that hold the underlying signals in the first place. ## Frequently asked questions ### What is renewal forecasting? Renewal forecasting is predicting, ahead of the renewal date, how much of a subscription business's recurring revenue will renew, expand, or contract for a given cohort or account. Unlike sales pipeline forecasting, it has to account for renewal being the default outcome rather than an event someone actively closes. ### How is renewal forecasting different from sales forecasting? Sales forecasting tracks discrete deals moving through stages toward a close date. Renewal forecasting tracks existing subscriptions where the default outcome is continuation, cancellation is often silent, and the revenue result can be flat, expanded, or contracted rather than simply won or lost. ### What data do you need to forecast renewals accurately? At minimum, billing and payment health, product usage trends, and contract or term data. A churn-risk score that weighs all three together, rather than a single flag like usage dropped, produces a materially more accurate forecast than any one input alone. ### Can a spreadsheet or CRM alone handle renewal forecasting? A CRM can hold contract and relationship data, but it doesn't natively see billing failures or usage decline unless someone manually enters that information. A spreadsheet-based forecast is only as current as its last manual update, which is why forecasts built this way tend to be optimistic and go stale between updates. ### Why does forecast accuracy get harder as a company scales? As subscriber count grows, the customer base diversifies and a single flat renewal-rate assumption fits fewer accounts. ChartMogul's research found net revenue retention becomes harder to sustain at scale, with only 6% of companies above 12,000 subscribers reaching 100% or higher NRR. ### What's the difference between forecasting renewal rate and forecasting net revenue retention? Renewal forecasting usually focuses on whether individual accounts continue. Net revenue retention forecasting nets renewal, expansion, and contraction together into a single revenue outcome, so a business can have strong account-level renewal and still show weak NRR if the accounts that renew are contracting. ### Does automating renewal forecasting replace the need for a customer success team? No. A data-driven forecast tells a team which accounts and how much revenue is at risk and why; it doesn't replace the relationship work of acting on that information. It changes where a CS or RevOps team spends attention, from manually reviewing every account to focusing on the ones the data flags as genuinely at risk. SA Samuel Adeyemo Revenue Systems, Revvly Samuel builds the systems that turn billing events into trustworthy revenue numbers. He writes about retention, pricing, and the operational details that decide whether the metrics hold up. Revvly Orchestrate the flow of revenue. © 2026 AcceleratorApp Limited. All rights reserved.