B2B churn: how to measure and reduce customer loss

B2B churn is the loss of customers and revenue from the base, and most companies discover the problem on the day the customer asks to cancel, when the customer’s decision is already months old. In the operations I diagnose, many of those cancellations started in the sale: a customer outside the ideal profile, a promise post-sales had no way to keep, or a handoff sent by email without the context of the discovery. To reduce churn, leadership needs to measure both losses, customers and revenue, log in the CRM the risk signals that appear before the notice, run a review of the at-risk portfolio with a plan for each account, and train the sales and Customer Success teams to act on those signals. A retention tool helps once the team knows what to watch and what to do with what it sees.

When a director calls me about churn, I ask for the list of customers who cancelled in the last twelve months and, for each one, three pieces of information: the recorded reason, the date of the first risk signal and the salesperson who closed the contract. In most companies the reason comes as free text, such as “price” or “other”, the date of the first signal does not exist and the salesperson has already left the company. Across more than 200 companies, I have seen churn drop when leadership started treating the cancellation as the end of a sequence of facts the CRM could have shown.

What is churn and how do you calculate the churn rate?

Churn is the share of customers or revenue a company loses in a period, calculated on the base that existed at the start of that period. In B2B the same word covers different measures, and the leadership meeting goes nowhere when CS talks about number of customers, finance talks about revenue and nobody agreed on the formula beforehand. Customers who joined during the period stay out of the calculation, because they were not part of the starting base. The table uses a hypothetical example: a company starts the quarter with 100 customers and R$ 500k in monthly recurring revenue, loses 6 customers worth R$ 25k, suffers R$ 10k in scope reductions on other accounts and sells R$ 40k of expansion to customers who stayed.

Metric	What it measures	Formula	Hypothetical example
Customer churn	How many customers in the base left	Customers lost in the period ÷ customers at the start	6 ÷ 100 = 6%
Gross revenue churn	How much recurring revenue left through cancellation and scope reduction	(Cancelled revenue + reduction) ÷ revenue at the start	(25 + 10) ÷ 500 = 7%
GRR, gross revenue retention	How much of the starting revenue stayed, excluding expansion	(Starting revenue − cancelled − reduction) ÷ starting revenue	465 ÷ 500 = 93%
NRR, net revenue retention	How much the base yields once expansion is added	(Starting revenue − cancelled − reduction + expansion) ÷ starting revenue	505 ÷ 500 = 101%

In the example, lost revenue (7%) exceeds the share of lost customers (6%) because of scope reductions, which a customer count cannot see: the customer who cuts half the contract stays on the list as active. The 101% NRR gives leadership the impression of a healthy base, while the expansion team had to sell R$ 40k to replace R$ 35k that left. I ask leadership to track the four lines together, every month, with the formula written in the team’s guide.

GRR or NRR: which one should leadership track?

Both, because each answers a different question and has a different owner. GRR never exceeds 100% and shows the operation’s ability to keep what it sold, which depends on CS, onboarding and the quality of the sale. NRR adds expansion and shows how much the base grows without new customers. In a McKinsey analysis from 2025 of B2B subscription software companies, those in the top quartile of valuation reached an NRR of 113%, against 98% in the bottom quartile. I apply the same reading to recurring service companies: a base that grows on its own reduces the pressure on the new business funnel.

Also calculate churn by entry cohort, by segment and by originating salesperson. The company average hides the pattern sales leadership cares about, and the cohort breakdown shows whether customers who joined in a tight-quota quarter leave earlier than the others. How to organize these indicators in a matrix with owner and frequency is covered in RevOps metrics.

Where churn starts: causes by origin

The reason a customer gives at cancellation is usually the last one in line. “Price” comes up when the customer no longer sees enough value to defend the contract inside their own company, and that value stopped showing months earlier. I classify causes by origin, because each origin has a different owner and a different signal to log.

Origin	How the cause shows up	What to log in the CRM	Who acts first
Sale	Customer outside the ICP, promise beyond scope, decision maker absent from the purchase	ICP fit, commercial commitments, contacts with defined roles	Sales leadership
Onboarding	First result delayed, kickoff without the sponsor	Expected and actual date of first value	Implementation and CS
Usage	Few active users, contracted features nobody uses	Monthly usage indicator per account	CS
Relationship	Sponsor left, contact limited to the operational team	Date of the last meeting with the decision maker, contact changes	CS and account owner
Price	Price increase applied without a value review, lower offer from a competitor	Date and percentage of the increase, discount requests	CS with support from sales
External factors	Acquisition, change of strategy, crisis in the customer’s industry	Event logged on the company record, with date	CS leadership and executive team

Why does so much churn start in the sale?

A salesperson with a new business quota closes whatever customer shows up, and without a written ideal customer profile nobody can tell them that contract will turn out expensive. The customer too small for the product and the operation without the structure to use the solution enter the base with an expiration date. How to define that profile and use it in qualification is covered in ICP, the ideal customer profile.

The promise is the second origin: the shortened deadline to beat a competitor and the integration that “can be done” reach post-sales as a customer complaint in the first month. The third is a handoff without the stated problem and without the success criteria, which forces CS to ask again what the salesperson already knew. The minimum content of that handoff and where each field lives in the CRM are in B2B post-sales.

External factors need to stay separate in the analysis, so leadership does not mix the churn the company controls with the churn it does not. I get suspicious when that category goes beyond a small share of the reasons, because it tends to become the destination of every poorly explained cancellation. In the exit interview, ask the decision maker for the concrete fact, such as the date of the acquisition or the announced budget cut.

Which risk signals should you log in the CRM?

A risk signal is an observable fact, with a date, that precedes the cancellation request. The analyst’s opinion about the customer’s mood goes in as a note, and the score relies on facts anyone on the team can verify in the CRM. I recommend starting with usage drop, critical ticket, change of decision maker, late payment and low NPS, adding the lack of contact with the decision maker, which shows up in most of the cancellations I analyze.

Signal	Logging criterion	Points
Usage drop	Monthly usage 30% below the average of the previous three months	3
Change of decision maker or sponsor	Departure communicated or decision maker contact disassociated from the company	3
Critical ticket	High-priority ticket open for more than 10 days or reopened	2
Late payment	Invoice overdue by more than 15 days	2
Low NPS	Score from 0 to 6 in the last survey, from any contact at the account	2
No contact with decision maker	No meeting logged with decision maker or sponsor in 90 days	1

The sum of points sets the account’s band: 0 to 2, healthy; 3 to 5, watch; 6 or more, at risk. The criteria and weights in the table are a starting point, and each company adjusts the numbers to its own usage and billing cycle. What must stay fixed is the logging rule: the analyst records the signal with a date on the day they observe it, and the score changes without waiting for the monthly meeting.

How to build a simple health score

Start with the six signals in the table and test the score against the past before using it to make decisions. Apply the criteria to what the CRM and finance recorded three and six months before each cancellation in the last year, and see how many of those accounts would have shown up as watch or at risk. Do the same with a sample of customers who renewed, and adjust weights and thresholds until the score separates the two groups.

How to reduce churn: the CS leadership rituals

CS leadership runs three rituals to act in time, all with the CRM on screen, and a parallel spreadsheet portfolio cancels out the effort, as I discuss in CRM adoption.

Weekly review of the at-risk portfolio

The meeting lasts 45 minutes and covers only accounts on watch and at risk, plus renewals in the next 120 days. Each account owner presents the signal that put the account on the list, the likely cause by origin, the next step and the agreed date. The leader asks for the logged fact and the date of the last contact with the decision maker, and an account that came onto the list without a plan leaves the meeting with one.

Action plan per account

Every account that reaches the at-risk band gets a plan logged in the CRM, with the cause by origin, the recovery goal, actions with owner and date, the people in the company who need to get involved and the criterion for the account to leave the risk band. If the cause started in the sale, the salesperson who closed the contract takes part in the plan. If the contract weighs on revenue, the executive team joins the conversation with the customer’s decision maker. The plan ends when the account returns to the watch band for two consecutive reviews or when the customer formalizes the exit.

Exit interview with a closed-option reason

When a customer asks to cancel, someone from leadership, preferably outside the account, talks to the decision maker within two weeks. The central question is when they started considering leaving and what happened at that moment. The reason goes into a closed-option field in the CRM, with options tied to the origins: expected result not achieved, low user adoption, change of decision maker, price or budget, move to a competitor, scope sold different from scope delivered, and change of strategy, acquisition or closure of the customer. A second field holds the detail as free text, and the “other” option requires a justification.

With the reason as a closed option, leadership adds up cancellations by origin every month and shows the sales manager how many came from scope promised beyond what was delivered, a count free text does not allow. The monthly ritual among the three leadership teams, where this number goes on the agenda, is described in SLA between marketing, sales and post-sales.

What sales, marketing and CS do to reduce churn

Sales leadership

The sales manager starts looking at each salesperson’s cohort churn, in addition to the closing quota. The ICP fit check becomes a qualification criterion, with a required field on the opportunity, and the commercial commitments field goes through the manager’s review before the deal moves to the won stage. Part of the salesperson’s evaluation can consider how many of the customers they closed made it past the first year.

Marketing leadership

Cancellation reasons show marketing which message attracted the wrong customer. If customers who leave early came from a price campaign or from content that promised fast results, marketing adjusts the message before generating more leads with the same profile, and segments communication to the base by health score band.

Customer Success: from support to portfolio management

Many CS teams came from support. Managing a portfolio and renewals calls for other skills: reading the account’s signals, running a value review with a director, negotiating a price increase without giving up the first discount and building a recovery plan with owner and date. In the same analysis, McKinsey found NRR around seven percentage points higher at companies with the most mature value realization and adoption practices, a group that made up 18% of respondents. Those practices depend on people prepared for the commercial conversation.

This is the point where training changes the churn number. The analyst who learns to run a value review books the meeting with the decision maker before the price increase, and the leader who learns to run the portfolio review asks for logged facts. Without that preparation, the company configures the score in a retention tool and the team keeps treating renewal as one more ticket.

A 90-day roadmap for leadership to reduce churn

1. Write the formulas for customer churn, gross revenue churn, GRR and NRR in the team’s guide and calculate the four numbers for the last four quarters.
2. Build the list of cancellations from the last twelve months with reason, entry cohort, originating salesperson and date of first signal, reclassifying reasons by the origins in the table.
3. Create in the CRM the closed-option reason field, the fields for the six risk signals and the ICP fit field on the opportunity.
4. Apply the health score to the past, adjust the weights and run the first assessment of the whole portfolio.
5. Start the weekly review of the at-risk portfolio and require a logged plan for every account in the at-risk band.
6. Bring the sum of cancellations by origin to the monthly ritual with sales and marketing and agree on an action for the origin that weighed most.
7. At the end of the quarter, compare GRR, NRR and churn by origin with the baseline from step 1.

Steps 5 and 6 depend on behavior, and they are where most retention initiatives stall after the second month.

Training to reduce churn: what HR and L&D need to assess

How to recognize the need

HR usually receives the request as “train CS in customer service”. Before accepting that framing, ask CS leadership whether the team knows how to calculate the GRR of its own portfolio, whether cancellations have a closed-option reason and whether the portfolio review happens with a plan per account. Three negative answers, combined with analysts who came from support and now carry a renewal target, point to a need for training in portfolio management and commercial conversations.

Who takes part

CS analysts and managers, the onboarding team and sales leadership take part, because some of the causes start in the sale, and marketing joins for the reading of churn reasons. Leaders go through their own cohort, and the team cohorts work with the company’s real accounts in the CRM.

How to measure

Take the baseline before the training: GRR, NRR, churn by origin, share of accounts with a health score updated in the month and share of at-risk accounts with a logged plan. In the first weeks, track behavior, such as signals logged with a date and the weekly review held in the CRM. GRR and NRR respond after a renewal cycle, and HR agrees on that timeline with the executive team before presenting results.

If your leadership wants to reduce churn by treating the cause at its origin, with sales, onboarding and CS working on the same signals, that is the work of the in-company training on CRM adoption and customer base management. I run the training on the CRM and the portfolio your company already has, design the health score, the portfolio review and the exit interview with the leadership team, and prepare CS for the renewal conversation. The company leaves with the CRM best practices guide and the portfolio and customer base management manual, written for your operation.

Frequently asked questions

What is churn in B2B companies?

Churn is the loss of customers or recurring revenue from the base in a period, calculated on the base that existed at its start. In B2B, the company tracks customer churn and revenue churn, which includes cancellations and scope reductions, and completes the picture with GRR and NRR.

What is the difference between customer churn and revenue churn?

Customer churn counts how many customers left, and revenue churn measures how much recurring money left, adding cancellations and scope reductions. The two numbers diverge when the company loses large accounts or when customers cut part of the contract without cancelling, and leadership needs to look at both.

How do you reduce churn in B2B companies?

Start with last year’s list of cancellations, classified by origin, to learn where the loss starts. Then log risk signals in the CRM with a date, build a simple health score, run the weekly review of the at-risk portfolio with a plan per account and bring exit reasons to sales and marketing every month.

How do you calculate GRR and NRR?

GRR divides starting revenue minus cancellations and reductions by starting revenue, and it never exceeds 100%. NRR uses the same calculation and adds the expansion sold to the base in the period. New customers stay out of both, because they were not part of the starting base.

Which signals indicate a customer is going to cancel?

The most frequent signals are usage drop, change of decision maker or sponsor, unresolved critical ticket, late payment, low NPS and lack of contact with the decision maker. They only help if the team logs each one in the CRM with a date, on the day it is observed.

How do you train the Customer Success team to reduce churn?

Train CS in portfolio management and commercial conversations: reading signals, running value reviews with the decision maker, negotiating renewals and price increases and building a recovery plan per account. Training works best with the company’s real accounts in the CRM, with CS and sales leadership in the same program and with the portfolio review starting the following week.