Client Retention and Churn

Churn is the number your multiple is built on

An MSP is valued on the durability of its recurring revenue, and churn is the direct measure of durability. A shop losing 15% of its clients a year has to resell 15% of its base just to stand still; a shop losing 4% compounds. The difference shows up in growth rate, sales cost, and the multiple a buyer will pay. Most small MSPs only compute churn when a broker asks. Compute it monthly.

Logo churn versus revenue churn

Logo churn is the share of clients lost. Monthly: clients who terminated this month divided by clients at the start of the month. Annualize by summing twelve months, not multiplying one by twelve – a single lumpy month distorts the picture.

Revenue churn is the share of MRR lost, in two flavors. Gross revenue churn is MRR lost to cancellations plus downgrades and seat reductions, divided by starting MRR. Net revenue retention (NRR) is starting MRR plus expansion – added seats, upsold services, price increases – minus churn and contraction, divided by starting MRR. NRR above 100% means the existing base grows without a single new logo.

The two diverge. Losing three ten-seat clients is 3% logo churn at a hundred clients and barely dents MRR; one 150-seat anchor is 1% logo churn and can be 15% of revenue. A client who drops from 60 seats to 35 shows up nowhere in logo churn while taking 40% of the contract with them. Pull both numbers from your PSA agreements and accounting each month, and put a name on every dollar lost.

What healthy looks like

Typical as of 2026: top-quartile MSPs hold logo churn under 5% per year, roughly 0.4% per month. The broad middle runs 5–10%; above 10–12% something structural is wrong – wrong-fit clients, weak onboarding, or a failing service desk. On revenue, gross revenue churn under 1% per month is good; NRR of 100–110% is healthy for an MSP that does annual price increases and grows with its clients, and NRR below 95% means the base is shrinking under you regardless of what sales reports. Track clients you fired on purpose separately; a buyer will ask about both.

The leading indicators

Churn is rarely sudden. The notice letter arrives six to twelve months after the relationship actually broke. Watch for:

• Ticket sentiment. Not volume – tone. Per-ticket CSAT dipping, terse replies, "again" in subject lines. Three negative surveys from one client in a quarter is a phone call, not a dashboard entry.
• QBR attendance. The client who reschedules the QBR twice and then sends a junior instead of the decision-maker has already downgraded you from partner to vendor.
• Declining seat counts. Each reduction is either a business in trouble or a business moving work somewhere else. Ask which.
• Payment lateness. A client who paid net-15 for three years and slides to net-45 is either cash-strapped or has decided your invoice matters less.
• A new internal IT hire. The most reliable predictor of all. Sometimes it means co-managed growth; often someone was hired to "evaluate our IT spend." Meet them in their first two weeks and make them successful, or they will make their case by replacing you.
• Silence. No tickets and no calls for 90 days is not a happy client; it is one who has stopped expecting anything from you.

Score every client quarterly; two or more flags puts them on an at-risk list the owner reviews personally.

The first 90 days decide most churn

Clients who leave in year one almost always decided in the first quarter. Onboarding is where sales promises meet delivery, and any gap – a migration that ran long, a ticket that fell through – becomes the story the client tells about you for the rest of the contract. Over-invest here: a named point of contact, a 30-day check-in with the owner, a 60-day ticket-volume review, and a 90-day mini-QBR that shows what changed. The client onboarding process covers the mechanics; the retention point is that the first 90 days are a sales activity.

The retention levers

QBR cadence. The single most effective retention activity is a QBR a business owner finds worth attending: risk register, roadmap, budget, what you did and what it prevented. Quarterly above roughly 25 seats, semi-annual below. The QBR process is worth running even when it feels like overhead; the clients who skip it are the ones who leave.

vCIO. A vCIO function moves you from ticket-taker to advisor. A client with a three-year roadmap built with you cannot leave without abandoning the roadmap.

Documented value reporting. Clients do not see the 200 patches or the backup that restored quietly. Send a one-page monthly or quarterly report that says what you did and what it prevented, in business terms. Invisible work is unvalued work, and unvalued work gets shopped.

Price increases done right. Not raising prices is itself a retention risk, because it erodes margin until you cut service. The clients who leave over an increase are the ones surprised by it. Write annual increases into the MSA at signing, notify 60–90 days ahead with a value summary, apply them to everyone on the same date, and never negotiate individually. Typical annual increases run 5–8% as of 2026, higher when tool costs jump.

When to fire a client

Some churn should be yours. Fire a client who stays unprofitable after a pricing correction, refuses baseline security controls, abuses your staff, or habitually pays late. Run the agreement-profitability report from your KPI review monthly and give the bottom client a choice: new price, or a referral to someone better suited. Give 60–90 days and run the client offboarding process cleanly. Deliberate churn of the wrong clients improves margin, morale, and the churn number that matters.

How churn feeds valuation

Buyers pay for recurring revenue and discount it by how sticky it is. Typical as of 2026: shops under $1M EBITDA trade around 4–7x adjusted EBITDA, larger and better-run ones 8–12x, and the spread is mostly explained by three things a buyer checks in the first week of diligence – recurring revenue percentage, logo churn, and NRR. Sub-5% churn and NRR over 100% support the top of the range; 12% churn and NRR under 95% push you to the bottom or into an earn-out. Client concentration compounds it: a top client over 15–20% of revenue is a discount regardless of how loyal they seem. A year of clean monthly churn data, with a recorded reason for every lost client, is one of the cheapest things you can bring to diligence.

Bottom line

Measure logo and revenue churn monthly from your PSA and accounting, name every dollar lost, and aim for under 5% logo churn and NRR above 100%. Watch the leading indicators – sentiment, QBR attendance, seat counts, payment timing, the new IT hire – and act on two flags. Win the first 90 days, run QBRs that owners attend, report value in writing, raise prices on a schedule, and fire the clients who cost you the rest. Churn is the multiple you will sell for.