Membership math: lifetime value, churn and the renewal curve most shops never draw

Most owners can tell you how many maintenance members they have. Far fewer can tell you what one member is worth, how long the average one stays, or how big the program will be in three years if nothing changes. Those three numbers come from the same place: the renewal curve. This post builds membership lifetime value from that curve, step by step, with a worked model you can copy into a spreadsheet tonight.

The short version: the blended renewal rate on your dashboard is probably hiding a weak first year, failed cards are churn you never decided on, and your program has a ceiling you can calculate today.

What a maintenance member is actually worth

Start with what a member contributes in a year, not what they pay. A member pays dues, but they also cost you two visits, a repair discount and some admin, and they buy repairs at a higher rate than non-members. The number that matters is contribution: dues minus the cost to deliver the plan, plus the gross profit on the work members buy.

Our post on memberships that renew prices a two-visit HVAC plan from scratch. It lands at about $332 a year to deliver against $420 of dues, so roughly $88 of plan margin. Say a member also spends $600 a year on repairs at a 50% gross margin after the member discount. That adds $300. Contribution: about $388 a year. These are example numbers. Replace each with yours; the HVAC pricing guide shows how to get the loaded labor cost underneath them.

Then lifetime value is contribution times how long the member stays, with later years discounted because a dollar in year six is worth less to you than a dollar today:

With a constant retention rate r, margin m and discount rate i, this collapses to a tidy shortcut that marketing professors Sunil Gupta and Donald Lehmann call the margin multiple: LTV = m × r ÷ (1 + i − r)¹. Their worked example, a brokerage keeping 95% of customers a year at a 12% discount rate, gives a multiple of 5.59. (Their version counts margin from the first renewal onward. If you collect the first year's contribution up front, as a membership does, add one more m.) The shortcut has one problem for a service shop: r is not constant. That's the whole point of the next section.

The renewal curve: why year one decides everything

If you look at a single cohort of members, say everyone who joined in spring 2023, you'll almost always see the same shape. A big drop at the first renewal, a smaller drop at the second, and then a slow, steady leak.

There's a good reason for it, and it isn't that members grow to love you over time (though some do). Peter Fader and Bruce Hardie showed that when a group of customers has a mix of underlying renewal probabilities, the group's renewal rate rises year after year even if no individual changes at all². The people who were always likely to leave do so early. Whoever's left is, by selection, the loyal end of the mix. Their paper's warning is blunt: fitting a single retention rate to early data will understate how long the survivors stay, and fitting it to a mature book will overstate how long new members stay.

For a maintenance plan, the "likely to leave" group is easy to picture: the homeowner who joined to get a discount on one repair, the one who never got their fall visit booked, the one who sold the house.

Here is a hypothetical cohort of 100 new members with a 70% first renewal, 80% second, and 85% every year after:

Add up the "still active" column all the way out and you get about 543 member-years per 100 members sold, so the average membership lasts about 5.4 years in this example. At $388 a year of contribution and a 10% discount rate, the discounted lifetime value comes to about $1,425 per member sold.

The blended rate is lying to you, politely

Now picture the whole book in a steady state, built from a few years of selling 300 new members a year under that same curve. In any given year, about 300 first-year members come up for renewal, about 210 second-year members, and about 1,120 members in year three or later. Count everyone who renewed and divide by everyone whose term ended:

An 82% renewal rate looks healthy on a dashboard. Meanwhile three in ten new members are leaving at their first renewal, and that's the group your techs worked hardest to sell. The blended number is dominated by veterans who would renew if you did almost nothing.

Split renewal rate by tenure. At a minimum, report first renewal separately from everything else. If you only fix one report this year, fix that one.

The ceiling on your program: members on the books = sales × lifetime

Here's the formula that surprises owners the most. In operations research it's called Little's Law: the average number of items in a system equals the rate they arrive times the average time each one stays. John Little proved it in 1961 for queues, and it holds for almost any stable system with arrivals and departures³. A membership book is one.

Sell 300 a year with a 5.4-year average life and the book settles around 1,630 members. Not 3,000 in ten years. About 1,630, and it gets there more slowly each year as it approaches the ceiling. If the owner's plan says "2,500 members by 2029," the math says you need either about 460 new members a year at today's renewal curve, or a curve that keeps members about 8.3 years. Those are very different projects, and it's worth knowing which one you're signing up for.

This is also why "we added 300 members and the program only grew by 90" is not a mystery. At 1,300 members and an 82% blended renewal rate, about 234 members leave every year. Sales have to outrun that leak before the count moves.

What one point of first-year renewal is worth

Because every later year is multiplied through the first renewal, a better first year is worth more than the same gain later on. Using the example curve and the 300-a-year sales pace:

Two lessons from this table. First, ten points at the first renewal is worth almost three times five points at the second. Second, the long-run rate looks like the biggest lever on paper, but it's the one you control least, for a reason that has nothing to do with service quality.

The churn you can't fix: people move

In the Census Bureau's 2023 Current Population Survey, about 9.3 million of the 226 million people living in owner-occupied homes had moved in the previous year, roughly 4.1%. For people in rented homes it was about 16%⁴. Every member who sells the house is a cancellation, no matter how good your visits were.

Two practical moves follow. Sell memberships to the home, not just the person: when a member sells, ask for the buyer's contact and offer the remaining term as a transfer. And be careful selling plans to renters or landlords' tenants, whose move rate is about four times higher. A landlord who owns ten rentals is a better member than any one tenant.

Card failures: the churn nobody decided on

On an autopay plan, a member can leave without ever deciding to. A card expires, gets replaced after a fraud alert, or hits its limit on the day you charge it. If nobody fixes it, the membership lapses.

The exposure depends on how often you charge. A monthly plan runs a card 12 times a year. An annual plan runs it once. Even a small failure rate per charge compounds:

These are worked examples, not benchmarks. But look at the second row: if just 1 in 100 monthly charges fails and never gets fixed, you lose more than a tenth of your monthly members each year to nothing but cards. Compare that with the 4% you lose to moves.

What do outside numbers say? There's no independent benchmark for maintenance plans, but subscription billing companies publish their own network data, and it points the same way. Stripe reported in 2024 that a quarter of lapsed subscriptions on its platform lapse purely because of payment failures, not a decision to cancel⁵. Recurly's network data for July 2026 puts involuntary churn at about a third of all monthly churn (1.25% of 3.60%)⁶. Both are vendor figures from software subscriptions, so treat them as direction, not as your number.

Some of the fix is plumbing. Visa and Mastercard both run account updater services that pass new card numbers and expiry dates to merchants who bill stored cards⁷. Whether you get those updates depends on your payment processor and on the card issuer taking part. The rest of the fix is process.

A recovery sequence that respects the member

• Retry on a schedule, not immediately. Many declines are temporary (insufficient funds, a bank's fraud check). Spread retries over about two weeks.
• Text a secure update-card link with every retry. The member fixes it in a minute without calling you.
• Call on day 10 or so. A CSR call recovers members a text won't: the ones who changed banks or are annoyed about something else.
• Hold a grace period before you cancel. Keep the member's visits and discount live for a couple of weeks while you recover. Then mark the membership lapsed, with the reason "payment," so it shows up separately from decisions.
• Count unrecovered failures as non-renewals. Leaving them in "pending" flatters the renewal rate for months.

TradesWork retries a failed membership charge 1, 3, 7 and 14 days after the first failure, with a text and an update-card link each time, and charges that never clear land in a Needs attention list for the office. Card-number refreshes from the networks' updater services happen at the payment processor: Stripe, which TradesWork uses, says it attempts to update saved cards automatically where the issuer takes part⁸. Not every issuer does, so a call from the office is still part of the sequence.

Monthly or annual billing? Run both curves

Monthly billing lowers the price at the kitchen table, which helps conversion. Annual billing has one card event a year, and the renewal is a single moment you can prepare for. Each fails differently:

The honest answer: track voluntary and involuntary churn separately for each billing type, then decide. If most of your monthly churn is payments, fix recovery before changing plans. If most of your annual churn is members saying no at renewal, the year-end value summary in memberships that renew is the fix. Renewal notice rules, including California's 15-to-45-day window for annual plans, are covered in the maintenance plans guide.

The membership numbers to put on one page

The renewal rate and churn definitions in our glossary match the math above. Use the same definition every month, or the trend means nothing.

What to do this week

1. 1

   Pull every membership ever sold

   Start date, end or cancel date, cancel reason, billing type. Most systems can export this. If yours can't tell a card failure from a decision, start tagging reasons now.

2. 2

   Build cohorts by start year

   For each year's cohort, count how many reached their first, second and third renewal. That's your curve. Don't average across cohorts until you've looked at each one.

3. 3

   Compute your average life and ceiling

   Sum the survival shares to get average years. Multiply by your yearly sales pace. Compare the ceiling with your plan.

4. 4

   Find the first-renewal leak

   Pull the first-year members who didn't renew. Did they get both visits? Was it a card? Did they move? Fix the biggest bucket first.

None of this needs software you don't have. It needs a clean export and an hour with a spreadsheet. The payoff is knowing, with numbers, whether your next dollar should go to selling more members or keeping the ones you sold last spring.