HVAC Maintenance Plans: How to Build Recurring Revenue That Actually Makes Money

The plan fee is the least interesting revenue in the program. A 400-agreement base at $200 a year is $80,000 — real money, but not why you do this. You do it because a plan member is a customer who has already agreed to let you inside their house twice a year, and who calls you first when something breaks instead of typing "AC repair near me" into Google. Shops that track it usually find plan members generate two to three times the annual revenue of a comparable non-member once repairs, accessories, and replacements are counted.

The bigger lever is equipment replacement. A tech standing in front of a 16-year-old condenser twice a year, documenting declining capacity with photos, is the cheapest replacement lead source in your business — no ad spend, no lead fee, and a customer who already knows the tech's name. In shops with a mature agreement base, a substantial share of replacement sales come out of the member list rather than from new leads, which is exactly why the members are worth more than the fee suggests.

And it smooths the calendar. HVAC demand is brutally seasonal — the phone rings off the hook in July and goes quiet in October. A tune-up backlog is work you can move into the shoulder months to keep techs billable when there are no emergency calls. That alone is often worth more than the plan fee, because the alternative is paying wages against an empty board.

Finally, it is the one thing a buyer will pay a premium for. When a shop sells, the agreement base is scrutinized line by line — active count, renewal rate, whether the agreements are transferable, and whether unused visits are a liability sitting on the books. A large base with a weak renewal rate gets discounted hard. A smaller base renewing above 85% on auto-pay does not. Even if you never sell, that same distinction is what tells you whether the program is real.

Residential plans in most markets land between $150 and $300 per year for a single system — roughly $13 to $25 a month if you bill monthly, which you should. Two visits per year is the standard: one pre-cooling in spring, one pre-heating in fall. Additional systems in the same house typically add $80 to $150 each, since the drive time is already paid for and the second unit is nearly pure margin.

The inclusions that cost you little and matter a lot to the customer are the ones to lead with: priority scheduling ahead of non-members during a heat wave, no overtime or after-hours diagnostic fee, a standard filter at each visit, and a repair discount. The inclusion that quietly destroys margin is an aggressive repair discount — 20% off parts and labor on a shop doing meaningful repair volume with members can hand back more than the plan fee. Ten to fifteen percent is plenty; customers rarely buy on the discount percentage anyway. They buy on priority and on not thinking about it.

Tier if it fits your market, but keep it to two or three. A basic plan at the low end with the two visits and priority, a mid tier adding the repair discount and a waived diagnostic, and a premium tier with a labor warranty or a replacement credit that accrues. What you are really doing with tiers is giving the customer a reason to choose rather than a reason to decide yes or no — most shops see the middle tier take the majority of sales, which is the point.

Bill monthly on a card, not annually by invoice. The monthly price reads smaller, the cash is smoother, and — the real reason — a card on file makes renewal automatic instead of a phone call somebody has to make. Shops that switch from annual invoicing to monthly auto-pay routinely pick up double-digit points of retention without changing anything else about the program.

Here is the calculation that decides whether your program makes money, and it takes four minutes. Take your fully loaded technician cost per hour — wage, payroll taxes, benefits, truck, insurance — which in most residential shops is somewhere between $45 and $75. Multiply by the real time a maintenance visit takes including drive: usually 1.25 to 1.75 hours, not the 45 minutes on your task list. That is $56 to $130 per visit, times two visits, plus $10 to $25 in filters and materials. Call it $120 to $285 a year in delivery cost.

Now compare that to a $179 plan. On the low end of cost you are fine; on the high end you are underwater before you have discounted a single repair. This is why so many programs feel like they are not working — they are not. The fee is not covering delivery, and the shop is relying on pull-through revenue it never actually measured to make up the gap.

There are only four ways to fix it, and you need at least two. Raise the price toward the $229 to $279 range, which the market bears far better than most owners believe. Cut real visit time with a tight, checklist-driven maintenance procedure so the visit is 60 to 75 minutes instead of drifting to two hours. Route by geography so members in the same neighborhood are batched into one day instead of scattered across the week — this is where the dispatch time actually goes. And measure pull-through honestly: tag every repair and replacement that originates from a maintenance visit so you know what a member is truly worth.

One accounting note worth getting right early: money collected for visits not yet delivered is deferred revenue, not profit. Shops that book the full annual fee as income in January and then deliver the fall visit in October are borrowing from themselves — it feels great in Q1 and hurts in Q4. Your bookkeeper can set this up in an afternoon, and it keeps you from making a hiring decision on money you have not earned yet.

Maintenance agreements are sold at the end of a repair or an install, by the technician, standing in front of the equipment, while the customer is still grateful. That is the moment. Not a postcard, not an email campaign, not the CSR on the phone. A trained tech with a simple offer converts a meaningful share of repair calls — shops that do this well land somewhere in the 25% to 40% range on eligible calls, and shops that do not run in the single digits with the same customer base.

The script is short and it is not a pitch: "I got you running today. The reason this failed is the coil was loaded up — that is what the twice-a-year visit catches. It is $22 a month, you jump the line in July, and you get 15% off repairs. Want me to set it up before I go?" That is the whole thing. It works because it connects the plan to the failure the customer just paid for, which is the only version of the argument that is actually persuasive.

Pay for it, but pay for the right thing. A $25 to $50 spiff per agreement is standard and it works. What does not work is paying only on the sale — that gets you agreements sold to people who cancel in month three. Pay a portion on the sale and a portion at the first renewal, and the incentive suddenly points at selling plans to customers who should have one. Track conversion per technician; the spread between your best and worst tech on this single metric is usually enormous and entirely coachable, which is exactly the kind of gap AI call and job review tools are built to surface.

Install is the other high-conversion moment, and most shops fumble it. A customer who just spent $9,000 on a new system will say yes to protecting it far more often than a customer three years into ownership. Bundle the first year into every replacement quote by default — not as an upsell line the salesperson can skip, but as part of the standard proposal. Then the renewal conversation in year two is with someone who already has the habit.

Every maintenance program looks great in year one. Year two is where you learn whether you built an asset or a list. The metric is renewal rate, and the benchmarks are unforgiving: below 70% you are filling a leaking bucket and your growth is entirely dependent on how many new agreements the techs sell this month. Between 70% and 85% is a working program. Above 85% — which is achievable, and is where auto-pay shops live — the base compounds on its own and every new agreement is genuinely additive.

The single biggest driver of that number is not service quality. It is whether the renewal requires a decision. An annual invoice forces the customer to re-buy, and a meaningful share simply do not get around to it. A monthly card on file renews silently. That mechanism alone commonly moves retention by fifteen to twenty-five points, which dwarfs anything you will accomplish by improving the visit itself.

The second driver is whether you actually delivered the visits. Unused visits are the quiet killer — the customer paid for two, got one because your fall schedule collapsed under emergency calls, and cancels in the spring. If your undelivered-visit count is climbing, your renewal rate is already falling and you will not see it in the numbers for another six months. Track scheduled-versus-delivered every month, not once a year.

And run a real save process on cancellations. A card decline is not a cancellation — it is an expired card, and it accounts for a surprising share of what shops record as churn. Automated retry plus a text, then a call, recovers most of them. For genuine cancellations, ask why and log it. Three quarters of the reasons will be "we sold the house" or "we replaced the system," and neither of those means your program has a problem — but you cannot know that unless someone is writing it down.

Here is the operational trap. You sell 400 agreements. That is 800 visits a year, and customers want them in the two windows where the weather changes — so 60% to 70% of the demand compresses into roughly eight weeks in spring and eight in fall. Meanwhile those same weeks are when your emergency call volume starts climbing. This is the point where most programs either burn out the office or start silently not delivering visits.

Push the season wider than the customer thinks it is. Start calling for spring tune-ups in February, not April, and for fall in August. Members who book in the shoulder are the easiest to schedule, and every one of them you pull out of the peak is a slot you did not have to fight for. Offer the off-peak slot as the default in your outreach rather than asking an open-ended question about timing.

Then automate the outreach itself, because doing this by hand is a full-time job you are not going to staff. Database marketing tools like Arch mine your own customer list for who is due, who is overdue, and whose equipment age makes them a replacement candidate, then run the campaign automatically. Your field service platform should be generating the visit list — FieldEdge has genuinely strong agreement management, and ServiceTitan, Housecall Pro, and Workiz all handle recurring visit generation and renewal tracking well enough that there is no reason to run this out of a spreadsheet.

Protect the phones during the rush, because the tune-up scheduling calls and the no-cooling emergencies arrive on the same line in the same week. This is where an AI receptionist earns its cost in a maintenance program specifically: routine "I need to book my spring tune-up" calls get handled and booked into the CRM without touching your CSR, while true emergencies route to a human. Numa handles this with native ServiceTitan, Housecall Pro, and Jobber integration at $249 to $379 a month for a typical shop; Rosie covers a solo or two-truck operation from $49 to $199. Either way the goal is the same — your dispatcher spends peak week on the jobs that pay, not on rescheduling filter changes.

Last, close the loop on the visit. A completed tune-up is the single best moment to ask for a Google review — the customer is happy, nothing is broken, and there is no invoice shock. An automated review request after every maintenance visit turns 800 visits a year into a steady review flow that no amount of ad spend buys.

First 30 days, do the math and set the price. Calculate your loaded labor cost and your real visit duration by timing ten actual maintenance calls door to door. Set the annual price at a number that covers delivery with margin — for most shops that is north of $200 — and build two or three tiers with monthly auto-pay as the default and only payment method you promote. Write the maintenance checklist so the visit is repeatable in 60 to 75 minutes.

Days 30 to 60, build the selling motion. Train every tech on the twenty-second offer, tied to the failure the customer just experienced. Set the spiff to pay half on sale and half at first renewal. Add the first year of the plan to every replacement proposal as a standard line. Configure the agreement in your field service platform so visits generate automatically and renewals are tracked — if you cannot pull an active-count and renewal-rate report on demand, you are not done.

Days 60 to 90, turn on the automation and start measuring. Set up the outreach campaign for due and overdue members, the automated review request after each maintenance visit, and the card-decline retry sequence. Then track five numbers monthly and nothing else: active agreements, agreements sold this month, renewal rate, visits scheduled versus delivered, and revenue pulled through from member visits. Those five will tell you within two quarters whether you built an asset or a list — and unlike almost anything else in this business, you can fix each one directly.