The CRM Play: finding service agreements in your own records
Your rep has about two hundred accounts in her book. Ask her which ones deserve an agreement conversation this quarter and she'll name six or seven, quickly and with confidence. They'll be the accounts she visited most recently, or the ones where somebody actually picks up the phone.
This isn’t a careless guess - it’s what’s fresh in her mind at this moment.
She's guessing because nobody ever handed her a way to see the whole book at once. That's what commercial service agreement sales looks like in most organizations, a memory exercise with a quota attached.
Every quarter someone asks why attach rate is flat. The answer sits in a database she doesn't have access to and wouldn't know how to query if she did.
70% of contractors lack fast access to warranty and service agreement data (ServiceTitan, 2025)
63% of contractors report that over half their customer base is secured through PMAs (ServiceTitan, 2025)
Median in-service age for a packaged rooftop unit is 16 years, based on surveys of commercial buildings (ASHRAE Service Life Database)
The United States holds roughly 5.9 million commercial buildings totaling 97 billion square feet (EIA, 2020)
Why the residential playbook for service agreements fails in commercial
Advice on landing service agreements is almost always the same.
"Train the technician to ask at the point of service." Why? The customer is relieved the unit is running again, the tech is standing right there, and the moment is warm.
That upselling model was built for a residential HVAC system. It breaks in commercial relationships for three reasons that have nothing to do with how well anyone asks.
The person who signs isn't on site. Your tech is talking to a building engineer or a facilities coordinator who has no signature authority. The property manager or director of facilities who approves the contract may sit in another state and may never meet your technician.
The decision isn't emotional and it isn't fast. Commercial agreements get evaluated against operating budgets and capital plans on an annual cycle.
There is no moment of relief to close into, and the U.S. Department of Energy's own operations and maintenance guidance frames the choice in exactly those terms, as a planned-versus-reactive cost comparison rather than a comfort purchase (DOE FEMP and PNNL).
The work is spread across buildings. One account can carry forty rooftop units across nine sites with three different contacts. Asking at the truck has no mechanism for aggregating any of that.
There’s another layer to consider as well.
An extended service agreement is a risk-transfer product attached to an equipment sale, usually administered by a third party and priced on claims exposure.
A PMA is a recurring labor relationship with a multi-year horizon. When contractor sales content treats them as the same sale, the residential model is leaking in through the vocabulary, and teams end up chasing maintenance contracts with a playbook built for a different transaction.
So the commercial version of this problem isn't a conversation problem.
Commercial service agreement. A recurring contract between a contractor and a building owner or operator covering scheduled maintenance across multiple assets, often across multiple sites. Distinct from a residential membership plan in scope, decision path, and budget cycle.
Preventive maintenance agreement (PMA). The vertical-native term for a scheduled-maintenance contract priced against a defined asset list. Most commercial teams use PMA and service agreement interchangeably, though a service agreement can also include repair coverage a PMA does not.
Sales play. A repeatable opportunity surfaced from operational data by a defined trigger, assigned to a named owner. The trigger is a query. The output is a candidate, not a qualified deal.
What your service history already knows about your customers
Every service call your company has “run” generated a record. Dispatch logged the site. The technician logged the asset and the fix. The parts desk logged what got consumed. Accounting logged what the customer paid.
All of it was captured for operational reasons, and it works fine for those reasons. Nobody ever pointed a sales question at it.
That's the gap. The information that should trigger an agreement conversation is already in the CRM you pay for every month, but it's stranded there until your team uses the right filters to access it.
And this seems to be an industry wide challenge. According to a recent industry poll, seven in ten contractors said they can't get to warranty and service agreement data quickly.
None of this helps a rep unless she can get to it in a few clicks. So the question becomes which list she gets first.
Start with the customers you've served and never signed
Start with the obvious one, and be honest that it's obvious. Customers you've done significant demand work for, with no service agreement in place.
This play is discussed openly in the industry and it's available in commercial field service tooling. If your team isn't running it, that's step one, and the rest of this article assumes you are.
"We have a customer here where we've done a significant amount of demand work, but there's no service agreement in place." - Leila Rookstool, ServiceTitan
Leila Rookstool, Senior Industry Advisor, Commercial at ServiceTitan and a fourth-generation MEP professional with more than fifteen years in the trades, describes this motion as software actively crawling the existing customer base to surface plays a salesperson can take ownership of.
Pull it against three years of history and you'll get a large list.
What converting demand service to a maintenance agreement actually takes
The list gives you an account name and a work history. It doesn't tell you whether the building still belongs to that customer, whether a competitor already holds the agreement, or whether your own team walked away from the account two years ago for a reason someone still remembers.
A surfaced name is a candidate - but not a deal just yet. Qualification is still a rep's job, and it might take them a bit of time to switch from a prospecting for outbound motion to customer re-engagement.
Why this list gets worked once and then abandoned
Sales reps love “net new.” Net new deals come with a world of opportunity. Installs, service agreements, and more. But for the company, recurring revenue drives the long-term growth.
So reps put all of their effort into prospecting new accounts, hoping to land whales, and tend to focus on them until pipeline slows down.
When it does, a sales manager pulls the coverage-gap report, hands two hundred rows to four reps, and asks for activity.
After a shift in the sales motion, reps burn two weeks and results land somewhere south of what that level of effort justified.
So the team concludes that mining service data doesn't work and shifts back to net new.
While the conclusion is incorrect, their experience has told them otherwise - so if the topic gets brought up as a strategy again, there’s team pushback.
The way to solve this is to work your existing book with the same targeting and messaging discipline as your existing prospecting workflows - but add an ABM layer.
The charge line that shows you who is already buying maintenance
Go back to the rep in the opener and her book of business. She couldn't see which of those accounts were already paying, piecemeal, for the exact work a program would cover.
But, that information was in the system the entire time, sitting one level below the one everybody queries.
To find it, segment the list on what the customer has been charged for.
"For any customer that we have effectively charged for a belt or a filter that doesn't have a service agreement in place, I want to surface that as a potential sales play." - Leila Rookstool, ServiceTitan
How charge-level targeting improves service agreement conversion
A customer buying belts and filters from you outside a program has already made the decision the agreement asks them to make. They believe the equipment needs upkeep - and, they're willing to pay for it.
What they haven't done is buy it in a way that generates recurring revenue.
Which means the conversation isn't about persuasion. It's about packaging the solution.
You're not convincing them that maintenance matters - they've been paying for it a line item at a time. You're showing them what it costs to keep buying it that way.
Now, the list is smaller. A book of two hundred accounts might surface a dozen. Twelve accounts with a documented spending pattern beats two hundred with a work history, and the rep can actually work all twelve before the quarter closes.
Rookstool frames this as thinking a layer deeper than agreement status, and how deep you can go depends on how much operational detail your system can reach. Charge-level data is the layer most teams own and never query.
There's one more question under it, and it stops being about agreements entirely.
When equipment age turns into a sales conversation
Building owners already use equipment age to plan capital budgets. Read the same number from the sales side of the table and it's an opportunity signal.
Almost all published equipment-lifespan content is written to the owner, which leaves the contractor's version of that question unclaimed.
The lifespan thresholds that matter
The number that circulates in contractor content is a fifteen-year life for a packaged rooftop unit. ASHRAE's service life database, built from surveys of commercial buildings conducted in the mid-2000s, reports a median in-service age of sixteen years for packaged rooftop units.
Treat that number as the opening of a conversation rather than a deadline.
A unit at sixteen years isn't a unit about to fail. It's a unit whose owner has started doing arithmetic about what comes next, and the contractor who is already in the building has a better seat at that conversation than the one who isn't.
Read alongside warranty expiration timing and install date, equipment age stops being a maintenance-scheduling field and starts looking like a rep re-engagement window.
Turning a maintenance record into a capital conversation
Once repair costs break over the 50% of replacement cost metric, an aging unit with a run of repair calls isn't an agreement conversation - it's a replacement conversation.
And, that’s why reps need to be engaged in actively reviewing this data.
Rookstool has described an end-of-life sales motion in which equipment records would be crawled to surface units past a fifteen or twenty year threshold, with those opportunities routed to salespeople.
Three queries, at three levels of precision:
The query | What it surfaces | Precision |
Demand work performed, no agreement on file | Every account you serve reactively that has never been offered a program | High volume, low precision |
Charged for a specific consumable, no agreement | Accounts already paying for upkeep piecemeal | Lower volume, materially warmer |
Installed asset past expected service life | Replacement and upgrade pursuits | Lowest volume, highest deal value |
Three queries, three lists, one book. Someone has to be able to run them.
How do you actually build these plays?
Four things have to be connected in your existing systems to run this playbook.
Work order history with charge detail.
An asset register with install dates.
Agreement status per asset, not per customer.
And the physical building, because a single customer can span nine addresses and the play only makes sense per site.
None of this requires a large team. Climate Engineering, a thirty-person HVAC contractor in Colorado, does a version of it by hand. They pull current and lost customers onto a map and build quick check-in lists from what they see.
And, if you’re using Convex, a sales manager can run the charge-level list and it comes back tied to buildings and verified contacts - ready for sales to work.
Instead of arriving as a spreadsheet of account names she'd spend a day reconciling against addresses - the sales team spends their time re-engaging contacts they haven’t spoken to in a while or building fresh touch points with new ones.
Timing matters here. This is work for the stretch when the phones slow down and reps start asking what to do with the week, which is exactly when most teams reach for a cold list instead.
What that changes is less about the manager than the rep.
What changes when the list arrives already worked
Back in 2022, Jay Byers, Director of Service Sales Programs at Schneider Electric, described thousands of sites where his company had installed equipment that had since atrophied or gone dormant. The data existed - but it sat in platforms that had been stitched together over the years, and Byers said the lack of connectivity between them was one of the challenges his teams faced.
The detail that lands hardest is what it took to get ready for a customer meeting.
A salesperson who wanted to know the last time his company had serviced a site had to call IT or an administrator who knew that particular system and have the data pulled out. Byers described the experience as difficult and painful for his salespeople. Before that, they found it by brute force, calling whoever used to own the account.
He also named the gap that matters most here. His team might know they had done work at a building without knowing there were twenty units on the roof.
That's the whole problem in one sentence. You can't run a charge-level query or an equipment-age query against a site when nobody can tell you what's installed there.
Byers called technology the only way to reach a goal his company had set of growing services 2x in three years, and the multiplier isn't the data. It's that a rep can see it without asking anyone's permission.
Picture the rep from the opener starting her Monday with twelve named buildings instead of a memory, each one carrying a documented reason she's calling. Her week becomes triage rather than recall.
Her book didn't change. Her visibility into it did.
None of it helps the contractor whose service data is scattered, which is the normal starting point.
Where to start if your service data is a mess
Fragmented systems are the normal condition in commercial services, not the exception. Work orders in one platform, agreements in a spreadsheet, asset lists living in a senior technician's head.
Waiting for a full consolidation project before running any of this is how teams stay stuck for three years. Pick the cleanest data you own and run one query against it. If charge detail is reliable and asset records aren't, run the consumable query and skip the equipment one.
One clean query against partial data beats a perfect query you never run. The cost of bad data is real, and it's also not a reason to postpone targeting until the cleanup finishes.
The starting point is smaller than most teams assume.
What systematic commercial service agreement sales looks like
Your work order history is a prospect list that already passed the hardest qualification test in the business. These people have paid you money.
The teams that get results from it aren't running a better version of the coverage-gap report. They're asking three questions at three resolutions, treating the output as candidates rather than deals, and accepting that the widest list is the least useful one.
Run it when the calendar gives you room. Pipeline doesn't pause for your busy season, and the book you already own is the cheapest place to rebuild it.
Your rep's book hasn't changed. What changes is whether she has to remember it.
Ready to see your book the way your reps need to?
If your service history is sitting in systems that can't talk to each other, that's the normal starting point, not a disqualifier. Schedule a demo and we'll walk through what your own data would surface.
FAQ: Commercial service agreement sales
How do you sell more commercial service agreements?
Start with targeting rather than pitching. Query your service history for accounts with demand work and no agreement, then narrow to accounts charged for specific consumables, then to assets past expected service life.
What data do you need to identify service agreement opportunities?
Four sources have to connect: work order history with charge-level detail, an asset register with install dates, agreement status recorded per asset, and the building each asset sits in. Missing any one of them limits which queries you can run.
How do you convert demand service customers into maintenance agreements?
Lead with their own spending history. A customer who has paid for repeated reactive calls has already funded a maintenance program without the coverage, scheduling, or pricing structure one would give them.
How many agreements should a commercial team expect to close from this?
Conversion benchmarks that circulate for service agreements come out of residential contractor content, where a technician asks at the point of service and the homeowner signs on the spot. Trace them back and most lead to a vendor blog with no study behind them. Measure against your own book, one query at a time, and compare each list to the one above it.
How long do commercial HVAC systems last?
ASHRAE's service life database, built from surveys conducted in the mid-2000s, reports a median in-service age of sixteen years for packaged rooftop units. Treat that as the point where replacement planning starts, not the point where equipment fails.
Why do commercial customers decline service agreements?
Most often because the timing missed a budget cycle, because the scope wasn't priced against their actual asset list, or because the person your team pitched had no authority to approve it.
How often should you contact existing customers about new services?
In a Convex webinar poll, most attendees said they do it only occasionally. Teams working targeted plays tend to move to a quarterly cadence per account, with the trigger determining the reason for the call.
Related reading
Share
