Most commercial HVAC shops still get referral work. Fewer realize the jobs getting sent no longer match the shop they are building.
Key Takeaway: A referral engine asks at the right moment, tracks who sent what, and stops treating "we get most of our work word of mouth" as a strategy.
A lot of your new work came in the way it always has: through somebody who knew somebody.
ACHR News reported that word of mouth accounted for roughly 71% of contractor business volume the prior year. That is real volume.
Still, what shows up in your own pipeline tells the real story. The buildings are getting smaller. The scope of work is getting thinner. Two of your top property managers retired last year and the FMs who replaced them have never met you.
The work still books with "referral" in the source field. None of it looks like the jobs you originally built the shop to win. Nobody noticed at first. The cash from the strong years was still landing.
Reality
You are still getting referral work. The problem is the jobs coming in no longer look like the jobs you built the shop to win.
That is what a referral habit looks like when it ages. A thank-you note from history. Yet, next quarter's pipeline still has to be built. The referrals you got last year are already on the books. The question is what you are doing to make sure a new property manager picks up the phone.
Many shops often do nothing. They wait. They hope. They tell themselves that good work creates referrals. Sometimes it does. More often, the property manager who would refer you forgets, or sends you to a friend who runs strip retail when you do hospitals, or hands you a price shopper who burns 90 minutes of your estimator's time on a job you never wanted.
A referral engine is built on purpose. It asks at the moment of value. It tracks who sent what. It turns good referral relationships into repeatable strong-fit work. It is built so that when you are not in the room, somebody on your team is still working it.
Here is how to build one.
The first lie to kill is that referrals are free.
Reality
Wrong-fit referrals feel free because they arrive through relationships. They are not free. They consume estimator time, site visits, proposal effort, and job-mix clarity.
A wrong-fit referral costs you the same things a wrong-fit paid lead costs you. An hour of your senior estimator on a building type you do not service. A site visit on a Friday that pushes a real proposal into Monday. A polite conversation with a property manager you cannot afford to insult and cannot afford to bid. A C-job that books because saying no to a referrer felt rude.
Marcus runs a $9M commercial shop in two metros. Service, retrofits, light controls. Last quarter he pulled his referral data for the first time. Six referral leads in 30 days. Four were off-profile: one residential, one outside his service area, one tenant fit-out under 5,000 square feet, one strip retail job with three weeks of free estimating buried in the ask. Two were close to fit. One booked at $42,000 GP.
The four off-profile leads cost his estimator 9 hours of phone time, 2 site visits, and a drafted proposal that died on review. Nobody put a number on those hours because they came from "a referral." The hours died anyway.
Paid leads at least come with a price tag attached. A referral comes with a relationship attached, which is harder to walk away from.
Fix It
Measure referral quality the same way you measure paid lead quality: booked strong-fit work, not lead count.
Free referrals are expensive when they fill your board with the wrong work. The discipline that protects you from paid-marketing waste is the same one that protects you here: measure every source by booked strong-fit work, not by lead count.
Pro Move
A sloppy referral stream does not just waste estimator time. It teaches the market to send you the wrong work.
The deeper problem starts when those wrong-fit referrals book. The market learns. The property manager who sent the strip retail job tells two colleagues you handled it well. Now you are the strip retail shop. Six months later, half your inbound referral flow looks like the work you took to be polite.
The shops that get this right do two things. They define what a strong-fit referral looks like before they ask for one. Then they make it easy for the referrer to send the right kind, not just any kind.
A referral engine has three inputs. If any one is missing, the whole system sputters.
Reality
Nobody owns the ask. The owner assumes the project lead asked. The project lead assumes the service manager asked. Nobody asked.
The single biggest reason shops do not get referrals is simple. Nobody asks. Six months later, the property manager has forgotten the moment when she would have said yes without thinking about it.
Fix It
Ask at the moment of value, not at invoicing.
The right time to ask is the moment of value. The invoice is too late. It is the day after you finished a clean rooftop swap and the property manager called your project lead to say the building came up faster than expected. It is two weeks after you closed a multi-site service agreement and the FM has just gotten her first month of clean reports. It is the week after a chiller you replaced got through a heat wave without a service call.
Those are the moments most shops let pass.
Pro Move
Use one sentence. No card. No swag. No automation.
"If you know another property manager dealing with the same problem we just solved, would you introduce us?"
Send it from the person who did the work, not from you, and not from an automated text message.
Reality
Without a profile, referrals drift toward whatever someone happens to need.
You cannot ask everyone for everything. Build a one-line referral profile that mirrors the A-job definition you are already using on your job mix. If you have not done that work yet, your fit problem is upstream of your referral problem. Read the job mix audit framework first.
Fix It
Use a one-line referral profile that tells your team who to ask and tells the referrer who to send
Sample profile for Marcus's shop: "Commercial property managers, 50,000 sq ft and up, in our two core counties, with a multi-site portfolio or one large building over 200,000 square feet."
That sentence does two jobs. It tells your team who to ask. It tells the referrer who to send.
Pro Move
Write the profile so clearly that it filters out the work you do not want before it ever reaches your estimator.
Most owners skip this and end up with a referral pipeline full of small residential jobs from their plumber's customer list.
Reality
If you cannot report referral source, fit, and booked outcome, you do not have a referral engine.
If you cannot answer "how many strong-fit referrals did we get last quarter, from whom, and how many booked," you do not have a referral engine. You're winging it.
Fix It
Add the fields and run the report monthly.
The minimum viable tracker is one CRM field on every new lead: Source = Referral, with a sub-field for who sent it. From there: Strong-fit yes or no. Booked yes or no. Gross margin if booked. Run that report monthly. Pay attention to the names that keep showing up.
Pro Move
Find the few people driving most of the work and protect those relationships on purpose.
Most shops have 3 to 7 referrers driving 80% of their referred work. They almost never know which ones. Find them. Those are the relationships you protect.
The referrals coming out of your shop matter as much as the ones coming in.
Every shop already does this informally. When a property manager calls and says she needs an electrician, a plumber, a roofer, a controls integrator, a general, somebody on your team gives her a name. Maybe you. Maybe your service manager. Maybe whoever picked up the phone. Those recommendations are leaving your shop every week.
That is already a pipeline. Most shops do not run it that way. They treat outbound recommendations as a courtesy instead of as infrastructure.
Pro Move
Treat outbound recommendations as infrastructure, not as a favor.
Once you treat it as infrastructure, the effect compounds. The plumber whose work you have recommended to four buildings over two years remembers you when his commercial customer asks who handles their rooftops. The controls integrator you put in front of a tenant fit-out project last fall sends you the next chiller swap.
The mechanism is precision. Vague outbound recommendations create vague return flow. "I think I know a plumber, let me dig up his card" produces nothing. A clean send, by name, with one sentence about why, produces remembered relationships. Trade partners cannot reciprocate something they cannot trace back to you.
The discipline is simple. Pick five trades you do not compete with. Vet them once. When the question comes in, give the name in one sentence. Track who you sent. Then send a one-line note to the trade partner the week after.
"Sent ABC Property Group your way for their plumbing. They are good people."
That note costs you 30 seconds. It is the highest-yield 30 seconds in your week.
Quick Win
You do not need a platform. You need four weeks, two CRM fields, and one meeting where referral activity gets reviewed.
Week 1: Define the Profile and the Ask
Sit with your sales lead and write the A-customer profile in one sentence. Write the ask in one sentence. Then pick the trigger event for the ask: completion of a strong-fit job, anniversary of a service agreement renewal, post-incident win on a covered building.
Week 2: Add the CRM Fields and Pull the Historical List
Add two new fields on your lead record: Source = Referral and Referred By. Then pull the last 18 months of leads and tag them. You will not catch all of them. You do not need to. Catch enough to see the pattern. Identify your top 7 referrers by booked job volume.
Week 3: Build the Contact Loop
Your top 7 referrers each get a real phone call from the most senior person they have ever interacted with at your shop. Not from you, unless they only know you. From their actual relationship.
The script is short.
"Wanted to thank you for sending [X] our way. We finished that job in [outcome]. If you know other [profile match] dealing with [the problem we solved], we would handle them the same way."
That is the whole script. No reward offer. No swag. The asset is your competence and the relationship.
Week 4: Set the cadence
Pick the meeting where referral activity gets reviewed. Most shops bolt it onto their existing weekly sales meeting. Track three numbers: referrals received this week, referrals booked this month, top three active referrers.
If those numbers do not move for two months in a row, the trigger event is wrong or the ask is not happening.
Run that for 90 days before you make any structural changes. The data will tell you whether you have a fit problem, an ask problem, or a follow-through problem. If the issue is fit, diagnose it against the fit-vs-follow-through frame before you change anything else. If the issue is follow-through, look at what happens in the first 60 seconds of an inbound call.
This is the moment most owners flinch. A referrer you cannot afford to lose sends you a wrong-fit job. Small. Off-profile. Awkward scope.
Reality
Relationship pressure makes a bad job harder to decline than it should be.
You have three choices. One of them is wrong.
The wrong choice is to take it, lose money on it, resent the referrer, and quietly stop returning the next call from that property manager. That is what most shops do. The referrer notices the cooling. The pipeline dies.
Fix It
Do not stay silent. Either take it as a deliberate relationship trade or decline it cleanly and route it to someone you trust.
The right choices are cleaner. Take it as a relationship trade, billed honestly, with the work scoped down to what the customer actually needs and the conversation framed as a favor that will be remembered. Or decline it cleanly and refer it to someone you trust who handles that work, with a thank-you to the referrer for thinking of you.
Both protect the relationship. Both protect your shop. The second one also activates the reciprocity engine.
Pro Move
Use the off-profile referral to clarify what on-profile actually looks like.
The mistake is staying silent. The referrer cannot read your mind. If she sent you something off-profile, it is because nobody told her what on-profile looks like. That is upstream from her. That is on you.
Quick Win
Track one number quarterly: percent of new strong-fit booked work that came in through referral.
Not total leads. Not total jobs. Booked strong-fit work, by source.
If that number is climbing quarter over quarter, your engine is real. If it is flat or falling while your overall referral count looks fine, your shop is being referred to the wrong work and the system is selecting for the wrong customers.
Either result is useful. A flat one tells you the ask is going to the wrong people, the profile was never clearly communicated, or your top referrers have moved on. A rising one tells you the system is doing what a system is supposed to do: producing predictable strong-fit pipeline that does not depend on you being in the room.
You keep getting referrals. Some of them are great. Most are random. Your top 7 referrers eventually retire, sell their portfolios, or get replaced by FMs who never met you. The pipeline that felt steady for a decade quietly thins out. You do not notice for a while because the booked jobs from the strong years are still cashing.
By the time you notice, the referrers are gone and the relationship capital cannot be rebuilt in 90 days.
A referral engine is the cheapest pipeline insurance you will ever buy. It costs you a CRM field, a four-week build, and a 30-second note to a trade partner once a week. The shops that do it have predictable strong-fit work coming in 18 months from now. The shops that do not are running on the relationships somebody built 10 years ago.
You know which one you are running.
Here are the top toolkits HVAC owners are reading, using, and sharing to work smarter every week.
Fix the first 60 seconds with two-lane intake, four-question qualification, and booked next steps so good leads stop dying in your own phone system.
Stop stuffing your calendar with C-work. Run the A/B/C job audit to protect crew time and margin from the jobs that quietly drain both.
Audit every channel by what it costs to book real work, not cheap leads. Finally see what earns its spot and cut the rest.
Stop chasing vanity clicks. Track what actually converts to booked jobs and margin so you know what's working and what's just noise.
No spam. Get real advice and proven tactics to win more jobs and keep more profit.
“This is the stuff no one tells you. We’re making more with less stress.”
— Louis, Texas
