Insulation leads have no single price. One vendor prints $54.99 per exclusive organic lead on a $399-a-month package, and $76 to $262 per Google Ads lead from its own clients' data; most sellers print nothing. AI cannot lower those prices, but an AI script that checks property type, access and existing insulation can stop wasted survey visits.
The figure to track is cost per job won, with cost per booked survey as a stepping stone to it. A screen that is too strict can cost more than it saves, because every wrongly declined lead is a lost job. The sums below show where the line sits.
What insulation leads cost, source by source
The four lead sellers that turn up for this search term are mostly silent about money. One prints a per-lead price for organic leads, a Google Ads range calculated from its own clients and package prices from $99 a month; the other three print no prices, and one of those mentions a "flat monthly rate" without an amount. Treat everything in the table as a vendor figure, because none of it is independent data.
| Source | What is printed | What to ask or check |
|---|---|---|
| Exclusive organic leads from a lead vendor | $54.99 a lead, flat, "exclusive, never shared", for customers on its Growth package, which its pricing page lists at $399 a month plus a $399 setup fee; organic leads are free on its $1,299-a-month tier | Its page counts a phone call or form submission as a lead and says there is no contract. Add the package fee to the lead price before you compare sources |
| Google Ads leads bought through a vendor | $76 to $262 a lead, the 10th to 90th percentile of what its own clients paid each month over the last twelve full months; on the Growth package a 15% management fee on ad spend comes on top | The vendor divides ad spend by vetted exclusive leads after removing duplicate and spam calls, so ask how your own account would be counted |
| Shared marketplace leads (Angi, HomeAdvisor, Thumbtack) | No published insulation price on their public pages | Log a month of what you actually pay per lead, and remember that other firms may be contacting the same customer |
| Bark | Credit cost shown before you respond; this page quotes no credit prices | See how Bark charges for leads and how to model cost per job |
| Google Local Services Ads | You pay for valid leads, such as calls and messages, not clicks. Google is moving these ads into Performance Max in phases from August 2026 to 2027, without a maximum cost per lead | Categories vary by area, so check Google's eligibility tool for insulation, and download your old reports, which do not carry over |
| Marketing agencies on a flat monthly fee | One says "flat monthly rate" and "No setup fees!" with no amount; the per-lead vendor above starts its packages at $99 a month plus a $249 setup fee | The monthly fee, the ad budget on top, the contract length, and who owns the leads and the ad account |
| Your own enquiries (website form, Google Business Profile, referrals) | No per-lead fee; the cost is your time | Count them separately so they do not flatter the paid sources |
A few cautions apply to every row. The $54.99 and the $76 to $262 range come from a company that sells exclusive leads, and its own shared-against-exclusive calculator argues the case with example inputs of $25 for a shared lead and $75 for an exclusive one, which are its inputs, not market data. One insulation agency's case study claims a 2700% return on investment and a 40% rise in gross revenue with no method behind them, so I have not used either. And the same lead vendor's roofing page prints $135 to $404 for a Google Ads lead, so the figure moves a long way by trade, and a range across all its clients hides the spread between markets and seasons.
The package fee changes which tier is cheaper, too. On monthly fees alone, $399 plus $54.99 a lead costs the same as the $1,299 tier with free organic leads at about 16 leads a month ($900 divided by $54.99). Below that the smaller package wins; above it the larger one does, before you count its higher setup fee and whatever else each tier includes.
From price per lead to cost per booked survey
Cost per booked survey is what you spend on leads divided by the surveys you book. It is the number that matches the way an insulation firm sells, because a survey visit is the first real cost after the lead. The conversion from lead to survey is the part you have to measure yourself, so this table assumes rates and invites you to replace them.
| Source and price | Assumed lead-to-survey rate | Cost per booked survey |
|---|---|---|
| Exclusive organic, $54.99 | 60% | $92 |
| Exclusive organic, $54.99 plus the $399 package spread over 12 leads ($88) | 60% | $147 |
| Google Ads, low end $76 | 60% | $127 |
| Google Ads, midpoint $169 | 60% | $282 |
| Google Ads, high end $262 | 60% | $437 |
| Shared marketplace, $25 (the vendor's example input) | 30% | $83 |
| Shared marketplace, $40 (assumed) | 30% | $133 |
The 60% and 30% rates are my assumptions, with the lower rate on shared leads because competitors are also contacting the customer. Your own records may say otherwise, and the table's message is the spread: the same survey can cost you $83 or $437 to book depending on source and rate. The package row shows how a monthly fee lifts a cheap-looking lead price by about 60% at 12 leads a month, and the Google Ads rows leave out the 15% management fee, so add both before you compare. Neighbouring trades take the same lead-to-survey question in solar lead generation, with cost per lead, cost per survey and AI follow-up and window replacement leads, replying fast and booking the survey.
One month for a small installer, with and without a screen
Take a small insulation firm, with figures invented to show the arithmetic. It buys 40 leads a month at an average of $90, so $3,600. A survey visit costs $110 (an hour and a half on site plus an hour driving, at $44 an hour for labour and the van). The average sale is $4,200, and each job leaves about $1,500 in gross profit.
| No screen | Good screen | Screen too strict | |
|---|---|---|---|
| Surveys booked | 24 | 17 | 17 |
| Jobs won | 7 | 7 | 6 |
| Lead spend | $3,600 | $3,600 | $3,600 |
| Survey visits at $110 | $2,640 | $1,870 | $1,870 |
| Total acquisition cost | $6,240 | $5,470 | $5,470 |
| Cost per booked survey, leads only | $150 | $212 | $212 |
| Cost per booked survey, all-in | $260 | $322 | $322 |
| Cost per job won | $891 | $781 | $912 |
| Survey-to-job rate | 29% | 41% | 35% |
| Gross profit left after acquisition | $4,260 | $5,030 | $3,530 |
Look at the cost per booked survey row. It goes up when you screen, from $150 to $212, because you book fewer surveys against the same lead spend. If that were your only measure, the screen would look like a failure. Cost per job won tells the true story: $781 with a good screen against $891 without, and 17.5 fewer hours on the road and in lofts and attics. The third column is the trap. The same screen that wrongly drops one winner leaves $3,530 of gross profit, which is $730 less than having no screen at all.
How strict a screen can afford to be
Skipping a survey saves you the visit cost and risks the profit on a job the lead might have become. So a lead is worth a survey whenever its chance of becoming a job, multiplied by the gross profit, beats the visit cost. Rearranged, the minimum chance is visit cost divided by gross profit. In the example that is $110 divided by $1,500, or 7.3%, about one in fourteen.
In practice, only decline the leads you honestly believe have under about a one in fourteen chance, such as a request with no owner's consent and no route to getting it. For the middle group, whose chances might be one in four or one in ten, use a cheaper step than a survey: a ten-minute phone call or a photo exchange. A ten-minute call costs perhaps $7 of time at $44 an hour, so it pays for itself on almost any lead with a realistic chance; the real question is whether to follow it with a $110 visit. When surveyors are fully booked in a busy season, a wasted slot also displaces a better lead, and the sum changes. If a slot spent on a one-in-ten lead pushes out a two-in-five lead, the expected loss is 40% of $1,500 minus 10% of $1,500, or $450, not the $110 visit cost. So in a full week the bar rises to the chance of the next lead waiting for a slot, and the ten-minute call does more of the sorting.
The five things to learn before a survey is booked
The qualifying script rests on five facts. None needs technical knowledge, and all of them change whether a survey makes sense.
- Property type and age. House, bungalow, flat or commercial building, and roughly how old. Age matters because older buildings can need extra checks before work starts, and the surveyor will know which.
- Who decides. An owner can agree to a survey. A tenant usually cannot, so the next question is whether the landlord or building owner has agreed. A flat adds a second question: which part of the building the customer controls.
- Access. For an attic or loft: whether there is a hatch or a fixed ladder, whether it is boarded, and how much is stored up there. For walls or floors: whether the outside and any under-floor space can be reached.
- Existing insulation. None, some, or unknown, and if some, roughly how deep or what it looks like. "Don't know" is a valid answer and is the reason for a survey.
- Timing and reason. When they want it done and what triggered the enquiry: cold rooms, bills, a sale, other building work. A grant or incentive scheme, if one could apply to your customers, belongs in this question too, and eligibility stays with the scheme's own checks.
Add one rule for the human. If the enquiry mentions damp, mould, condensation, a roof leak, old wiring in the roof space or anything the customer thinks might be asbestos, it goes straight to a person and the surveyor looks before anyone quotes. AI should not decide any of those.
The scoring prompt and the first message
Use two prompts, both with a person sending the output. The first scores an enquiry, the second drafts the reply. Paste the customer's first name and job details only, and leave phone numbers, email addresses and full addresses in your own system.
You triage insulation enquiries for a small installer. Never say what insulation is suitable, never promise savings, and never mention a price.
Score each check 0 (unknown or poor), 1 (partly) or 2 (clear):
1. Job type (attic or loft, wall, floor, spray foam)
2. Decision maker (owner, or tenant with the landlord's agreement)
3. Access (hatch or ladder, boarded or not, stored items, crawl space)
4. Existing insulation (none, some, unknown) and approximate property age
5. Timing (within 8 weeks)
6. Inside our service area: [list your area names here]
7. Photos usable (hatch view or floor view)
Total out of 14. Decision: BOOK SURVEY (11 or more), CALL FIRST (6 to 10), POLITE DECLINE (5 or fewer).
List the missing facts. Add the flag SURVEYOR FIRST if the enquiry mentions damp, mould, condensation, roof leaks, old wiring or suspected asbestos.
If the enquiry does not say, score 0. Do not guess.
Enquiry:
[paste first name and job details]An illustrative result for an enquiry that read "Semi-detached house, we own it, attic about half boarded, hatch on the landing, some old insulation, want it done this month, bedrooms are cold, two photos attached":
Job type 2, decision maker 2, access 1 (boarded half, stored items unknown), existing insulation 1 (some, depth unknown, age unknown), timing 2, service area 2, photos 2. Total 12 of 14. BOOK SURVEY. Missing: age of the property, how much is stored in the attic, whether the boarded half covers the hatch. No flags.
Two corrections before acting. The assistant scored the service area 2, but the enquiry gave only a town name and the prompt listed the areas you cover. Check it against the list yourself, and if you forgot to paste the list, the score is a guess. Second, it scored the photos 2 because the enquiry said two were attached, but it saw only the words and not the images. Glance at them yourself, because a photo of the landing and not of the attic is an easy mistake to make. If you do upload photos to the assistant, use it only to say whether a photo shows what you asked for, such as the hatch view or the floor, and not to judge the insulation in it; that call belongs to the surveyor.
Write a reply to this insulation enquiry, from a person at [firm name]. Under 90 words. Thank them, quote back one detail from the enquiry, ask the two most important missing questions (one sentence each), and say what happens next. Do not give advice on what insulation is suitable, do not mention savings, grants or prices, and do not promise a date.
Missing facts: [paste the list from the score]
Enquiry details: [paste]Three enquiries, three outcomes
Here is how the same script sorts three sample enquiries; the chance of a job is my own guess.
| Enquiry | Score and decision | Chance of a job | Next step |
|---|---|---|---|
| A: owner, semi-detached house, half-boarded attic, some old insulation, wants it this month, photos sent | 12, BOOK SURVEY | About 40% | Offer two survey slots and ask for the missing property age |
| B: flat, second floor, "wants it insulated", landlord owns the building | 5, POLITE DECLINE by the thresholds | About 10% | A person overrides to a call: ask who decides and which part of the building the customer controls; a call costs about $7, a survey $110 |
| C: tenant in a rented house, wants a price to send to the landlord | 4, POLITE DECLINE | About 3% | Send the hold message; invite the owner to get in touch |
Enquiry B scores 5, which the thresholds would decline, but at about 10% it sits above the 7.3% line from the earlier sum, so a person overrides the score and makes the ten-minute call. Enquiry C sits well below the line. A hold message for C might read:
Hi [first name], thanks for asking. We can't give a price without a survey, and we only survey when the owner has agreed to the work, so I'll hold off for now. If the owner would like to talk, ask them to get in touch and we'll arrange a visit. [your name]
It is polite, it does not say no for ever, and it never gives a price or advice, which matters because insulation proposals should explain options without promising savings. The wording works for any lead that fails on decision maker alone.
Four habits keep the script from costing you jobs.
- A person reads every decline in the first month. The cost of a false negative is a whole job, and a scoring prompt will sometimes mark down an enquiry for the way it was worded and not for what it said.
- Scores are advice, not rules. The thresholds sort the pile. The person who knows that a half-written enquiry from a landlord's agent is actually a multi-property job can override them.
- Keep a log of overrides. If you override the same kind of score often, change the prompt, not the habit.
- Capacity changes the line. With ten survey slots a week and fourteen leads scoring 11 or more, the top ten by score get the slots and the rest get a call. In a quiet week, the same leads all get a survey.
Who sends the first reply, and how fast
The script is only as good as the reply it leads to. For a firm with a handful of leads a day, the easiest set-up is manual: the owner or office person pastes each lead into a chat assistant, reads the score, edits the reply and sends it. A free tier or a plan around $20 a month is enough at that volume. Automating the first reply, so a website form triggers the message, comes later, once the script has run for a month and you trust its scores. Qualifying, booking and confirming site visits with AI goes through the automated version.
For phone leads, an AI receptionist can ask the same five questions. Smith.ai's AI receptionist, for example, is free for 25 calls a month and then $3 a call at the time I read its page, and its page describes transfers to its own live receptionists, so ask how a hand-off to your team would work. Whatever you choose, tell callers they are talking to an automated assistant, and make sure anything about damp, a leak or a smell reaches a person the same day.
Reading the numbers after a month
Keep one row per lead: date, source, cost, score, decision, whether a survey was booked, whether it happened, whether you won, and the gross profit. After a month, sort by source and work out three numbers for each: cost per booked survey, survey-to-job rate and cost per job won. Then check the screen separately: of the leads you declined or sent to a call, how many came back on their own, and how many would have been worth a survey?
Here is a separate month for a similar firm across four sources, with the survey visit at $110 and gross profit at $1,500 a job. The exclusive organic row includes the vendor's $399 package fee, and the all-in column adds the visits to the lead spend.
| Source | Leads | Lead spend | Surveys | Jobs | Cost per booked survey, leads only | All-in cost per job won |
|---|---|---|---|---|---|---|
| Exclusive organic leads at $54.99, plus the $399 package | 12 | $1,059 | 8 | 3 | $132 | $646 |
| Google Ads leads through an agency, about $150 each including its fee | 15 | $2,250 | 9 | 2 | $250 | $1,620 |
| Shared marketplace leads, about $40 each | 13 | $520 | 4 | 1 | $130 | $960 |
| Own website form and referrals | 8 | $0 | 6 | 3 | $0 | $220 |
Against $1,500 of gross profit, the Google Ads source loses $120 on each job it wins, the marketplace leaves $540, the exclusive organic source leaves about $854 even after its package fee, and the firm's own enquiries leave $1,280. The cheapest lead is not the cheapest job: the $40 marketplace leads cost $960 a job all-in, against $646 for the $54.99 exclusive ones, because fewer of them reach a survey. The figures are invented to show the method, and your own month could reverse every row. What the table cannot show is time: if the Google source also costs an hour a week of someone's attention, add it.
Two decisions follow. If a source has a cost per job won above what you will pay, stop buying from it and move the spend to a source that beats it, whatever the sales call promised. If the screen declined leads that later booked with someone else, loosen it by a point and run another month. If your sources include a vendor's exclusive leads, run them for a month on the no-contract terms some vendors print, with the package fee inside the sum, and keep your own enquiries in the same sheet as the benchmark that paid leads must beat.
Insulation leads: questions before you buy or screen
How much do insulation leads cost?
There is no single price. One lead vendor's insulation page prints a flat $54.99 for an exclusive organic lead on its $399-a-month package, and a Google Ads range of $76 to $262 per vetted lead, the 10th to 90th percentile of what its own clients paid over twelve months. Other sellers print nothing, and shared marketplaces publish no insulation price. Measure your own cost per booked survey.
Are exclusive insulation leads worth the higher price?
Only if they lift your survey-to-job rate by more than they raise your cost. One lead vendor's calculator makes the case with example inputs of $25 for a shared lead and $75 for an exclusive one, which are its own inputs, not market data. Compare cost per job won across at least a month of leads from each source, not price per lead.
Can AI qualify insulation leads on its own?
It can score an enquiry and draft the follow-up questions, and I would start with a person sending the message. It should not judge insulation condition from photos or say what is suitable, because that is the surveyor's job. Send anything mentioning damp, mould, roof leaks, old wiring or possible asbestos to a person.
What should I ask before booking an insulation survey?
Five things: the property type and approximate age, who decides on the work, how the area is reached, what insulation is already there, and when they want it done. Add a photo of the hatch or the floor if the customer can take one. A short script covers all five in two or three messages.
Do Local Services Ads work for insulation?
Check before you plan around them. Categories vary by area, so use Google's eligibility tool for insulation first. You pay for valid leads such as calls and messages, not clicks, and Google is moving these ads into Performance Max in phases from August 2026 to 2027, which removes the option to set a maximum cost per lead. Old reports do not carry over, so download them.
Further reads
- Foundation Repair Leads: Screen Crack Photos With AI Before the Visit — Screening customer photos before a visit, in another trade.
- Garage Door Leads: Qualify and Answer Them Faster With AI — Another trade's qualifying questions and fast replies.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: one lead vendor's insulation, roofing, pricing and shared-against-exclusive pages, and three other insulation lead sellers' sites (October 2026); Google Ads Help on Local Services Ads moving to Performance Max; Smith.ai's AI receptionist pricing page. The worked examples use invented figures.