
Pay for whichever lead type gives you the lowest cost per booked job after refunds, not the lowest price per lead. Run a short, capped test of each, track every lead to a booked job or a no, and keep only the source that pays for itself.
Lead sellers love to argue about exclusive versus shared. The honest answer is that neither is always better. A cheap shared lead that four other roofers also call can cost more per job than an expensive exclusive one, and an exclusive lead that is not really exclusive is the worst deal of all. The way out is to judge every source by the same math.
Know what you are actually buying
Exclusive leads
The seller promises the lead goes only to you. You pay more per lead because you are not racing anyone. The catch: "exclusive" can mean different things. It might mean exclusive for 24 hours, exclusive within one platform while the customer also filled out three other forms, or exclusive only in your zip code. Ask the seller to define it in writing.
Shared leads
The same request goes to several contractors, often three or four. Each lead costs less, but the customer gets several calls within minutes. The contractor who calls first and sounds organized often wins. If your office cannot call back fast, shared leads tend to look worse than they really are.
Leads you generate yourself
Calls and forms from your own website, Google Business Profile, ads, and referrals. You pay for the traffic and the site, not per lead, and nobody else gets the customer. These take longer to build but you own them. Compare bought leads against this baseline when you can.
Read the duplicate and refund rules first
Before the price, read the fine print. These rules decide your real cost.
- Duplicates. If the same homeowner submits twice, or you already had them as a customer, do you pay twice? Ask how the seller detects duplicates and how long you have to report one.
- Bad leads. Wrong numbers, out-of-area addresses, and requests for services you do not offer should be refundable. Ask what counts, how to claim it, and the deadline.
- Disconnected or fake contacts. Ask whether you get credit when the phone number does not work or the email bounces.
- Contract terms. Minimum spend, auto-renew, and notice periods to cancel. A cheap lead locked into a long contract is not cheap.
Keep a simple record of every claim you file and every credit you receive. If refunds are slow or keep getting denied, count that time as a cost.
Compare by cost per booked job
Price per lead hides the only thing that matters. Use this calculation for each source:
Total spend for the period, minus refunds received, divided by the number of booked jobs from that source.
Example numbers. Say a remodeler tests two sources for a month.
Source A sells shared leads. They spend $1,200 on 40 leads, get $120 back in refunds, and book 5 jobs. Cost per booked job: $1,080 divided by 5, which is $216.
Source B sells exclusive leads. They spend $1,500 on 15 leads, get $0 back, and book 6 jobs. Cost per booked job: $1,500 divided by 6, which is $250.
Here the cheaper, shared source wins, even though its close rate is lower. Change one number, say Source A books only 3 jobs because the office calls back late, and the cost per booked job jumps to $360. The math moves fast, which is why you have to measure your own results instead of trusting a seller's averages.
Then compare cost per booked job with what a booked job is worth to you after materials and labor. The U.S. Small Business Administration offers a free break-even point calculator if you want to check how many jobs you need to cover fixed costs. If a lead source costs more per booked job than the job earns, it is losing money no matter how busy it keeps the phone.
Run a pilot with stop conditions
Do not sign a long contract to find out if a source works. Set up a short test you control.
Set the cap and the time
Pick a fixed budget and a fixed window, such as four to six weeks. Write down the budget before you start. When it is spent, the test is over.
Track every lead to an outcome
Every lead gets one of three labels: booked, not booked, or refunded. Use the same sheet or CRM for every source so the numbers compare fairly. If you run ads too, track those calls the same way. Our guide on tracking which Google Ads calls turned into booked jobs covers that side.
Write the stop rules in advance
Decide before the test what ends it early. For example: no booked job after the first 15 leads, refund claims denied more than twice, or the same lead showing up from two sources. Writing it down first stops you from talking yourself into another month.
Fix your side before blaming the source
Speed and follow-up matter, especially for shared leads. If your team calls back hours later, any source will look bad. During the test, aim for a fast first call and at least two follow-ups. That way you are measuring the lead, not your office's busiest week.
When each type makes sense
Shared leads can work if your office answers fast, you have a strong follow-up habit, and your margins allow you to win a smaller share of a larger pile.
Exclusive leads can work if your jobs are high value, your callback is slower because crews are in the field, or your close rate improves when you are not competing on the first call.
Your own leads make sense for almost everyone over time, because you are not renting attention. Building them takes a website that turns visitors into calls, a Google Business Profile that is kept up, and some patience. Our lead generation and digital marketing work focuses on that side: owning the source instead of renting it.
Related guides
Common questions
Are exclusive leads worth the extra cost?
Sometimes. They are worth it only if their cost per booked job, after refunds, is lower than your other options. Test both with the same budget cap and compare the results.
How do I know if a lead was really exclusive?
You usually cannot see it directly. Ask the customer how they found you and whether other companies called. If many say yes, ask the seller to explain their exclusivity terms or give credit.
What refund rules should I insist on?
Clear credits for duplicates, wrong or disconnected numbers, out-of-area requests, and services you do not offer, with a reasonable claim window. Get them in writing before you pay.
How long should a lead test run?
Long enough to get enough leads to judge, often four to six weeks, with a fixed budget. End it early if your written stop conditions are hit.
Should I stop buying leads completely?
Not necessarily. Many contractors use bought leads to fill gaps while they build their own website, reviews, and search presence. The goal is to rely on bought leads less over time, not to drop them overnight.
Next step
Want a free plan for comparing your lead sources by cost per booked job? We will help you set up the tracking sheet, the test budget, and the stop rules. Get my free plan



