Where to buy purchase mortgage leads

Purchase leads get sold as one product. They are four different things, and the difference is not quality.

It is how much of the work was already done, and how many other lenders are dialing the same number as you.

This page is about buying them. If you are choosing between buying leads at all and building referral relationships with agents, that is a different and usually better question, and it is covered at the end.

Where purchase leads actually come from

Every purchase lead on the market starts the same way. Somebody who wants to buy a house fills in a form or answers an ad. What happens to that enquiry next is the entire product difference.

A fifth route belonged in this table a year ago. Prescreened lists built from the fact that somebody had a mortgage credit inquiry run, sold to lenders with no existing relationship to that consumer, stopped being lawful in March 2026.

Source: Homebuyers Privacy Protection Act, Public Law 119-36, 119th Congress. govinfo.gov · Plain reading: the trigger lead ban in plain terms.

Why purchase is harder to buy than refinance

Purchase is structurally harder to buy than refinance, and the reason is the clock rather than the quality of the lead. Most buying guides skip it. It is also the thing that decides whether a purchase lead budget survives its first quarter, because a desk that measures purchase the way it measures refinance will conclude the source failed.

A refinance enquiry is a decision somebody can act on today. A purchase enquiry usually arrives while the person is still looking at houses. The National Association of Realtors puts the median home search at ten weeks before an offer is accepted, and ICE Mortgage Technology reported 36.8 days from application to closing on purchase loans in March 2026. The application is the end of the search, not the beginning of the relationship.

Source: National Association of Realtors, Profile of Home Buyers and Sellers, on median weeks searched. nar.realtor · ICE Mortgage Technology, March 2026, days from application to closing on purchase loans. Both figures describe the market, not any one vendor's leads, and neither is a promise about a particular file.

Three consequences follow, and all three are budget decisions rather than opinions.

• A thirty day measurement window will tell you purchase leads failed. Measured on that window they nearly always will, because the median consumer has not chosen a house yet.
• The follow-up cadence matters more than on refinance. A refinance lead that goes cold is gone. A purchase lead that goes quiet is often still shopping, and the lender who is still there in week seven is the one who takes the application.
• Pre-approval is the real first conversion, not the loan. Measuring purchase spend against funded loans in the same month measures your patience, not the source. What can and cannot be known before you pull credit is its own question, taken apart in what is actually for sale under the heading of pre-approval leads.

Once the file is in the pipeline, how the lead arrived stops mattering almost entirely. ICE reported a 79.8 percent closing rate on purchase applications for December 2021, against 77.5 percent on refinance. The whole difference between lead sources lives before the application, which is another way of saying it lives in the follow-up.

Source: ICE Mortgage Technology (formerly Ellie Mae) Origination Insight Report, December 2021. The series has no traceable data point after 2021 and the report is no longer publicly hosted, so the year travels with the figure. Denominators and caveats in what the data supports between a lead and a funded loan.

The five questions to ask before you pay

Ask all five in writing. The answers separate vendors more reliably than any price list, and a vendor who has measured their own product can answer them in a paragraph.

1. How many lenders receive this same lead? Ask for a sentence, not an adjective. The useful answer is a number. If the answer is one, ask them to write that down.
2. What was the consumer asked, and what did they agree to? You are going to call this person. What they consented to is your compliance position, not the vendor's.
3. Which criteria was this screened against, and who sets them? If the vendor says you set them, ask what happens to a lead that misses. If the criteria are theirs and fixed, ask for the list. Ask to see a redacted delivered lead too, so you know the shape of the record before you buy a hundred of them; ours is published in what a mortgage lead looks like.
4. How long between the consumer pressing submit and the lead reaching my system? Ask for the median and the tail, not the best case. Minutes and seconds are different products.
5. What happens to a lead that misses the criteria? The answer tells you whether screening is a real gate or a description of the average.

Consent, and why it is your problem rather than theirs

The lead arrives with a consent record attached. The call is made by you, so the exposure is yours.

• Ask which parties were named. Under the Telemarketing Sales Rule, a written agreement permitting calls to a registry listed number has to evidence authorization for calls by or on behalf of a specific party. Ask to see the list you were on.
• Scrubbing is a separate and continuing obligation. The version of the federal Do Not Call registry you scrub against has to be no more than 31 days old at the time of the call. This is the rule that quietly turns cheap aged data expensive.
• Bundled consent is still lawful. The FCC adopted a one-to-one consent rule that would have forced a consumer to agree to one seller at a time. On 24 January 2025, three days before it took effect, the Eleventh Circuit vacated it. So the shared model is not going anywhere, and checking what a consumer actually agreed to sits with the desk making the call.

Source: 16 CFR 310.4(b)(1)(iii)(B)(1) and 310.4(b)(3)(iv), Telemarketing Sales Rule. ecfr.gov · Prior express written consent and its required disclosures at 47 CFR 64.1200(f)(9). ecfr.gov · Insurance Marketing Coalition Ltd. v. FCC, No. 24-10277 (11th Cir., 24 January 2025). media.ca11.uscourts.gov

Comparing price without fooling yourself

Price per lead ranks the four routes in one fixed order. Cost per funded loan can rank them in a different one, and it is the only ranking that pays anybody.

• Divide the price by your own closing rate. Not the seller's, and not a figure off a benchmark page. Lead to funded on a screened, singly sold purchase lead realistically runs around 3 to 8 percent, and vendor figures above that band are usually deflated by a factor of somewhere between one and a half and three when you check the denominator.
• Add the hours. A lead at a quarter of the price that needs four times the dials costs the same, unless those hours were already idle, in which case it is a bargain.
• Expect the bands to be far apart. Across the market, cost per funded loan on singly sold screened leads tends to land in the region of 1,200 to 2,000 dollars, against roughly 5,000 to 15,000 dollars and up where the same enquiry went to several lenders. Those are industry bands rather than quotes, and your own file will move them.
• The order flips on capacity, not on quality. Hours you already pay for are free. Hours you do not have cost more than any price per lead on the list.

What we could not source

Four things on this page we could not source to a primary figure, named rather than papered over. Each is a place where we went looking for a measurement, found only material published by a company selling one of the four routes, and decided to say so instead of borrowing their number.

• There is no independent price index for purchase leads. The bands above are our read of what gets quoted in this market. That is not a measurement, and no vendor publishes one.
• There is no audited study comparing conversion by purchase lead type. Every table we chased was published by a company selling one of the four routes, including the tables that favour ours.
• The ICE closing rate series stops in 2021. We use it because it is the best primary figure that exists, and we date it every time rather than implying it is current.
• Contact rate and pre-approval rate have no primary source at all. Not disputed, absent. Taken apart in what the data supports.

Where we fit, and where we do not

We sell the second route, so read this section knowing that. We have tried to write it the way we would want a competitor to write theirs.

• One lead goes to exactly one client. That sentence is the whole meaning of the word, and it is the thing worth getting in writing from anybody.
• Screened against our criteria before it is sent. The criteria are ours and they are fixed. A lead that misses them is never delivered and never billed.
• Delivered while the consumer is still on the page. Median delivery under one second, and 97 of 100 inside ten seconds.
• Phone verified on our landing page leads. Scoped deliberately, because that is the lead this page is about and the one we push.
• Consent captured on the form. TrustedForm certificate by ActiveProspect, captured on the landing page form, and pulled on request.

Where we are the wrong answer, which is a real list.

• You want maximum conversations for minimum spend. Buy aged data and work it properly. We lose that comparison and we should.
• You want the phone to ring already connected. That is a live transfer product, not this one.
• You need volume above everything. Criteria that hold are criteria that say no, and a desk sized for volume will find the flow smaller than it wanted.
• You have strong agent relationships and spare time. Referrals convert better than anything on this page and cost nothing per lead. Buying makes sense when you need flow that does not depend on somebody else's pipeline, not as a substitute for the relationships.

Our own clients report closing 5% to 10% of our leads to funded loans, with our best client up to 14.6%.

Source: client self-reports given on sales calls, not a platform measurement. The loan is originated in the client's own system, so we record no closings of our own and most delivered leads never get a status pushed back to us. Read these as what buyers told us about their own files, not as an audited figure.

Choosing, in one pass

• Spare calling hours and a tight budget → aged data, worked properly.
• Genuinely competitive rate and instant answering → the shared route.
• Nobody free to dial, somebody free to talk → live transfers.
• Fewer conversations, more of them worth having, and the first call → screened and sold once.
• Strong agent relationships already → feed those first, and buy to fill the gaps rather than to replace them.
• No spare hours and no spare budget → fix the follow-up process before buying anything. None of the four survives a desk that calls twice and stops.

Related reading: the four lead types compared, what shared leads do to a borrower, and how many calls it takes to reach a mortgage lead. Buying questions are answered on our FAQ.