How to get real estate leads without torching your budget

Where real estate leads actually come from

Start with the number most lead-gen advice skips. In NAR's 2025 Profile of Home Buyers and Sellers, 91% of sellers used an agent, and 66% of them hired someone referred to them or an agent they had already worked with. Eighty percent interviewed only one agent before they signed. Buyers acted almost the same way, at 58%.

Read that again. The majority of business goes to whoever the seller already trusts, and the decision is usually made before any ad, form, or bought lead enters the picture. That is the 80/20 rule in its real shape: a small slice of past clients and their referrals drives most of your closings, and every cold source fights over what is left.

This does not make paid leads worthless. It means a plan that spends everything on strangers and nothing on the people who already know you has the ratio upside down. So before you buy a single lead, put a name and a birthday reminder on every past client and ask each one, out loud, who they know that is moving. Then add a channel.

The rest of this page treats leads as three separate problems: buyer leads, seller leads, and off-market seller leads for investors. They cost different amounts, close at different rates, and reward different work. Blending them is why so much advice reads identically and helps nobody.

The real cost of buying leads

Buying leads is the fastest route to a pipeline. It is also the easiest way to spend $4,000 on one closing without noticing. The prices below come from HousingWire's 2026 vendor roundup, updated in June. Watch the per-lead column, because that is where the money actually leaks.

Here is the part the vendors leave off the landing page. HousingWire, citing NAR, puts online lead conversion at 0.4% to 1.2%. Take Zillow at a middle price of $40 a lead and a generous 1% conversion. That is 100 leads, or $4,000, for one deal, before you count the hours spent calling people who filled out a form at midnight and forgot by breakfast. On a shared product, two other agents got the same name and are dialing it too. Shared is the word to fear. Zillow Premier Agent can sell that same lead to several agents at once, so your 1% got cut into thirds before you dialed.

None of that makes portal leads a mistake. Fast, cheap-per-lead volume is useful when your calendar is empty and your follow-up is sharp. Just price the channel by cost per closing, not cost per lead, and the ranking of your options changes overnight.

Free and low-cost sources

Free leads are not free. They cost hours instead of dollars, and for a new agent with more time than money, that is the right trade. Four sources carry the load.

Your sphere of influence is first, and it is not a euphemism for pestering friends. It is a list of every person who would take your call, contacted on a schedule, with something useful each time. Open houses are second: you meet unrepresented buyers in person, and the sign-in sheet is the only lead capture that happens face to face. Third is content plus a Google Business Profile, which turns your name into the answer when a neighbor searches at 11pm. Fourth is plain consistency, the boring engine behind the referral numbers above.

Put rough numbers on it. A busy open house might collect 15 to 25 sign-ins, of which a handful are real and unrepresented, so four Saturdays a month is a steady trickle, not a flood. A sphere of 200 contacts, worked properly, tends to throw off enough referral business to anchor a year once it is warm. Neither shows up on a dashboard the week you start. That lag is the whole reason people call these channels free.

The catch is speed. Sphere and content pay off in months, not weeks, so a brand-new agent who needs a check in 60 days cannot live on them alone. Pair the slow free channel with one fast paid one and let the free work compound underneath.

Seller leads are the scarce half

Buyer leads are everywhere and cheap. Seller leads are the scarce, valuable half, and the market got tighter: NAR clocked for-sale-by-owner at a historic low of 5% of home sales in 2025, which means 95% of sellers hand the listing to an agent. Winning that listing is mostly a question of being known and being early.

Being known is the referral engine again. Being early is a data problem, and it is where geo-farming earns its keep. Pick a neighborhood, become the obvious name in it, and show up before the owner has decided to sell. Our guides on getting listings and building a real estate farming plan break down the routine. The short version: repetition in one ZIP beats a scattershot blast across five.

Off-market leads for investors

Investors and wholesalers are not chasing pre-approved buyers. They want owners who might sell below market, often before the property ever lists. That flips the whole method.

Driving for dollars still works: you log distressed-looking houses block by block and skip trace the owners. Our driving for dollars guide covers the apps and the routine. Beyond that, the classic pulls are absentee owners, tired landlords, and pre-foreclosures, the same motivated seller lists every wholesaler competes over. The problem is that everyone buys the identical list from the identical vendor and mails it the identical week. Sameness is the enemy.

Here is the concrete version. Pull the absentee owners in one county and you might get 8,000 names, the same 8,000 your three nearest competitors also bought. Skip tracing them runs a few cents to a couple of dollars a record, so the cost is real and the differentiation is zero. What separates a full voicemail box from a dead one is which 500 of those 8,000 you contact first. The edge is ranking, not raw count, which is the next section.

Direct mail and predictive lists

Direct mail is cheap to send and expensive to waste. A retail First-Class postcard stamp is $0.65 as of the July 2026 rates, and USPS Marketing Mail starts near $0.227 a piece at commercial volumes. Postage is never the reason mail fails. The list is. The list is the product, and the stamp is a rounding error.

Mail 5,000 random owners and you pay to reach 4,900 people who are not selling this year. The fix is ranking every owner in a market by how likely they are to move in the next 6 to 12 months, then mailing the top of that list instead of the whole thing. This is the same predictive category SmartZip sits in, and it is exactly what Farmrix does: it scores owners on likelihood to sell, ranks them, and prints and mails postcards to the ones at the top. Less mail, more deals. The point is not more volume, it is a shorter, sharper list, so a 500-piece drop to ranked owners can beat a 2,000-piece drop to a raw pull.

The economics are easy to check. Farmrix's smallest package is 500 ranked owners and 500 postcards for $1,195, and the ladder climbs to 16,000 owners and 16,000 postcards at $19,995. Set that against a raw 5,000-piece blast at $0.65 a stamp, which is $3,250 in postage alone to reach mostly the wrong households. Fewer, better-aimed pieces is the cheaper road to the same number of phone calls.

The follow-up gap that wastes leads

The reason most lead budgets underperform is not the source. It is that the lead gets one call and never a second. A form fill at a 1% conversion rate only pays off if you actually chase the other 99, and most agents quit after touch one or two. Build the cadence before you build the budget. Same-day first contact, then a planned run of five to twelve touches over the following weeks by call, text, and email. A cheap lead you call eight times beats a premium lead you call once. That one habit moves your cost per closing more than switching vendors ever will.

The pay-at-closing trap

Here is a common piece of advice that is wrong for most working agents: chase the $0-upfront, pay-at-closing lead sources because they feel risk-free. Run the arithmetic before you believe it.

Sold.com charges a 30% to 35% referral fee at closing. Say you close a $400,000 sale at a 2.5% listing commission. That is $10,000 to you, minus a 35% fee, so $3,500 walks out the door on that one deal, and you still did all the calling, showings, and paperwork. Close ten of those in a year and the channel cost you $35,000, quietly, with no invoice to flinch at. A $40 portal lead looks reckless and a 35% cut looks safe, yet if you convert consistently the cut is the pricier choice by a wide margin. Pay-at-closing suits agents who close rarely and want zero fixed cost. It punishes the ones who close a lot.

Match the channel to your stage

The right answer depends on how much time and cash you have, not on which vendor has the loudest ad. A quick map:

• New agent, more time than money: sphere plus open houses plus content, one ZIP to farm, no paid leads until you have a follow-up system you actually run.
• Established agent with cash: keep the referral engine, add a ranked seller-lead or farming channel, and treat portal leads as fill-in volume, not the plan.
• Investor or wholesaler: a ranked, likely-to-sell mailing list beats a raw absentee pull, and it pairs with driving for dollars for the deals that never hit any database.

Whatever you pick, measure one number: cost per closed deal. A channel that produces cheap leads you never close is more expensive than a pricey one that converts, and only your own tracking, over 90 days, tells you which is which. Your market and your follow-up will move these numbers, so trust your log over any blog.

What to do next

Pick two moves for the next 30 days. First, call your last ten clients and ask, by name, who they know that is thinking about moving, because that is the 66% source doing the heavy lifting. Second, add exactly one paid channel that matches your stage from the map above, and set a spreadsheet to track its cost per closing from day one.

If seller leads are the half you want, skip the raw list. Look at how seller leads get sourced, then let Farmrix rank the owners in your market by likelihood to sell and mail the top of the list for you. You send less mail, you chase fewer dead names, and you spend your hours on the owners most likely to actually move.