Cost per lead is the number almost every real estate investor tracks and the number that most often leads them to the wrong decision. Cost per deal is the one that decides whether a channel deserves next month's money.
The two can rank channels in exactly opposite orders, and the gap between them is wider in this business than in almost any other.
Why the Cheap Leads Are Often the Worst
The mechanism is simple once you see it. Cost per lead measures how cheaply you can get someone to fill in a form. Nothing in that measurement has any connection to whether the person has a reason to sell.
A social ad offering a free home value report will produce leads at a very low cost, because a lot of people are curious about what their house is worth and almost none of them intend to sell. A probate mailing costs far more per response, because you are paying for postage to a filtered list, and the people who respond have a property they need to deal with.
Run those through to closings and the ranking inverts completely. The channel that looked ten times cheaper turns out to cost more per deal, sometimes by a lot, and an investor optimizing cost per lead will systematically shift budget toward the worse channel while believing they are being disciplined.
Working the Arithmetic
Take two channels over a quarter.
Channel A. Two thousand dollars spent, two hundred leads. Cost per lead is ten dollars, which looks excellent. Of those two hundred, thirty were reachable and interested, six became appointments, and one became a contract. Cost per deal: two thousand.
Channel B. Two thousand dollars spent, twenty five leads. Cost per lead is eighty dollars, which looks poor. Of those twenty five, eighteen were reachable and interested, nine became appointments, and three became contracts. Cost per deal: six hundred and sixty seven.
Channel B costs eight times more per lead and produces deals at a third of the cost. An investor tracking only the first number moves money to Channel A and their business gets quietly worse over the following two quarters, with no single moment where the mistake is visible.
The other thing hidden in those numbers: Channel A consumed two hundred contacts worth of your time to produce one deal, and Channel B consumed twenty five to produce three. Cost per deal understates the difference, because it does not price the hours.
Calculating It When Your Sample Is Small
The obvious objection is that three deals is not a sample. That is correct, and there are ways to get a usable number anyway.
Use a longer window. Quarterly at minimum, and a rolling twelve months for any channel you have run that long. A month of closings tells you about timing, not performance.
Match the spend to the leads, not to the calendar. If your average cycle from first contact to closing is ninety days, this quarter's closings came from last quarter's spend. Comparing simultaneous spend and closings is the single most common calculation error here, and it makes growing channels look bad and shrinking ones look good.
Use an intermediate stage when closings are too rare. Cost per appointment or cost per qualified conversation happens five to ten times more often than cost per deal, and it correlates well enough to act on. This is the practical answer for any channel with fewer than about ten closings, and it lets you make a call in weeks rather than a year.
Use the median, not the mean, for deal value. One thirty thousand dollar assignment among four deals will distort an average past the point of usefulness.
The Number Underneath Both
Cost per deal still is not the end of it, because deals differ in what they pay.
A channel producing deals at three thousand each that net eight thousand is worse than one producing deals at six thousand each that net twenty five. The measure that resolves this is profit per dollar spent, or simply gross profit from a channel divided by spend on that channel.
This matters because channels genuinely differ in the kind of deal they produce. Distressed and situational niches tend toward larger spreads, since the seller is trading price for certainty. Broad-market channels tend toward thinner ones, because the people responding have options and know it. Both facts are visible in your own numbers once you split them by channel, and invisible if you do not.
What Cost Per Lead Is Still Good For
It has a real use, and discarding it entirely is the overcorrection.
Cost per lead is a fast feedback signal. It updates in days rather than quarters, which makes it the right measure for spotting that something broke. If your cost per lead triples overnight, something in the campaign, the page or the platform has changed, and you want to know that today rather than next quarter.
So the working rule: use cost per lead to monitor, use cost per deal to decide. The failure is only when the fast number gets used for the slow decision.
Cost per lead is also the right measure when comparing two versions of the same thing within one channel, because lead quality is roughly held constant. Two landing pages against the same traffic can be judged on conversion and cost per lead, which is how testing at low volume stays tractable, as covered in split testing when you do not have much traffic.
The Tracking You Need
You cannot calculate any of this retroactively, which is why most investors cannot produce the number when asked.
Four things, recorded at the time. Spend by channel by month. Every lead tagged with its source when it arrives, not reconstructed later. The stage each lead reached, with dates. And the net profit on each closed deal.
The tagging is where it breaks in practice, because a lead that arrives by phone with no source recorded is permanently unattributable. Distinct phone numbers per channel fix most of it mechanically, and asking every caller how they heard about you fixes much of the rest. Both are covered in attribution for real estate investors.
Where these records live matters less than that they exist in one place, which is the argument in the guide to investor CRMs.
The Version of This Number for Time, Not Money
An investor doing their own calling has a second cost that never appears in a spend column, and ignoring it distorts the comparison badly.
Take the two channels above. Channel A required roughly two hundred contact attempts to produce one deal. Channel B required twenty five to produce three. On money, Channel B is three times better. On hours, it is closer to twenty times better.
That gap matters most for the investor who is the constraint in their own business. If you can make forty calls a day, your capacity is the scarce resource rather than your budget, and the right measure is deals per hundred contacts rather than deals per dollar.
The practical consequence: a channel that is slightly more expensive per deal but far less demanding of your time is usually the better choice while you are still working leads yourself. That inverts once you hire, because a paid caller converts your time constraint back into a money constraint, which is one of the real arguments in when to hire your first acquisitions person.
Worth calculating once a year: total contact attempts divided by deals, by channel. Most investors have never seen this number and it frequently changes which channel they want more of.
What to Do With the Answer
The decisions follow more directly than investors expect, with one asymmetry worth naming.
If a channel produces deals below your target cost, spend more until it stops doing so. Channels have a ceiling, and finding it is the point.
If a channel produces deals above target, the question is whether the problem is the channel or your handling of it. A poor contact rate or a slow response will ruin a good channel's numbers, and switching channels does not fix either, which is the diagnosis in why your leads are not closing.
If a channel has produced no deals at all, the question is whether it has had enough volume to prove anything. Zero deals from thirty leads is not evidence. Zero from three hundred is.
The asymmetry: require more evidence to kill a channel than to scale one. Over-funding something ordinary is a recoverable mistake that announces itself in a quarter. Shutting down something that was working is not, because nothing later tells you it was a mistake. Why that caution is mathematically justified rather than merely cautious is in why small sample marketing numbers mislead, and how cost per deal fits the wider set is in marketing metrics for real estate investors.