Email marketing for real estate.

Key takeaway

A real estate list gets chosen from about once a decade: sellers had owned their homes a median of 11 years before selling. So the program's job is to still be arriving, and still be welcome, in year nine. Build four sends rather than seven, segment by where somebody sits on the deal timeline, and read your own state commission's advertising rule once rather than guessing at it. Then I'd check whose domain you send from, because 5,000 a day counts subdomains.

What email marketing looks like when the product is a house.

Email marketing for real estate is sending commercial email on a schedule, to a database of people who gave you their address, from a domain you control. That definition is the same for a realtor as for a software company. What's different in this vertical isn't the writing. It's the arithmetic.

A database of eight hundred past clients, neighbors, and inquiries contains perhaps a dozen transactions in any given year, and you can't tell in advance which dozen. That's what the email is for. It isn't a campaign that converts, it's a standing invitation to be the person they call when the year turns out to be theirs. The question isn't whether last month's send produced a listing appointment, it's whether all eight hundred still open the mail and still want it.

The agents I've watched struggle with this aren't bad writers. They send four good newsletters, get busy with a closing, come back nine months later, and mail a cold database from a domain nobody ever authenticated. It lands in spam, and the conclusion drawn is that email doesn't work in this business. Two separate things broke there, and the rest of this page is about both.

The sales cycle is eleven years long, and your list has to survive it.

The list is everybody who might transact, plus everybody who might refer somebody who transacts. That's a wider thing than a lead list and it behaves differently: most of it will do nothing for years, and the part that's about to move gives very little warning.

The scale of that is measured. In NAR's 2025 profile of buyers and sellers, sellers had owned their homes a median of 11 years before selling, an all-time high, and 88% of buyers and 91% of sellers worked with an agent. Almost everybody uses an agent, then, and they pick one roughly once a decade, so the whole task is to still be arriving, and still be welcome, in year nine.

Judge the program, then, on whether the list is still deliverable and still opted in after five years, because that's the horizon it operates on.

A decade-long cadence is exactly the thing a person doesn't sustain by hand. Nobody remembers in 2031 that the family at number 14 bought in 2020 and is two years past the median. Something has to hold that clock, and what the something is turns out to be the real platform question, which is the last section of this page.

The four sends worth building before anything else.

Four, and they're worth building in this order. Each of these four does a job the other three don't, and an agent who only ever ships the first two is still ahead of most of the market.

1. The introduction email: fires when somebody joins the database. Who you are, what you'll send, and how often. It's the only email whose timing the reader chose, which is why it gets read.
2. The monthly local note: one thing about their actual streets that a portal can't tell them. What sold, what it went for, what changed. Monthly is a real answer, and it's the cadence most agents can hold.
3. The activity email: a listing, a sale, an open house, written or templated by you when there's something worth saying. Useful to the segment that's moving now, and noise to everybody else.
4. The anniversary and valuation check-in: runs off the purchase date rather than off your memory. Year one is a thank-you, year seven is a valuation, and the date arithmetic does the remembering.

The fourth is the one everybody skips and the only one that compounds, because it's the only email here still arriving in year six, which is where the eleven-year median leaves most of your database. Build it early even though it pays late. None of the four is worth building, though, until you know who gets which.

Segments that match how a deal actually works.

The four segments that matter here aren't demographic. They're positions on a timeline, and the timeline is the deal: somebody who bought two years ago, somebody approaching the tenure line, somebody touring houses on Saturday, and somebody who filled in a portal form and never replied. Only one of the four changes weekly.

1. Past clients and sphere of influence: the referral engine, and the segment with the longest half-life. Low frequency, high personal content, and the one that most rewards being left alone between touches.
2. Sellers, or people who will be: anybody approaching that eleven-year line. Equity, valuation, and what the street is doing, which is the one segment where a market number genuinely helps rather than decorates.
3. Active buyers: the only fast-moving segment, measured in weeks. Timeline, price band, and area. It goes stale fastest and needs pruning hardest, because a buyer who closed in March is not a buyer.
4. Cold and unworked leads: portal and advertising inquiries who never replied. Long horizon, low frequency, and the segment your complaints come from if you mail it the way you mail the first one.

The mistake is rarely too few segments. It's letting somebody sit in the wrong one for a year: the buyer who closed and stayed on the buyer list, the past client who's eleven years in and still getting open-house mail. Whatever moves people between the four has to run more often than you remember to run it.

Your marketing email is an advertisement, and two rulebooks say so.

Two rulebooks apply to the same message, and passing one tells you nothing about the other: the federal law governing commercial email, and the state commission that issued your license. The second is the one most agents have never read. I'd read it once rather than guess at it, because the part it governs is your footer, and a footer is written once and sent eight hundred times.

Start with the licence, because it's the half that's specific to you. Most state commissions treat any communication designed to attract the public to your brokerage services as advertising, and most name email expressly. What that usually means in practice is a footer carrying your name or the team's plus the broker's, with the broker's name given real prominence rather than tucked into small print. The useful half is the carve-out: mail from a licence holder to that licence holder's current client generally isn't an advertisement. So the market update going to eight hundred past clients is covered, and the email to the buyer already under contract isn't. The wording is your own commission's, and it's worth reading once rather than guessing.

The other half isn't specific to you at all, and it's shorter. Honest headers and a subject line that matches the message. A reader who can tell it's a promotion before opening it. An opt-out that works on the first click. And a real contact address in every send, which is a decision rather than a checkbox for a sole practitioner, because the business address is often the house and it's going out on eight hundred emails. Use the brokerage's address if it's yours to use, or rent a box.

Both rulebooks are about the message, not about whether it arrives. You can name your broker, print an address, honor every opt-out, and still be filtered before a human sees any of it. That's a third gate, and it belongs to somebody else.

What the mailbox providers require, and the brokerage-domain trap.

Gmail, Yahoo, and Microsoft decide whether the message is delivered at all, and they publish what they want. It's short, it's enforced by machines that never read your subject line, and most of it is one-time DNS work a competent person finishes in an afternoon. Google's sender guidelines set the bar in two tiers, and the second is about volume.

That 5,000 is where this vertical gets caught, because it's counted per primary domain over 24 hours and subdomains roll up into the same number. One agent mailing eight hundred contacts a month is nowhere near it. A brokerage sending on behalf of forty agents from a single domain crosses it every week, and no individual agent has any idea they're inside a bulk-sender regime with a complaint ceiling attached. Reputation pools the same way: one agent's bought list is measured against the domain everybody else sends from. If the domain is the brokerage's, the sender reputation isn't yours, and neither is the fix. The fastest way one agent tips that pool is frequency: complaint rates spike hardest when a list used to getting almost nothing suddenly gets three or four emails a week, and on a shared brokerage domain that spike is everyone's problem.

Who actually does this on a Tuesday.

The platform question for an agent turns on four things, and none of them is which dashboard looks nicest. Whether you can export the database and its history in full, without asking anybody's permission. Whether the sending domain is yours or the brokerage's. Whether the bill scales on contacts held or on messages sent. And whether the person who has to run it has any time.

Every platform in this category was designed for a human clicking buttons, and assumes that human has spare evenings. In a one-agent business that human is the agent, after showings. In a brokerage it's the operations or marketing lead, running it for forty of them alongside everything else. So the honest question isn't which interface you prefer, it's who does the work: authenticating a domain, importing a database, keeping four segments accurate as people move between them, scheduling a monthly note eleven years running, and pulling this month's anniversary list. Every one of those is the same operation with a different argument each time.

That's the bet we've made. Nitrosend is MCP-first: every capability is an API endpoint and an MCP tool before it's a screen, so the anniversary pull and the segment update are commands rather than an evening. Contacts are unlimited on every plan, including Free, so a database that doubles doesn't double the invoice. BYO sending keys on Pro and above cover Amazon SES, Resend, Postmark, Mailgun, and SendGrid, which leaves the sending in your own provider account. Nitrosend specializes in email: it's the sending side, and the deal record stays wherever you already keep it.

None of that decides what to say to a past client in year seven, and no tool knows which of your eight hundred contacts wants a phone call instead. The judgment is the job. The repetition around it isn't.

Start with the send you'd otherwise forget.

A real estate database has to stay deliverable, opted in, and welcome across the decade between one transaction and the next, and it's run by somebody whose actual job is selling houses. That's the case for making it a command instead of an evening. Our free tier is 8,000 emails to start and then 500 a month, with unlimited contacts and the MCP server, API, and CLI switched on. If you already work through an agent, hand it this month's anniversary pull and see what comes back.

Related reading

• Email Deliverability Best Practices That Actually Matter
• Email Authentication Protocols: SPF, DKIM, DMARC, and BIMI
• How Do I Authenticate My Email? The Setup Steps

Sources

• NAR, 2025 Profile of Home Buyers and Sellers: the median 11-year tenure before selling, an all-time high, and the 88% of buyers and 91% of sellers who worked with an agent.
• Google, Email sender guidelines: SPF or DKIM and a Postmaster Tools spam rate under 0.3% for every sender, and SPF and DKIM and DMARC plus one-click unsubscribe on marketing and subscribed messages above 5,000 messages a day.
• Google, Email sender guidelines FAQ: the 5,000 count taken per primary domain over 24 hours with subdomains rolling up, and the recommendation that unsubscribes be fulfilled within about two days.

Common questions