Market Update Email Prompt

What this prompt does

A market update email prompt turns the numbers already sitting in your MLS report into the monthly email your database opens — median price, homes sold, days on market, translated into sentences a first-time buyer's parent can repeat at dinner. You supply the data; the prompt supplies the structure and the plain language. It never supplies the numbers.

That last rule is the whole design. AI models will happily invent a plausible median price for any zip code you name, and a wrong number in a market update doesn't just embarrass you — it follows you into the listing appointment three months later. So this prompt works from your pasted figures only, and where a number is missing it writes a [VERIFY] placeholder instead of guessing.

The newsletter intro prompt is this page's closest neighbour, and the division of labor is clean: that one writes the opening hook that pulls a reader into any newsletter; this one writes the market update itself, from the data out. The follow-up email prompt is the other cousin — a one-to-one message that needs a value piece to justify sending. This email, sent to the whole database, is the value piece.

The prompt

Write a market update email for my real estate contact database, using ONLY the numbers I give you below.

**My market:** [CITY OR NEIGHBORHOOD + PROPERTY TYPE — e.g. "single-family homes in Maple Grove"]
**This month's numbers:** [PASTE FROM YOUR MLS REPORT — median sale price, homes sold, active listings, average days on market, list-to-sale price ratio]
**Comparison period:** [THE SAME NUMBERS FOR LAST MONTH OR THIS MONTH LAST YEAR]
**One thing I saw on the ground:** [SOMETHING THE NUMBERS DON'T SHOW — e.g. "three of my last four listings under $500K got multiple offers"]
**Data source and month:** [e.g. "Northstar MLS, August 2026"]

Rules:
- 200 words maximum, greeting to sign-off.
- Use only the numbers I provided. If a figure you would normally include is missing, insert [VERIFY: what's needed] — never estimate it.
- Pair every statistic with a plain-language translation in the same breath (next to days on market: "homes are taking about a week longer to sell than they were in spring").
- Describe what happened this month. Do not predict what happens next, and do not tell readers whether they should buy or sell now — if a sentence drifts into advice or a forecast, cut it.
- Include my on-the-ground observation as one sentence, clearly framed as what I'm seeing, not as a statistic.
- Cite the data source and month in a short line at the bottom.
- Close with an invitation to reply with questions — no pitch, no urgency.
- Give me 3 subject line options under 50 characters, no exclamation marks.

How to use it

1. Pull the numbers before you open the chat. The five figures in the prompt come straight off your MLS's monthly market report — the one your board already emails you. Two minutes of copying beats any amount of prompt cleverness, because the model can structure data brilliantly and source it not at all.
2. Give it the comparison period. "Median price $487,000" means nothing on its own; "$487,000, down from $492,000 in July" is a story. Month-over-month catches turning points, year-over-year smooths out seasonality — pick one and keep it consistent so your readers learn how to read you.
3. Don't skip the on-the-ground line. The observation only you could make (what happened at your last three showings) is the reason this email comes from you and not from a portal's automated report. It's also the line replies quote back.
4. Check the [VERIFY] flags, then delete them. Each one marks a spot where the model wanted a number you didn't give it. Either add the real figure or cut the sentence. An email with one solid stat beats an email with four shaky ones.

Example output

Inputs: single-family homes in Maple Grove; August 2026 — median sale price $487,000, 41 homes sold, 112 active listings, 23 average days on market, 99.1% list-to-sale ratio; July comparison — $492,000, 47 sold, 17 days on market; observation: three of the last four listings under $500K drew multiple offers; source: Northstar MLS.

Subject: Maple Grove, by the numbers — August

Hi [First name],

Here's what actually happened in Maple Grove last month.

The median single-family sale price came in at $487,000, a touch below July's $492,000 — a 1% move, not a shift. 41 homes sold, down from 47, and the average sale took 23 days against 17 in July, so homes are sitting about a week longer than they were at the start of summer.

Sellers are still getting 99.1% of asking on average. And from my own files: three of my last four listings under $500K drew multiple offers, so the entry end of the market hasn't cooled the way the topline suggests.

Numbers from Northstar MLS, August 2026.

If any of this raises a question about your own place, just hit reply — I read every one.

Notice what isn't in it: no "the market is expected to," no "now is a great time to." What happened, what it means in plain words, one thing only this agent could see, and a source line — that's the entire formula.

Variations

The quarterly version

"This is a quarterly update, not monthly. Here are three months of numbers — summarize the quarter's arc in one extra sentence at the top, still no predictions, still 250 words maximum."

Split by audience

Run the prompt twice with the same data — once adding "my list is mostly homeowners who bought in the last decade; lead with what the numbers say about the value of what they own," once with "my list is mostly renters and first-time buyers; lead with days on market and the entry price point." Same month, two emails, each one relevant. The email subject line prompt earns its keep here, because the subject that opens a homeowner isn't the one that opens a first-timer.

The new-neighborhood teaser

"I'm expanding into [NEIGHBORHOOD] and this is my first update for that list. Add one opening sentence that says I'm new to covering this area — establish that honestly rather than implying a track record I don't have."

Common pitfalls

• Don't let the model fill a data gap. If you didn't paste the condo numbers, the email doesn't mention condos. Every figure the model produces unprompted is a guess wearing a stat's clothing, and the [VERIFY] rule only protects you if you leave it switched on.
• Don't slide into forecasting. "Inventory is up 12%" is reporting; "expect prices to soften this fall" is a prediction your readers may act on — and the version of you that shows up at a listing appointment in January doesn't want to defend it. Describing the month is the job; readers draw their own conclusions.
• Don't bury the stats in agent-speak. Absorption rate, months of inventory, list-to-sale ratio — your readers didn't take the licensing exam. Nielsen Norman Group's usability work found that even highly educated readers prefer plain, scannable language over specialist phrasing, and your database skews busy, not dim.
• Don't send twelve identical months in a row. If nothing moved, say that in the first line — "August looked almost exactly like July, which is itself worth knowing" — a flat month stated plainly builds more trust than a manufactured trend.

Who uses this prompt

• Solo agents: the monthly email that used to eat a Sunday evening now takes fifteen minutes — pull the report, paste the numbers, fact-check the draft. More in prompts for real estate agents.
• Small teams: one person pulls the MLS data, every agent runs the same prompt with their own on-the-ground line, and the whole team's updates go out the same week without sounding cloned.
• New agents: no sales history yet, but a clean, sourced market summary every month makes you the contact who knows the numbers — a reputation you can build before your first closing.

ChatGPT, Claude, and Gemini all run this prompt unmodified. Its two working parts come straight from Anthropic's prompting best practices: give the model the context it cannot infer (your numbers, your market, your observation), and constrain the output to the job — in this case, constrain it hard enough that the model's fluency can't outrun your data. The email that results is short, sourced, and yours. That's the point.