We Measured 695 Luxury Real Estate Videos in Marin. Here's What Actually Got Watched.
Everyone in Marin real estate has an opinion about what content works. Post more Reels. Hire a better photographer. Get on video. Be consistent.
We got tired of the opinions, so we measured it instead.
Over several weeks we pulled the exact view, like and comment counts on 695 videos across large luxury real estate channels in Marin County. We read seven competitor websites down to the code. We went through Instagram accounts post by post. And we asked the AI assistants we could test the questions a Marin buyer actually asks before they make an offer, to see which agents get named.
Some of it confirmed what you'd guess. Some of it was genuinely strange. Here is what came back.
A note on names: no team, brokerage or account is identified anywhere in this article, and figures that could point at an individual account are rounded or left out on purpose. Marin is a small market. The patterns are the useful part, and the patterns hold regardless of who produced them.
The short version
Across Marin luxury real estate, the most expensive content is the least watched. Three teams have published 538 professionally produced videos between them, and all of that work together has earned roughly 181,000 lifetime views. One agent posted a single video of themselves talking, with no property in it at all, and it did more than 30,000 views on its own.
The pattern repeats on Instagram, where landscape footage outperforms styled interiors by more than ten times inside a single account. And when we asked AI assistants a high stakes Marin buying question, no Marin agent and no Marin real estate site appeared in the answer at all.
The county's content problem is not a production problem. It's a funnel problem.
What we measured, and how
Every figure from our own study traces to public data we recorded ourselves, and the one outside statistic in here is cited:
- 695 videos across large luxury real estate channels in the county, with exact view, like and comment counts on each
- Seven competitor websites, read down to the code, including their structured data, blog cadence and internal linking
- Instagram accounts reviewed post by post, not sampled
- Live AI answer tests across six buyer questions, recording verbatim which businesses got named and which pages the assistants pointed at
We deliberately excluded the widely circulated real estate AI statistics you'll see quoted in vendor decks. Most of them trace back to nothing. If we couldn't verify a number ourselves, it isn't in here.
Finding one: the most produced category is the least watched
Three Marin teams have published 538 videos between them. Professional photography, drone work, the full production, over a period of years.
Together, all of it has earned roughly 181,000 lifetime views.
Then one agent in the same market posted a single video of themselves talking to camera, explaining what was happening in the market. No property in it at all. The thumbnail looks like a screenshot. That one video has done more than 30,000 views.
One person, no crew, reaching a meaningful fraction of what three teams produced across years of listing tours.
That's one example rather than a trend, and we'd say so plainly. But it points the same direction as everything else we measured: nobody is waiting for another property tour. There are already more of them than anyone can watch. What people can't get anywhere else is what you know about this market, said out loud, by the person who knows it.
Finding two: place beats property, by more than ten times
Inside a single luxury account we reviewed, one clip of the Marin landscape on a summer morning outperformed that same account's styled interiors by more than ten times.
Same account. Same standard. Same month. The most expensive thing they made reached a fraction of the audience the cheapest thing did.
This is the finding most worth acting on, because it costs nothing to apply. Nobody follows a Marin real estate account because they want to see houses. They follow it because they want to live here, or they already do and they like being reminded why. Sell the place and the houses come with it.
Finding three: the post that beat the photography was a mechanic
Here's the pattern almost nobody in the county is using.
One of the strongest posts we measured wasn't a photograph at all. It was an off market listing whose caption asked people to message one specific word for the details, and the comments filled up with people doing exactly that.
Comments and direct messages are what these platforms reward hardest right now. That post didn't win because it was prettier. It won because it gave people a reason to respond, and then made responding take one word instead of a phone call.
Compare that to the standard Marin real estate caption, which ends with "call me" or "DM for details." That isn't a reason. It's a request. The difference between the two is most of the gap in this study.
Finding four: in April, pretty became a distribution problem
If your listing posts stopped reaching new people this spring, there's a specific reason.
On 30 April 2026, Instagram extended its unoriginal content penalty from video to photos and carousels. Templated and reposted content stops being eligible for algorithmic recommendation. Your existing followers still see it. New people stop seeing it entirely.
The originality bar the platform published is unique text, creative edits and voiceover, with a perspective added.
Read that against a typical luxury listing post: a silent montage on a trending sound, with a templated caption. It has none of the three.
Original footage still counts, and beautiful still matters. But polish used to be neutral and now it's a distribution question, and it's written into the rules rather than left to taste.
Finding five: we asked AI a Marin buying question and no Marin agent appeared
This is the finding we didn't expect, and it's the one with the most money attached.
There's a question Marin buyers ask constantly before they make an offer: whether a home with a particular mailing address actually sits in a particular school district. In this county the two don't always match, and the answer can move what a house is worth by a serious margin.
So we asked the AI assistants.
The answer came back with a high school in Kansas. Then a school in Ohio.
Not one Marin real estate site appeared. Not one Marin agent. On the single highest stakes local question in the market, the assistants had nothing local to read, so they reached for whatever matched the words.
Meanwhile 45% of consumers now ask AI for local business recommendations, up from 6% a year earlier, according to BrightLocal's 2026 research.
This is what answer engine optimization is actually for. Not ranking, but being the source the machine quotes when it answers. Right now, on the questions that decide Marin transactions, that source doesn't exist. Whoever writes those pages first will own them, and the window where that's cheap is open now.
Why the ceiling isn't talent
Here's the encouraging part of the study.
Almost every serious team we measured is stuck in the same band, a couple thousand views per post, and follower count barely moves it. Teams with far larger audiences aren't getting proportionally more reach. An account that broke out of that band did it by changing what it posted, not by spending more.
Which means the ceiling in this market isn't talent, and it isn't budget, and it isn't your photographer. It's method. And method is the only one of those you can change this month.
What we'd fix first
If we were handed a Marin luxury real estate account tomorrow, this is the order:
- Put a reason to respond in every post. One keyword, one instruction, one thing that takes a viewer three seconds. This is the cheapest change with the largest measured effect in the whole study.
- Shift the ratio toward place. Keep the listing content, and stop making it the whole feed. The landscape footage is cheaper and it travels further.
- Get the principal on camera. Not a tour, an explanation. The one explainer video in our whole dataset reached a meaningful fraction of what three teams earned across years of listing tours, and nothing else we measured points the other way.
- Answer the questions nobody has answered. School boundaries, neighborhood distinctions, the local specifics that change price. This is the AI search work, and in Marin real estate these questions are still unanswered.
- Make sure a response has somewhere to land. A reply that arrives the next morning has already lost. This is the funnel underneath the content, and it's the part that turns reach into a conversation.
Notice that four of those five cost nothing but a decision.
The offer, which is genuinely free
We'll run this exact analysis on your business. Your team, and your three closest competitors, measured the same way: what's working, what isn't, and where the opening is. Including the AI answer test for your specific market, which is the part that surprises people every time.
No charge, and you don't need to meet with us to get it.
Email [email protected] with your team name and we'll start. We'll send it back as a finished report, and the fixes in it are yours to run whether or not we ever speak.
If you'd rather talk it through than read it, you can book a conversation instead.
Tracey Bauer, who ran this study, spent 33 years on the wholesale side of luxury optical with Chanel, Tiffany & Co. and Bulgari before founding Lens on Luxury in San Rafael. What she keeps seeing in this market is beautiful work that nobody is ever asked to respond to.
That's a fixable problem, and it starts with knowing where you actually stand.
