Homes whose online listings showed more photos with plants in them sold for higher prices, according to an analysis of more than 40,000 US home sales. The gap held up after the researchers accounted for the size, condition and location of each house, how well it was photographed and what the listing site itself expected it to sell for.
Listings with more plant photos also drew more likes and more booked tours, and they tended to sell a little faster.
The study, “Visual Plant Imagery and Property Valuation: A Descriptive Analysis,” was written by Yuqian Chang of Ivey Business School at Western University, Nathan Fong and Maureen (Mimi) Morrin of the School of Business at Rutgers University Camden, and Ning Ye of Stockton University. It was published online in Marketing Science on October 5, 2026.
Agents rarely think about plants
The team started by asking 98 real estate agents what advice they give sellers. When asked open-ended questions about general advice and presentation tips, only 2% and 5% of them mentioned plants.
The agents were then shown 15 common tips for sellers and asked how often they recommend each one. Adding plants indoors and outdoors came last on both how often agents recommend them and how important they are. Decluttering and deep cleaning came first.
Half a million listing photos
The researchers then picked 1,000 US zip codes at random and collected every home sold in them between June and December 2019 from a major real estate website. That gave them 40,294 properties and 569,773 listing photos. The median sale price in the sample was $259,000, close to the national median for 2019.
To sort the photos, they tested 13 image classification methods against pictures labeled by people. The most accurate one was the AI model GPT-5-mini, which identified whether a photo contained a plant about 95% of the time, so they used it on the full set.
Plants turned out to be everywhere. In the 31,031 listings used for the main analysis, the average home had 18 photos, and about 12 of them showed at least one plant.
More plant photos, higher prices
Before any adjustments, the number of plant photos rose with the sale price, while the number of photos without plants did not.
The researchers then added a long list of other factors: bedrooms, bathrooms, size, age, lot size, school ratings, walkability, how green the surrounding area looks from satellites, the direction the house faces, computer-rated photo quality, cleanliness and decor, the text of the listing, the zip code, and the website’s own price estimate. Each step shrank the plant effect, but it never disappeared.
In the most complete model, each extra plant photo went with a sale price about 0.16% higher. On a home at the sample’s median price, that works out to roughly $400 per photo, by our calculation. Photos without plants showed no reliable link to price.
The pattern held in 64 different combinations of control factors, with other ways of counting the plant photos, and with other ways of measuring price.
Where plants seem to matter most
The link was stronger when the plants appeared indoors than outdoors, and stronger when the plant photos came early in the listing.
It was also strongest in areas where few listings showed plants. Those neighborhoods had lower family incomes, about $76,000 against $94,000 where plant photos were most common. Nearby homes were worth about half as much there, and listing photos were of lower quality.
The authors suggest that plants may stand out more where they are unexpected, such as in a living room, near the top of a listing, or in a neighborhood where few sellers stage with greenery. Their starting point is phytophilia theory, the idea that people are naturally drawn to plant life.
A pattern, not proof
The authors are clear that this is a descriptive study of observational data. It does not show that adding a fern to a listing will raise a home’s price, and they make no recommendations to sellers.
Some factor the data did not capture could explain both the plants and the higher price. Two extra statistical checks, double machine learning and matched comparisons, gave the same direction, but the researchers say a properly designed causal test is still needed.
The data also come from 2019 sales on one website, and the site no longer publishes the view, like and tour counts used in the study.
The researchers suggest the same question could be asked about Airbnb and other rental listings, where photos do much of the selling. The work was funded by the Social Sciences and Humanities Research Council of Canada.
Study Details:
- Title: Visual Plant Imagery and Property Valuation: A Descriptive Analysis
- Authors: Yuqian Chang (Ivey Business School, Western University), Nathan Fong (Rutgers University Camden), Ning Ye (Stockton University), Maureen (Mimi) Morrin (Rutgers University Camden)
- Journal: Marketing Science
- Publication Date: October 5, 2026 (online)
- DOI: 10.1287/mksc.2025.0145