🎓 New discovery: we can predict if a restaurant will stay in business - 3 years in advance - based on the photos that people post about it. In comparison, reviews predict that only 1 year in advance. 🔍 How? Researchers analyzed 755,758 Yelp photos of 17,719 restaurants. People mainly post photos in three categories: - Food - Interior (e.g. ambiance) - Exterior (e.g. the entrance) Which of these predicts a restaurant’s success? ➡️ Food photos 🍛🍜🥘 The higher the proportion of food photos a restaurant receives, the more likely they are to survive and stay open. 🧠 Why? People post photos of food when it matches their expectations. And food is the most critical aspect of a restaurant. If people don’t post food photos, it probably means they didn’t like it. Nice decor alone won’t cut it, no matter how many photos people post of it. *Of course, these are predictions of probability, not certainty.
Also an interesting element to sniff out listings with fake reviews: great ratings and high volume but disproportionate volume of photos and content of them too little of food to be real? Great gage for what is often gamed. It's a constant battle and one Yelp specifically has done lots to creatively correct for.
Very cool, but I wonder if this is a generalizable rule or if it's only relevant to Yelp data and Yelp users at a particular point in time
Genuinely interesting!
Excellent explanation!
🎓 Science-based marketing insights | Founder @ Ariyh | Ex-Google
1yExcellent research by Mengxia Zhang and Lan Luo: https://doi.org/10.1287/mnsc.2022.4359