Beautiful restaurant photography does not translate to delivery platforms. Here is why, and the 5 rules I follow on every single one of my restaurants to stop the scroll.
I've reviewed over 200 restaurant listings on Uber Eats and DoorDash. The single most common mistake, by a wide margin, is photos. Not bad photos. Restaurant photos.
The problem is that restaurants hire a photographer, get beautiful images, upload them to delivery apps, and wonder why nothing moves. The photos are beautiful. But they're the wrong format for the wrong medium.
Delivery photos ≠ restaurant photos
A restaurant photo is designed to be seen printed on a menu or on an Instagram feed, large, full-bleed, in a context where the customer is already committed to being there.
A delivery photo is seen as a 380-pixel thumbnail on a phone screen, next to 19 other thumbnails, in less than 2 seconds of scrolling. The rules are completely different.
The 5 rules of delivery-first photography
Rule 1: The food fills the frame
White space is your enemy. On a small thumbnail, a photo with 40% empty plate becomes unrecognizable. Tight crop. Food taking up 80%+ of the frame.
Rule 2: Only one hero element
A burger photo should be about the burger. Not the burger, the fries, the drink, and the sauce on the side. On a thumbnail, the brain needs to identify one thing in under a second.
Rule 3: Warm colors only
Green bowls, blue plates, cold-tone backgrounds, they all read as "sad food" at thumbnail size. Warm tones (orange, red, yellow, gold) trigger appetite. This isn't opinion, it's decades of food psychology research.
Rule 4: Bold texture over fine detail
A photo of perfectly julienned vegetables doesn't work. Melty cheese, charred grill marks, visible crispy edges, glistening sauce, that's what reads at 380px.
Rule 5: No hands, no faces, no distractions
Every element that isn't the food dilutes the message. Including "lifestyle" shots of people eating. Save those for Instagram.
Take any of your current photos. Resize it to 380px wide. Look at it on your phone while walking. Can you instantly identify what it is and feel hungry? If not, the photo is failing.
Real numbers from our own restaurants
When we rebuilt the photo strategy at Tito's Kitchen in 2023, here's what happened over 45 days:
- Click-through rate: 1.9% → 5.8% (+205%)
- Conversion rate: 22% → 28% (menu aligned better with hero shots)
- Average order ranking: #52 → #7 in our delivery area
- Monthly orders: +340
We didn't change a single recipe. We didn't change prices. We didn't run a new promo. Just the photos.
AI photos vs real photography, the honest truth
Everybody's debating this right now. Here's my actual view after testing both extensively:
AI-generated food photography, done correctly, outperforms most real photography for delivery thumbnails. Not because AI is "better" in an artistic sense, but because it lets you control every variable (lighting, angle, cropping, color, ingredient visibility) without $2,000 photographer bills per product.
Real photography is still essential for your first hero banner, your website, and marketing. But for the 40+ item photos inside the menu, AI is now the pragmatic choice.
The key word is "done correctly." Bad AI photos look obviously AI. Good AI photos are indistinguishable, and they work.
The opportunity cost
Every day you're running weak photos is a day you're paying for customer impressions that don't convert. Delivery platforms send you traffic. Bad photos waste that traffic.
If your restaurant does $15,000/month on delivery and your CTR is 1.8%, doubling it to 3.6% isn't a 100% revenue jump (platform supply caps that). But a realistic 40% lift is $6,000/month extra, without spending a dollar more on ads.
That's what's on the table. Every month you wait.