I traveled to New York and Toronto, and finally got to test Savor in the real world. Last week I visited more than 20 restaurants, and it was the first time I could test the app outside the simulator and in a completely new context. Using the product in practice changes how you think a lot. Some things I thought would matter didn’t, and things I hadn’t even considered turned out to be essential.

At the start of the trip I went to an incredible Japanese restaurant in Manhattan, logged it quickly with the lock screen widget, took a few photos, and moved on. On my last day in NYC, I wanted my final meal at the restaurant I’d liked the most, so I opened the app to see which one it was and where. I could also see exactly what I’d ordered, and that gave me a feature idea: a chat where I can ask something like “In the last 5 days, which restaurant did I like the most?” and it answers with the restaurant, what I ordered, and so on, really quickly.

Another moment was when some friends started asking me about the trip, and instead of trying to remember everything from scratch and just telling them over text, I showed them a few screenshots from the app and could quickly walk through everything I’d logged about the restaurants I’d been to. That also sparked another feature idea: shareable lists of restaurants I’ve already been to or want to go to, so other people can see them even if they don’t have the app installed.

Now, something I missed: being able to give half ratings to restaurants. I went to a place that had incredible dishes and others that were only okay in the same spot, and sitting between 4 and 5 didn’t capture what I meant. I also missed being able to rate other things about the restaurant, like service and cleanliness, and rate each dish, because sometimes a restaurant can be really good but have one average dish. Having that separated would be really useful.

I tried Beli, a well-known app for real-people restaurant reviews that’s strong in NYC and other cities in places like Canada, France, London, and so on, but using it was really frustrating. The app locks the features that actually matter until you invite people and those people start using the app. I was stuck with little more than “top 10 pizza in NYC” lists when what I wanted was to find something good that was open right then near where I was.

I left Beli more frustrated than when I opened it, and with a very clear idea of what Savor will never do: never lock core functionality behind invites. If the product is good, people share because they want to, not because they’re forced to.

But the frustration with Beli also made me notice something missing in Savor. I have a great place to keep food memories, but I still can’t discover new places inside the app based on my taste. During the trip I needed both, and I had to use different apps for each.

Before the trip I mostly thought of Savor as a personal diary. I came back with a clearer, stronger conviction that Savor will be your personal food assistant. Where you log the places you’ve been, what you ate, what you thought, and over time it learns your taste and starts suggesting where to go. Not based on strangers’ reviews, but based on you and your taste.

The difference between Google Maps full of mediocre restaurants with 5 stars and what Savor can be became really clear to me. A stranger’s review has no context. Savor has your context, your history, your preferences, what you love and what you avoid.

The trip gave me enough clarity to start structuring what Savor can really be. In the next posts I’ll share how I’m thinking through each feature, defining the architecture, and building it from scratch.