Tinystores — for Tobias

Zürich · May 2026

Dear Tobias,

Since we had lunch on the 19th I've been building you something in the evenings. It is a small piece of software that scores any Swiss commercial-lease address for how well it would work as a Tiny Fish store — using the same kind of signals your intuition already uses, but at a scale a brain cannot.

This page is the first version. It is still rough at the edges, and it is private — just for you. The point is not the page; the point is to show you a direction and ask if you want me to push it further.

Jonathan

What this is, in eight questions

What is this?

A scoring tool for choosing the next Tiny Fish store. You give it a Swiss commercial-lease address; it returns a score from 0 to 100, with the reasons behind that score.

The reasons are the things that actually drive a small fresh-sushi point-of-sale: how many people walk past, how wealthy the catchment is, what the lease costs, how easy the location is to reach by transit, what direct competition exists, and now also whether the place can be reached by train if you ever want to serve cities the kitchen van cannot.

Why does this matter?

Today the next-store decision is mostly intuition plus the broker's book. That works at fourteen stores. At thirty — and especially at the international franchise scale you described to me — picking a location should be a model, not a vibe.

Every lease is a multi-year commitment in a market where rent between two seemingly similar blocks can vary three-to-one and weekday foot-traffic ten-to-one. The wrong store doesn't kill anyone. But it locks up management attention and chips at the brand discipline that makes Tiny Fish credible.

What kind of data does it look at?

Only public, freely available data — no broker fees, no expensive panels. The categories are:

How many pedestrians walk past the address. How wealthy the neighborhood is, including a specific household-income figure that captures wealthy Swiss districts cleanly. What the building and the street actually look like, read automatically from a recent street-level photograph. What the restaurants around the address charge for a meal. What commercial rent costs in that postcode. How fast you can drive from Altstetten. How far you can walk in five minutes. How busy the nearest train and tram stops are. Whether you can deliver to that city by passenger train if you wanted to.

Plus a few curated lists I built by hand — Swiss private banks, top employers, universal luxury brands — used to detect when one of those is within walking distance of the candidate.

How does it know how wealthy a neighborhood is?

Six different signals are combined: how many luxury shops are nearby, how many private banks are nearby, whether any premium employer — UBS, McKinsey, Pictet, Goldman Sachs — has its headquarters within walking distance, what the surrounding restaurants charge for lunch, the official household income for the district, and the automatic read of a recent street photograph.

The sweet spot is not pure luxury. A street with only Patek shops and no offices is wrong too. The sweet spot is wealthy people with a lunch break — professional districts with serious money and serious meal-break schedules. Bleicherweg. Bern Spitalgasse. Stadelhofen.

Does the tool agree with your existing stores?

Mostly yes — and the disagreements are useful.

Running the same unbiased tool on all fourteen existing stores ranks Löwenstrasse, Bern Spitalgasse, Bleicherweg and Stadelhofen at the top — exactly where your customer review counts put them. The tool's ranking agrees with your customers' ranking strongly.

One existing store comes out at the bottom. The tool says it was a stretch. I take that as the honest version of validation. A tool that only ever agrees with the chooser is useless; this one earns the right to be trusted by also surfacing the picks that are weaker.

What is the most interesting finding so far?

Bleicherweg — your flagship by review count — only ranked in the middle of the pack until two signals were added: the household income at the Zurich-district level, and an automated analysis of an actual recent photograph of Bleicherweg from the street.

The photograph analysis came back as professional lunch crowd, modern premium building, UBS signage detected, affluence 85/100. Once those two signals were in, Bleicherweg climbed to second of the existing fourteen — right behind Löwenstrasse, which sits on top of Switzerland's busiest station. That is exactly the correction the model is designed to make.

What about cities the van can't reach?

This is the real strategic question, and the answer changed my mind during the build.

With one Altstetten kitchen and a three-hour driving radius, the model says full national coverage caps around eighteen to twenty stores. The Romandie and Tessin are mostly outside the window.

If Tiny Fish commits to rail-based delivery — passenger trains plus insulated cold-chain boxes plus a local bike-courier handover at the destination station — then Geneva, Lausanne, Basel, Bern and Lugano all become viable. The thirty-store plan stays on track from one kitchen.

I wrote you a one-page brief on this. It compares three options: status-quo van, second kitchen at around one million capex, or a thirty-to-ninety-day rail pilot at around thirty thousand. The recommendation is the pilot. Happy to send the brief.

Can this scale to the world?

Yes — and that is the part I'm most confident about.

The Swiss build is the proof. The same engine can score Munich in 2030, Paris in 2032, Manchester in 2034 without redesign. Luxury brands are luxury everywhere. Restaurant pricing works in any currency. Public mapping data and government statistics exist in every European country. Recent street photography is available globally.

The signal sources are international by construction. The only piece that requires local work per country is the last-mile delivery partner network — and even there the playbook transfers. The hard part is not the model; the hard part is finding the local courier in each city. Solved once per country, the rest comes for free.

Where Tiny Fish is today

The fourteen stores, numbered, plus the SV Group fridges and the Altstetten kitchen. Click any pin to open it on Google Maps. This is the footprint the tool was honest-checked against.

Numbered store (1-14) SV Group Fridge Altstetten kitchen

Your fourteen stores, scored by the same tool

Same model, applied to the stores you already chose. No special weighting, no preferential treatment. The tool's ranking lines up strongly with how often your customers actually leave reviews.

loading…

Top candidates the tool is converging on

Score is the model's composite 0-100. Cities with a yellow score (50-65) are mid-tier. Red (under 50) means a hard filter penalty kicked in — usually either the kitchen is too far by van, or the candidate sits inside another existing store's catchment.

loading…

How this was built

This dashboard, the scoring engine behind it, and the validation against your fourteen existing stores took me four evenings.

I use a custom scoring engine I wrote myself in Python, a vision model from a leading AI lab to read street-level photographs, the public Swiss federal data infrastructure, open global mapping data, and automated lease-listing scraping. Everything runs on free or near-free public services plus a few thoughtful paid tools I already operate. Hosting is on Cloudflare's edge network at zero marginal cost.

The pace is the point. Five years ago, the same build would have taken months and a small team. The combination of modern AI tools, the maturity of public data, and a clear understanding of what to ignore makes the four-evening version possible. It's the same leverage that will compound across every product decision in front of Tiny Fish for the next decade.

Who I am, and what I'm considering

I'm a software developer and data scientist, fully self-taught over fifteen years, currently running my own software studio. I lean heavily on every modern AI tool — vision models, code agents, large-language-model reasoning — to build at a pace that wasn't possible five years ago.

I'm considering offering you my help, but only if our vision, our philosophy and our incentives align. What I have in mind is a specific shape: an advisory role plus software-for-equity. Not consulting hours. Not a service contract. A real seat at the table, and the code that compounds with the business.

If this — and the much more I haven't written here yet — could be of interest, I'd be happy to talk seriously. If not, no harm done. Tiny Fish is yours; you decide.

Jonathan

This page will cease to exist on 1 August 2026. If you want to keep it past then — or replace it with something more — just say the word.