Esteban Araya-Santiago

01·Bio

I’ve spent ten years building products people use on ordinary days — local news and weather for 36 stations, publishing tools at Snapchat, a wellness app for Latin American audiences, and doula-led maternity care for women having their first baby.

Twice in that decade I’ve been handed a confident premise by a room full of executives, and twice the data said they were wrong. At NBC it was “nobody reads articles on this app.” At Pacify it was “our members are no-showing.” Both rooms were about to spend real money solving a problem that didn’t exist. The most useful thing I did in either job was ask for two weeks and go look.

The other half of the work is what comes after — building the system that lets everyone else act on what’s true without waiting on an engineer. Fifty partner organizations running their own sites in twenty languages. Thirty-six newsrooms curating their own front pages inside a budget they can’t overspend. A clinical team editing the product copy themselves. Same move, three industries.

I grew up in Costa Rica and studied audiovisual production. I came to product sideways, through design and copywriting at a nonprofit where I was the only person doing any of it. That’s why I still design — and why the interface is usually where I go to find out whether an idea is any good.

02·Principles

Kill the premise before you build on it.The most expensive thing in product is a room of smart people confidently solving the wrong problem. The cheapest intervention is one person saying “give me two weeks” before anyone opens Figma.

Build the budget so you never have to say no. A constraint imposed by a person feels like a veto. The same constraint expressed as a system feels like a creative choice. Give people a real budget and they will spend it well.

AI collapsed the execution gap, not the judgment gap.A model can reproduce what a decision produced. It cannot reproduce the reasoning that produced it — unless someone wrote the reasoning down. That is most of the job now.

A finding that makes someone the villain doesn’t get acted on. Being right is the easy half. The truth has to be sayable in a room where the person it implicates is sitting.

Say what you can’t prove.Label the claim that didn’t survive verification instead of quietly dropping it. Confidence you haven’t earned is the fastest way to lose the confidence you have.