May 2026 · 6 min read
I've been a fan of Teresa Torres's Opportunity Solution Tree for years. It has given me focus and direction when I had a cloud of notes and feedback. If you haven't come across it yet: it's a visual framework for continuous product discovery that connects a desired outcome to the opportunities your customers have, the solutions you might build, and the experiments you'd run to validate them.
Every team I've worked on has eventually ended up building their OST in Miro. And often, someone creates a beautiful, colour-coded tree on a Monday. By Thursday it's a mess of misaligned sticky notes, orphaned nodes, and a canvas that requires three minutes of zooming and panning just to find the outcome you defined two weeks ago.
Miro is a great whiteboarding tool. But it's a blank canvas, it has no opinion about your OST. It doesn't enforce the hierarchy. It doesn't know that an experiment should hang off a solution, not float in space. It doesn't stop you adding a fourth level of nodes that Teresa Torres would raise an eyebrow at. The structure only exists if you maintain it manually, and in a fast-moving team, that discipline evaporates fast.
There are also the softer frustrations. Miro is a collaboration tool, which means it's optimised for teams, boards, and workspaces. If you want to sketch out a personal OST during a discovery session, or share a clean read-only link with a stakeholder, you're navigating a tool that wasn't really built for that. You're also paying for it.
I wanted something opinionated. A tool that knows the four layers of an OST: Outcome, Opportunity, Solution, Experiment, and enforces them, so I can focus on thinking rather than formatting. I wanted to be able to open it in a browser tab, add a node in two seconds, and export the whole thing as a clean Markdown file I can drop into a doc or a Notion page.
I also wanted it to live on my own site, because I've been on a bit of a journey lately of building the tools I use and writing about what I learn.
The OST builder is a single HTML file. No login, no workspace, no monthly fee. You open it and you get a blank tree with four node types, colour-coded, clearly labelled, and structured so the hierarchy is always visible.
You click + Outcome to define your goal. Then you add Opportunities beneath it, the customer needs or pain points you've identified through research. Below those go Solutions, and below those, Experiments. The tool doesn't let the layers bleed into each other: you can drag-reorder nodes within a level, but the four-level structure is fixed (that constraint is the point).
When you're done, or mid-session, you can export the whole tree as Markdown, CSV, or JSON. The Markdown export is particularly useful, it produces a clean numbered list per layer that you can paste straight into a doc, a slide, or a Confluence page.
A few things I deliberately left out: there's no collaboration, and no bells. This is a thinking tool, not a project management system. If you need a source of truth your whole team edits, Miro or FigJam might still be the right answer. But for individual discovery work, structured thinking sessions, or just getting your head around a problem space, I think lightweight wins.
Your tree is saved automatically in your browser, so it'll be there when you come back, no account needed. One thing worth knowing: the save is per browser, per device. Your tree lives in your Chrome on your Mac; a colleague opening the same URL gets a clean slate and their own private tree. It's not a shared whiteboard, it's a personal thinking space. If you switch to a different browser or device, you'd start fresh there too. For syncing across devices you'd need accounts and a backend, which felt like the wrong trade-off for a tool meant to be this simple.
Here's how I'd actually run a 45-minute discovery session with this tool.
Start with the Outcome. Be specific: "increase weekly active learners by 20% in Q3" is an outcome. "Improve engagement" is not. The constraint of one outcome node forces the focus that most OST sessions lack.
Then spend the bulk of your time on Opportunities. This is where Teresa Torres's framework is at its most powerful, and where most teams rush. Opportunities are not features. They are customer needs, pain points, or desires, grounded in something you've heard or observed. "Users lose their streak motivation after missing one day" is an opportunity. "Build a streak recovery feature" is already a solution. Add as many as feel real, you can always prune.
Solutions come next. For each opportunity you care about, what are the plausible ways you might address it? Again, resist the urge to jump to your favourite idea. Two or three solutions per opportunity is healthy. One is usually a sign you've stopped exploring.
Finally, Experiments. What's the smallest thing you could do to learn whether a solution is worth building? A fake door test, a prototype, a customer interview with a specific hypothesis, each one goes here. The discipline of writing these down before you build anything is the whole point of the framework.
When you're done, hit Export → Markdown, and you have a structured artefact you can share, review, or drop into your team's documentation in about three seconds.
The thing I'm most interested in exploring next is an AI coaching layer. The idea would be: you describe your outcome in plain language, and the tool suggests opportunities you might have missed, based on what's known about similar product spaces. It'll act as a prompt when you're stuck or when you've been staring at the same three opportunities forever.
I've already built a version of this in a Lovable prototype. It works well enough that I'm thinking about how to bring it to this tool without making the whole thing feel like it's doing the discovery for you.
The tool is live at helena.cool/ost.html. It works on desktop, it's free, and it doesn't know your email address. If you use it and have thoughts, I'd genuinely love to hear them: hi@helena.cool.
And if you haven't read Teresa Torres's book Continuous Discovery Habits, it's the best thing I've read on product discovery. Everything I've built here is downstream of her thinking.
Powered by unicorn magic and fairy dust