Journal
What we've decided, built and got wrong, written down. Build logs from our own ventures, teardowns, and the method behind every service. Each piece says where its evidence comes from: our method, our own ventures, a teardown, or observation.
All articles
Ideate: how we find out whether a problem is real before anything gets built
Stage one of Stuck-to-Scale, from our own ventures: the evidence we look for before anything is built, and what makes us stop.
Define: how we decide exactly what to build, for whom, and what to leave out
Stage two, from our own ventures: turning a real problem into a buyer, a first release, a written list of what's cut, and a stop line.
Build: how we turn an agreed scope into a product that works for real users
Stage three, from the AI Product Build and our own ventures: real data before architecture, rules before models, and tests before launch.
PMF: why paying customers aren't product-market fit, and what we measure instead
Stage four, from otlo's own label: why paying customers prove willingness to pay but not fit, and the evidence we wait for instead.
Market: how we make a working product reachable, repeatably, by buyers and by AI search
Stage five, from our own website and ventures: a position that names the buyer, launches packaged every time, and a product machines can describe.
Grow: why the last stage is Grow, not scale, and what a product needs to compound
Stage six, and the one our ventures haven't reached yet: why we named it Grow, and what a product needs to compound on its own.
How to evaluate a venture studio before you back one
Stage labels, kill discipline, ownership and room for founders: the checks that matter to experienced investors, run on our own ventures first.
Is it worth building? How to decide before you pay for the build
Four things to write down before any code, six questions to answer, and the cheapest verdict you'll ever get: don't build it.
Teardown: our old website, read the way buyers and AI search read it
Mislabelled work, overstated stages and retired offers, found by reading our own site literally, and the rules that replaced them.
What your family business has that no AI startup can buy
Customers, data, distribution and trust: how to turn what the business already holds into a venture, and protect it from the core.
White-label AI product work: how agencies say yes without losing the client
Say yes to AI product work you can't staff, keep the client and the relationship, and scope it so the margin survives.
Your AI-built prototype works. Here's what breaks when real users arrive
Messy data, permissions, costs and inputs nobody tested: the seven checks to run before launch, and when to start again.
Nothing matches that search.
Every article, at its stage.
Each article sits at the Stuck-to-Scale stage where its question first comes up. Six stages, one question each.
Ideate
Is there a real problem worth solving here?
Define
What exactly are we building, for whom, and what are we not building?
Build
Does the thing exist, and does it work?
PMF
Do the right people want it badly enough to keep using it?
Market
Can we reach them repeatably, and will they pay?
Grow
Does it compound without us pushing it?
Start from where you are.
Not sure what to build yet
Is it worth building? How to decide before you pay for the buildThen: Venture Diagnosis
Launching a product
Your AI-built prototype works. Here's what breaks when real users arriveThen: AI Product Build
Starting a venture beside your business
What your family business has that no AI startup can buyThen: AI Venture Build
Making the product look and sell like it works
Teardown: our old website, read the way buyers and AI search read itThen: AI Creative Build
Running an agency
White-label AI product work: how agencies say yes without losing the clientThen: Partners
Weighing a venture studio
How to evaluate a venture studio before you back oneThen: Our ventures