I sell at local pop-up markets through my brand, Upcycled Supply Co. And one of the most frustrating parts of being a vendor is finding markets to apply to.
They're scattered across Instagram. Deadlines are buried in captions. There's no central place to search, filter, or track applications. So I built one.
What is STOA Markets?
STOA Markets is a two-sided marketplace for discovering and managing local pop-up markets.
- Vendors can discover and apply to local markets in one place
- Organizers can list markets and manage vendor applications through structured workflows
The initial focus was on the vendor side — a discovery MVP to validate whether centralized listings would drive real demand.
The Problem I Was Solving
Vendor pain:
- Markets scattered across Instagram with no searchable database
- Application deadlines easy to miss
- No way to compare fees, dates, or requirements across events
Organizer pain:
- Manual application review with no structured tooling
- Vendor mix balancing done in spreadsheets
- Waitlists managed through DMs or Google Forms
- No operational dashboard
The opportunity: centralized infrastructure that serves both sides.
Initial Hypothesis
If we centralize listings, vendors will return to discover and apply to markets.
I started with vendors because the pain was acute and visible, the MVP was faster to build, and structured listings would create leverage for the organizer side later.
I deliberately did not build application workflow management, vendor mix tooling, waitlist automation, or organizer dashboards in v1. Scope discipline was the only way to ship fast enough to learn anything.
How I Built It
Stack: Airtable (database) + Softr (front-end)
What I built:
- Structured market directory with 100+ listings
- Searchable, filterable discovery experience
- Organizer submission form to add new markets
- Vendor-facing browse and apply flow
Where AI made the difference:
Building this fast required using AI as a force multiplier throughout:
- Used Airtable's AI agent to generate structured database schemas from messy, unformatted market data
- Used LLMs to extract event details from unstructured Instagram posts and convert them into normalized, machine-readable fields
- Reduced manual database design and listing entry time by roughly 70%
This wasn't just a productivity trick. Converting freeform social content into structured data was the actual product — it's what enabled filtering, search, and discovery. And it surfaced where AI could add value in future versions: auto-tagging vendors by category, clustering applications to reduce organizer review time, recommending vendor mix based on past event composition.
Early Results
Over roughly 4 months, with zero paid acquisition:
- 224 active users
- 1.8K tracked events
- 31 seconds average engagement
Vendors reported discovering and applying to markets through STOA. The hypothesis had signal.
But retention was low. Discovery generated curiosity — it didn't create habit.
The Core Constraint
Here's what I got wrong.
I assumed that centralizing listings would be enough to pull organizers onto the platform over time. It wasn't.
- Organizer onboarding required manual submission from me
- There was no self-serve dashboard for organizers to manage their listings
- Discovery alone was insufficient to overcome organizer switching costs
Key insight: Without operational value for organizers, supply growth is slow and defensible scale is unlikely. Vendors come for the listings. If organizers don't have a reason to keep listings fresh and accurate, the whole thing stalls.
What I'd Do Differently
If I were rebuilding today, I'd flip the wedge entirely.
Instead of starting with vendor discovery, I'd build an organizer operating system first:
- Self-serve organizer accounts
- Centralized vendor application dashboard
- Waitlist automation
- Vendor category auto-tagging
- Vendor mix visualization
Why: Organizers control supply. If the product makes their job meaningfully easier, they bring their markets — and vendor discovery becomes a pull effect rather than something you have to manufacture.
Monetization: Paid organizer subscription with a free vendor discovery layer.
Primary activation metric: 80% of newly listed markets self-managed by organizers within 30 days of onboarding.
What I Actually Learned
I underestimated organizer switching costs and overestimated the pull of discovery alone.
The MVP worked as a validation tool. It confirmed that vendors have real pain, that centralized listings generate organic interest, and that the problem is worth solving. But it also showed clearly that vendor-side demand alone isn't enough to bootstrap a two-sided marketplace.
The decision to pause rather than keep building was deliberate. Scaling before solving the organizer wedge would have meant spending more time on a structure that couldn't sustain itself.
v2 will start on the organizer side. When organizers have a reason to be here, vendors will follow.