The Execution Marketplace: A Three-Sided Order Book for Small Business

Every CMS is a bet about what you'll point it at.

As part of Auteurs — our guild of engineers, designers, and product folks who turn conversations into shipped product — we've been building our own. Not a WordPress skin: a real one. Server-rendered Rust, a JSX-like templating macro, Stripe checkout that never leaves the page, every asset compiled into a single binary. It's the engine behind this very site and the merch store. Point an MCP at it and one person can mass-create blocks; we've been writing skills so the AI builds them faster and breaks them less. One person can comfortably run five or six sites this way.

That's a good tool. But a tool is only worth what you aim it at — and I want to aim ours at Project FootTraffic.

Looking for the first place to point it, I found Rune and Board: a family game shop out in Hillsboro, a play on "Room and Board." The site looks like it was assembled between shifts, but underneath it's already doing a ton — real inventory, real commerce. Our CMS could make this place badass. And shops like this have the money; a busy one runs a few thousand dollars a day through the register. What they don't have is a dev. We have a CMS, and a way to run a lot of them at once.

But this post isn't about selling one shop a nicer website. It's about what the CMS becomes once you point it at a whole portfolio — and that's the upgrade I want us to build next.

The CMS We've Been Building

Start with what's real today. The CMS renders blocks. An operator describes a section in plain language, an MCP turns it into blocks against our component library, and a skill keeps the output on-spec instead of misshapen. Swap in a shop's Stripe keys and checkout works; every order flows upstream into an admin panel the owner can actually read. Add a discount code, a gift card, a loyalty loop, a post-visit text — each one becomes a reusable block the next shop inherits.

Nobody building that was describing a website. They were describing an assembly line. And an assembly line with idle capacity is just a market nobody has opened yet.

A Standing Inventory of Demand

A CMS, on its own, sells one site at a time. FootTraffic's whole thesis is that you don't sell sites — you treat a plaza like real estate and raise the value of the place. Point the CMS at that and the math changes.

Once we own a portfolio of these businesses across Portland, LA, Austin, Detroit — dozens today, hundreds as we grow — every one of them is a faucet of recurring work. A booking flow that needs wiring. A menu redesign. "Can someone fix our Google listing this week." A new-release Friday that really should have been a pre-order drop.

Each of those is a unit of work attached to a real business that already trusts us and already pays. That's not a backlog. That's demand we already own — the single hardest thing for any marketplace to manufacture.

It's a marketplace waiting to clear.

Three Sides, One Board

The model I keep drawing looks like rentahuman.ai — but pointed inward, at our own real estate. Three sides.

Side one — the businesses. The CMS already gives an owner push-button control of the easy stuff: the booking widget, the loyalty loop, the post-visit text, the discount code and gift card we'd otherwise wire up by hand. No human required.

Side two — the open network. When an owner hits something the buttons don't cover, they don't file a ticket and wait on an account manager. They vibe-submit it in plain English, it drops onto a board, and it's first come, first serve. Any operator — or AI-assisted human — in the network can claim it and ship it.

Side three — us, the house. This is what makes it ours and not just another freelancer site. We're not neutral plumbing between a buyer and a freelancer; we're a participant on our own board. We claim the brand-defining, high-margin work ourselves, set the quality floor the whole network has to clear, and capture every reusable block that falls out of a job — straight back into the CMS, so the next hundred shops inherit it.

What the Owner Touches

For the owner, none of that machinery shows. They get a console — the same CMS surface, in two modes. The easy things are buttons. The hard things are a text box.

That text box is the whole UX bet. The owner shouldn't have to know whether "pre-order Friday" is a tool toggle, a block tweak, or a two-day build. They describe the outcome they want; the system figures out who clears it.

The Board

Behind the text box is the board. This is where supply meets demand.

First come, first serve, with a lock so two operators never grab the same job. A business chip so whoever claims it has the context. A tag for the kind of work. An SLA so "this week" actually means something. And the house, sitting quietly in the same feed, claiming the jobs worth doing ourselves.

Routing Is the Whole Game

Everything I just showed you is window dressing on one hard problem: when a request comes in, where does it go? That decision — tool, open network, or house — is the core IP. Get it right and one person really can run a territory of a hundred businesses. Get it wrong and we're a worse agency with extra steps.

The classifier asks three things in order. Can a CMS tool already do this? Is it commodity work the network should clear? Or is it brand-defining enough that the house should take it and turn it into a block the next hundred shops inherit? That router is the thing I most want us to get obsessive about.

Why We're the House — and Why It Compounds

Trading on our own board is what turns regional ownership into a liquidity engine.

Every job the house claims should leave the CMS more capable than it was before. The discount-code flow we wire up once becomes a button every shop gets forever. The MCP-driven block creation and the skills we keep reaching for aren't side quests — they are the supply side's productivity, and ours. The better they get, the cheaper execution gets; the cheaper execution gets, the more businesses we can hold; the more businesses we hold, the deeper the board.

This is the Uber model fused to the agency model — which, after twenty years, is honestly all I've ever worked inside, so it's all I know how to build. I can't help myself but try to mimic it. The one difference that matters: we're not starting cold. We don't have to beg for either side. The real estate hands us the demand on day one, and we've already proven one person plus AI can stand up a real store on the supply side.

What We Upgrade Next

So here's the brick I want us to lay next — the upgrade that turns the CMS into the spine of the marketplace. Roughly in order:

• Intake parser — vibe-submit text into a structured, routable job. This is where the magic has to feel like magic.
• The board — post, claim, lock, SLA, status. Boring, load-bearing, has to be bulletproof.
• The router — the classifier that sends a job to a CMS tool, the network, or the house. The IP.
• Self-serve tiles — booking, loyalty, post-visit, discounts and gift cards, exposed as one-tap CMS blocks. The assembly line we've been sketching all along.
• Block capture — every house job ends by promoting its reusable part back into the CMS.
• Billing — subscriptions on top of the Stripe wiring the CMS already has. We don't beat Shopify's margin, and we're not trying to.

We have the CMS. We have Stripe. We have a game shop in Hillsboro that deserves better than what it's running. That's enough to lay the first brick.