Most agencies sell advice they never have to live on. The campaign ends, the report gets presented, and whether the thinking actually worked becomes somebody else’s problem. We wanted a different arrangement, so we built one. Extra Medium Labs is the product arm of this studio, the part of the business where we build and run our own products, with our own budget, on our own deadlines.
Four products are in the lab right now. Magpie, our flagship, is a live platform for local creator marketing: it connects local businesses with creators in their own market, ranked by how local their audience actually is. Hapsta, live at hapsta.co, runs small, capped live video rooms where fans hang out or play games with famous people; the celebrity sets the price and keeps 80 percent, and it’s currently in its waitlist stage. Roon turns the photos, videos, and voice notes a family already has into stories; it’s in prototype, with a waitlist open. JAWNRA, a global music discovery and chart platform that measures momentum instead of popularity, is in development.
Here’s what running them does to the client work.
It keeps the advice honest
When you market your own product, there is no client to blame and no scope to hide behind. The positioning either works in the market or it doesn’t, and the invoice for finding out arrives either way. That changes how you give advice. Recommendations get shorter and more specific, because you know exactly which ones you’d fund yourself and which ones sound good in a deck.
Every agency has opinions about what a launch needs. A lab turns those opinions into a track record you can inspect. When we tell a client their offer needs work before their identity does, that’s not a philosophy. It’s a conclusion we paid to reach.
It’s easy to be brave with somebody else’s budget.
Every lesson gets paid for twice
Client work teaches you things once, in someone else’s context. Your own products teach you the same lesson again, in yours, with your money. Pricing, onboarding, the gap between what users say and what they do, the features that demo well and die quietly: you learn these differently when the P&L has your name on it.
The transfer runs in both directions. A pattern we prove on a client engagement gets tested against our own products, where the feedback is fast and unfiltered. A mistake we make in the lab becomes a warning we can give a client with a straight face and a specific story attached. Paid for twice, used many times.
It’s a standing prototype shop
A working lab is also infrastructure. We have live products with real signups, real payment flows, real support questions, and real analytics, which means new ideas don’t have to be argued in the abstract. A landing page structure, an onboarding sequence, a pricing presentation: we can build it, put it in front of actual users on our own properties, and see what happens before a client spends anything on it.
That speed matters most at the start of an engagement, when the questions are biggest and the evidence is thinnest. Instead of bringing a client our best guess, we can often bring them something we’ve already watched succeed or fail at small scale, and the conversation starts from evidence instead of taste.
The same work, for your product
None of this makes the lab a side project. It’s the reason the studio side works the way it does. The Launch, Reinvent, and Grow engagements we run for clients use the same process we use on Magpie, Hapsta, Roon, and JAWNRA, because it’s the only process we have: settle the offer, tell the story plainly, prove it in the market, then make it look like what it is.
If you’re building something and want a team that has to follow its own advice to stay in business, that’s the arrangement we’re offering. The lab door is the same door as the studio door. Start a conversation and we’ll show you both sides of it.