The Idea
Roughly a decade ago, Zach Dunn dropped out of college, drove to a bank in south Georgia with his father and a business plan, and left with $250,000 to put cotton and tobacco in the ground. It held up for about two seasons. Then one ruinous year took the farm out entirely and left him with the debt still sitting on his name.
He enrolled again, this time as a business major, and sold cars to chip away at what he owed, climbing to general manager and then finance manager on the forecourt. His old college roommate, Chason Rabitaille, was working the same lot. Between test drives, the pair kept getting served the same thing on Facebook: ads for hat brands. Low cost of goods, no retail buyer to convince, no gatekeeper.
In April 2020, with the country shut down, Dunn put what was left of his savings, $15,000, into 15 hat colours, 40 patches and one industrial sewing machine, set up in a spare bedroom.
This was the beginning of Mad Hatter.
The Execution
- Circa 2016: Dunn and his father borrowed $250,000 to start a cotton and tobacco farm in rural Georgia. Two decent years, then a single bad season ended it and left him personally in the red.
- 2016 to 2020: He went back to college at Abraham Baldwin Agricultural College in Tifton and sold cars to clear the debt, moving up to finance manager at the dealership where he met Rabitaille.
- April 2020: The two launched Mad Hatter with $15,000 of Dunn's money. Rabitaille was 21 and still studying, stitching two or three hats a day by hand until volume forced the machine purchase. Everything was cut and finished to order, which is still true today.
- 2020 to 2021: They outgrew the bedroom and shifted production into a garage, funding growth almost entirely through Facebook and Instagram ads for vintage-style patch hats.
- April 2021: Apple shipped App Tracking Transparency with iOS 14.5, and the paid social maths stopped working. Margins on ads thinned out and the brand had no other way to buy customers.
- 2021 to 2022: Dunn pulled the ad spend and gambled the business on unpaid reach, learning Instagram through repeated flops before opening a TikTok account as attention moved. He also abandoned the outdoors aesthetic every competitor was copying, all antlers and bass logos, and switched to one-line joke hats with slogans like Mama Tried and I Am The Problem. Funny hats were nothing new; nobody was building a content engine around them.
- September 2023: TikTok Shop opened in the US, by which point Mad Hatter had two years of daily posting and a live affiliate network already in place. That channel now accounts for around 65% of revenue, on a platform forecast to do $23.41bn in US sales this year.
- 2025: Growth masked a problem. The accounts showed the business close to break even, so Dunn pulled the discounts on TikTok, held pricing, and traded top-line vanity for margin.
- Today: Mad Hatter runs from a 10,000 sq ft facility in Chula, Georgia, decorating every order within 24 to 72 hours off held blank stock, so a viral spike never turns into a 90-day factory wait. Its main TikTok account carries 397k followers, apparel is now roughly half the business, headcount sits at 40 to 45 across Georgia and Mississippi, and revenue is tracking to eight figures this year.
The lesson?
Most brands treat going viral as a marketing outcome. Dunn treats it as a warehouse problem. Holding blanks and decorating in-house means a spike in attention converts the same week instead of evaporating into a backorder queue, which is the difference between a moment and a business. The same logic saved him from himself in 2025, when the numbers were going up and the margin was going nowhere. Attention you cannot ship is just noise, and revenue you cannot price is just work.