Nine Companies, Four Exits, and the Only Filter That Matters

I have started or acquired nine companies. Four exits. The rest are either still running or taught me something expensive enough to be useful.

The first real business was Simple Truth Chiropractic. I built it from an empty room into a thriving practice in Grand Rapids and sold it in 2017. At the same time I was starting agencies out of the basement of that clinic. Grand Apps, which became an exit in 2021. FireEye Design. ClinicVitamins. StopFidgeting, an Amazon brand making sensory toys for autistic children, which I grew and exited in 2013. Later Signal Profits and the trading education work. Now BNGR and a few ecom brands. I even bought an HVAC company, Air Titans, grew it from $500,000 to $4.5 million in revenue in two years with more than 2,800 customer reviews, and sold it, all while staying focused on ecom and agencies.

Source references in the timeline and recent posts including https://x.com/JacobCanfield/status/2029939413740900532

Ideas are cheap. Systems are the product.

Every company taught the same core lesson in a different way. Ideas are cheap. Execution and systems are everything. You can have the best product in the world and still die from bad cash flow, bad hiring, or no distribution. You can have average products and win with better systems and better creative.

Michael Gerber wrote the definitive book on this thirty years ago, The E-Myth, and the core insight still kills more businesses than any recession: most owners work in the business instead of on it, and so they never actually build a business at all. They build a job with overhead. I lived both sides. The clinic only became sellable when the systems ran without my hands. The companies that stayed dependent on me stayed small exactly as long as they stayed dependent on me. There were no exceptions, and in twenty years of watching other founders I have still never seen one.

The other pattern the graveyard taught me: businesses rarely die of one dramatic wound. The U.S. Bureau of Labor Statistics numbers have been stable for decades, roughly half of businesses are gone within five years, and the autopsies are boring. Cash flow timing. A hire that should have happened six months earlier. Distribution that never got built because the product was so lovable it was supposed to sell itself. Nothing glamorous. Boring wounds, left untreated, in businesses run from the founder's head instead of from written systems.

The filter

BNGR is the current expression of the lesson. We ship over a thousand statics, UGC, and AI animation ads a month. The only scoreboard is verifiable performance. Tracked spend. Real results. No hype. No borrowed credibility. Just shipped work and audited numbers. That filter came from years of watching companies talk a big game and then fail to deliver.

I built an audience of nearly a million on a platform I did not control, and it taught the same lesson from a different angle. Owned audiences and owned systems beat rented ones. Everything I build now has that in the DNA. The work has to stand on its own even if the distribution channel disappears tomorrow morning without a warning email. Because sometimes it does.

Tuition, itemized

Failure is tuition. I paid it in full. Domains bought for businesses that never launched. Courses that sat unopened. Partnerships that looked good and then required more energy than they returned. Each one left a clear invoice. The only way to make the tuition useful is to extract the system that would have prevented it next time.

The businesses that lasted were the ones where I treated them like real companies from day one. Clear offers. Clear metrics. Clear processes. Taxes planned for. Journals kept. Goals written. The ones that stayed hobbies died. I have written a whole separate piece on that rule because it deserves one.

Run your own audit

If you are building something right now, apply the same filter I now apply to everything, including acquisition targets and my own new ideas. Is the performance verifiable, can you show the numbers without storytelling? Do you own the system, or are you renting it from a platform, a partner, or a person who can walk? And if you disappeared for two weeks, would it still run? Those three questions will save you years. They cost me about fifteen to learn.