The Borderline co-founder returns to Deconstructing Data with a blunt playbook for scaling past the idea stage, from business goggles to why your ad targeting is probably too wide.
Julie Mattson’s eight-year-old son had a business idea. A dog rinse pad, suction cups and all, that wipes a dog’s paws and dispenses a treat when he behaves. Mattson took it to Claude, and within two hours she had a manufacturing plan, a prototype sketch, and a photo of the product in action. “My kid’s a genius. We are going to make so much money,” she thought. Then she called a friend who actually knows dogs. No dog, he told her, would ever willingly step on a suction pad. “Funny, Claude never mentioned that,” Mattson said on this episode of Deconstructing Data.
That story is the whole thesis of Mattson’s second appearance on the show. She’s the co-founder of The Borderline, and she’s spent enough years watching founders build entire strategies on top of an AI that only ever tells them yes, to know exactly where the pattern breaks down.
Why does AI make founders overconfident about a bad idea?
Mattson calls it business goggles. You run an idea past Claude or ChatGPT, it comes back enthusiastic, and suddenly you can’t see the idea clearly anymore.
“You get clouded. Is there something called beer goggles with business? Business goggles. And then you start to think something’s a great idea, but that is something to be careful about.”
— Julie Mattson, Co-Founder, The Borderline
The problem, as her co-host David Finkelstein pointed out, is structural. Most AI models are trained to be supportive. “Most of the AIs are just trained to agree with you,” he said. Tell it you’ve got a great idea and it will happily draft your marketing plan, no matter whether anyone will actually buy the thing. It doesn’t have the tools to tell you your idea won’t work, and it doesn’t try. That’s a job for a human who’s actually built something before.
Mattson’s own workaround: use AI as a tool for execution, never as the source of strategy. “It’s important not to use AI as your strategy… I use it when I’m paralyzed. I’ve got writer’s block, and asking some questions gets the ball rolling,” she said. Good for copy, good for crunching numbers she can’t do herself, bad for deciding whether the business should exist in the first place.
What should founders expect emotionally in the first year?
Mattson runs a mentorship track for early-stage entrepreneurs, and she’s watched the same emotional arc play out enough times to map it out loud: excitement, telling everyone your idea, hope, fear, envy when you see someone else doing something similar, depression when it’s not working, avoidance of the people you told, and then, eventually, more motivation. “There’s a lot going to happen. And I know this sounds horrible to say to someone, but you got to know what you’re getting into,” she said.
Her practical fix isn’t emotional, it’s financial. Keep six months of savings, or a loan, set aside before you lean on the new business. Without that cushion, she said, founders start making knee-jerk decisions, pivoting too fast, and spending unwisely just to keep the lights on. When she started her ad agency, she made sure it kept running while she funded the startup on the side, precisely so one wouldn’t have to cannibalize the other.
How do you avoid building something you don’t actually care about?
Mattson’s test is simple: would you still want to talk about this thing in six years? “If you’ve gone out and created a vacuum cleaner that’s super robotic and better than anything on the market, you better like to vacuum, okay? Because you’re going to be talking about vacuuming for like the next six years of your life,” she said.
That same instinct extends to letting go of ideas that aren’t working. Mattson learned this one the hard way. She started an e-commerce company in 2004, before people trusted buying online, convinced that if she built it, customers would come. Almost none did. “I was so tied to my idea… that nobody knew better than I did, and that was a big mistake. I paid very dearly for that mistake with a lot of debt for the next 15 years,” she said. Her advice now: launch fast, learn, and don’t wait for perfect before you find out if anyone wants what you’re building.
What’s the single biggest marketing mistake early-stage founders make?
Casting too wide a net. Mattson sees founders take out an ad with no filters, hoping volume makes up for precision, and end up marketing a niche product to people who were never going to buy it.
“You got to go to a small pond, where the people who care about your product are there, and you spend efficiently, not just spend to spend.”
— Julie Mattson, Co-Founder, The Borderline
It compounds with a second mistake: wildly optimistic financial projections. Founders see an average Shopify conversion rate quoted somewhere and assume it applies to them, without ever working out true cost per acquisition. Mattson’s rule is to take whatever number you land on and cut it down hard. “Whatever you think you’re going to make, drop it down a couple points decimal,” she said. Finkelstein, who’s advised plenty of founders on this exact problem, agreed it’s one of the most reliable ways to protect your own morale: undersell yourself on paper, and actually hitting your numbers feels like a win instead of a letdown.
The same logic applies to positioning. Mattson pushes founders to know their barrier to entry before they build anything, because a clever feature with no defensible brand around it is an invitation for a bigger competitor to absorb it. Know your story, she said, the way a bakery that leans into a single family recipe outperforms one trying to be everything to everyone.
The bigger picture
None of this is anti-AI. Mattson runs her business on tools that didn’t exist a decade ago: Figma for creative collaboration, Julius for the number-crunching she can’t do herself, Claude for getting unstuck. What she’s arguing for is sequencing. Use the tools to move faster once you know what you’re doing. Don’t let them decide what you’re doing in the first place. Financial discipline, a genuine passion for the problem, and a brutally honest read on who actually wants your product still have to come from a human, running the numbers conservatively enough to survive being wrong.
Julie Mattson is offering Deconstructing Data listeners a free 30-minute business consultation (normally $150) with code BDEX50 at theborderline.co — search “Julie Mattson” to book.
Ready to bring that same discipline to your data? Visit bdex.com and click “Talk to an Expert” to get started.
Watch the full episode: From Startup to Scale: Smarter Marketing, Fewer Founder Mistakes
This article was adapted from an episode of Deconstructing Data, BDEX’s weekly podcast on data-driven marketing. Tune in live every Thursday at 4:15 PM Eastern on LinkedIn.
About BDEX: For companies that need clean identity data to power their products, BDEX offers unmatched quality and execution. Visit bdex.com and click “Talk to an Expert” to get started.
