Second Time Smarter: Discovery, Discipline, and the Startup Learning Machine
This article was originally posted on the Momentum Mode Podcast Substack channel, featuring Impruve CEO Mike Shannon.
There’s something almost poetic about startup life—especially when “the next round” isn’t about raising capital, but raising your game. In this episode, I sat down with Mike to unravel the invisible shifts that come when you play the founder role again—and this time, play it smarter.
From Forcing a 50-Page Plan to Unearthing the Opportunity
Mike recalls his first startup, Packback, launching in 2011 armed with a 40-page business plan. Their ambition? Textbook rentals that dashed into a brick wall of indifference. It wasn’t until years and pivots later that they discovered the real opportunity: discussion tools for education.
Contrast that with the second venture—Impruve. Mike cites the wisdom of Stephen King: rather than forcing the narrative, a founder becomes an archaeologist, dusting off what’s already laid bare. The goal shifts from execution of a plan to discovery of what actually resonates.
Build the Learning Machine, Not the Monument
Old play: “Here’s our vision—now build it.”
New play: “What’s the smallest thing we can test today?”
Mike shares how, for Impruve, their first prototype (“feedback-on-the-fly” tool) was a janky, seven-minute-loading sidekick. But it got into customers’ hands within 132 days—enough to create a learning engine. They kept the team lean, burned barely anything, and got fast signals to inform the next draft.
Next Draft, Not Pivot
Every iteration for Impruve is anchored in reality—not just chasing a new vision. With eight-week cycles, they’ve sifted through seven or so “next drafts,” evolving from a middle-manager coaching tool to an AI OS for independent financial advisors.
They refuse the word “pivot”—too abrupt, too jarring. This isn’t a sudden channel flip; it’s a sculpting of insights, refining one draft after another until the story lines up with real traction.
How to Start Lean and Stay Lean
Low-fi before you code: Use Google Forms and Figma click-throughs to test four early ideas—before building anything. Ask five targeted questions. Ask why.
Network as your lab: The Density Collective in Chicago gave early beta testers from day one. That network fuelled honest feedback and vertically focused discovery.
Follow the hero: Obsess over the user’s pain points. What workflows feel clunky? Where does frustration creep in? Then sit with them, shadow them, watch the tape.
Lean, lean, lean: 300K raised, 15 employees at year two. That same size today. They’ve built powerful features with a tight team and AI amplification, avoiding the cost traps of scaling too soon.
Founder 2.0: Capital-Efficient, Data-Driven, Long-Game Ready
First-time Mike took in millions and learned slowly. Second-time Mike raised far less and learned exponentially faster—burning $10K (not $3M) to uncover fatal signals. He’s built something that doesn’t rely on an exit timeline or investor “hold period.” Instead, he’s free to play long, stay boutique, and evolve as the data demands.
So What?
If you’re building again—or thinking about it—this episode offers an invaluable lens:
Drop the business plan; pick up the brush and uncover the canvas.
Build a learning machine, not a monument.Embrace next drafts—not panic pivots.
Stay lean. Stay close. Stay learning.
Startups aren't stories you write in advance. They’re narratives you excavate—draft by draft, signal by signal, draft by draft.
