The Product I Killed on Purpose

At Strategy Compass I owned an internal venture called Mindlantic — a microlearning product for the soft skills professionals actually get judged on: reading a room, building a presentation that lands, persuading the people who hold the budget. I ran it end to end. A designer and five developers reported to me; I set the strategy and the roadmap, did the hiring, ran the discovery. By the time the real decision arrived, we had put it through roughly sixty discovery interviews.

The Setup

Mindlantic began as a pet project of the CEO and his business partner — two people who had spent years learning how to win inside a consulting business and wanted to teach it. My job was to turn that conviction into a product: a real team, a real roadmap, and a real plan to either spin it out as its own business or prove it couldn’t be one.

On the merits, it worked. The first users genuinely loved it. The differentiator was content quality — lessons crisp enough that every one earned its place, very high signal-to-noise — and people felt that immediately. Stop the story here and the call writes itself: demand is real, spin it out, go.

The Fork

The obvious move was to spin it out. I didn’t.

What I kept looking at wasn’t the demand — it was the production model behind it. The thing that made Mindlantic good was the quality of the content, and that quality came from somewhere that didn’t scale: two people distilling decades of hard-won experience into lessons crisp enough to teach. Every lesson was hand-built to a bar we had set deliberately high. To become a standalone business we would have had to multiply content output roughly tenfold — and there was no way to do that without lowering the bar that was the entire reason anyone liked it.

So the real question was never “do people want this.” It was “can we produce this at the quality our edge depends on, at the volume a standalone business needs.” The answer was no, and no amount of demand validation changed it.

I made the call to fold Mindlantic back into the portfolio. We narrowed it to the soft skills that compounded with our core Office products — how to build a presentation, how to deliver it — where we already had the content, and shipped it as a learning helper inside Quickslide, QuickDocs, and QuickSheets. The portfolio got stronger. The team’s work didn’t get spun out to die slowly.

The part I am most deliberate about: I was both the product manager for Mindlantic and the person making the portfolio call about it. It would have been easy to protect my own venture. I tried to see it exactly as it was — an experiment that had taught us something real — and to hold it to the same evidence I held everything else to.

Validating demand is necessary, but it isn’t sufficient. You also have to validate that your production model scales at the quality bar your differentiator depends on.

Why This Call Still Holds

The constraint that kills a product is rarely the one in the demo. The demo proves people want it. What decides whether it ships at scale is whether you can produce it — to the quality your advantage depends on, at the volume your buyer needs. A buyer won’t take what you can’t produce to spec, however well it solves their problem on paper.

And it’s worth knowing which kind of constraint you’re up against. Ours wasn’t tooling. The quality came from two people distilling decades of experience — the kind formed in rooms where the real conversation never gets written down. You can’t hire your way to ten times more of that, and — the modern version of the question — you can’t prompt your way to it either: that experience was never in a training set, because it was never documented. The obvious bottleneck looked like content production; the real one was the scarcity of the judgment underneath it — and that kind of scarcity doesn’t scale, then or now.

I followed that thread — what stays scarce once AI makes the building part cheap — in Moats in the Age of AI.