← Writing

Why Your Fix Made Things Worse: Feedback Loops for Founders

JB Ong

You shipped the fix. For two weeks, it worked. Then the same problem came back — bigger, and now tangled up with three new ones you didn’t have before.

That’s not bad luck, and it’s not a people problem. It’s a feedback loop. And once you can see them, a lot of “unfixable” founder headaches stop looking like fires to stamp out and start looking like a board you can actually read.

New to this lens? Start with What Is Systems Thinking? A Founder’s Guide — the plain-English foundation this essay builds on.

The trap: you’re fixing events, not structure

Most early-stage problem-solving happens at the level of events. Signups dipped — run a promo. Support is drowning — hire another rep. A key hire is unhappy — give them a raise. Each move feels obviously correct, because each one targets the thing right in front of you.

The catch: events are the smoke. The structure underneath is the fire. When you fix the event without seeing the structure, the system quietly routes around your fix — and often uses the energy you added to make the original problem stronger.

Three loops show up again and again in young companies. Learn to spot these and you’ll catch the backfire before you fund it.

Loop 1: The relief that removes the pressure to fix the real thing

The pattern: A painful symptom pushes you toward a quick fix that relieves the pain — which removes the urgency to solve the underlying cause, so the cause grows.

Founder version: Your best engineer is buried, so you keep pulling them onto every fire. It works — the fires go out. But because they always go out, you never fix the reason fires keep starting (no docs, no ownership, no process). The engineer becomes more indispensable, more buried, and more of a single point of failure. The “fix” was quietly buying down your motivation to build the thing that would actually end the fires.

What to watch for: any solution that makes the symptom disappear fast but touches nothing about why the symptom appeared. Relief is not repair.

Loop 2: The fix that shifts the cost somewhere you’re not looking

The pattern: You solve the problem in one part of the system by pushing the cost into another part — usually one that shows up on a different dashboard, or no dashboard at all.

Founder version: Growth is flat, so you crank up discounting to hit the number. Revenue moves. But you’ve just trained the market to wait for discounts, attracted your most price-sensitive (and highest-churn) customers, and squeezed the margin that was funding support. Next quarter the top-line looks fine and the business underneath is weaker. The problem didn’t leave — it moved to a column you weren’t watching.

What to watch for: wins that appear in one metric while a quieter metric — margin, morale, retention, tech-debt — starts sliding. Ask “if this worked, where did the cost go?” before you celebrate.

Loop 3: The success that eats itself

The pattern: Something works, so you pour more into it — but growth hits a limit you didn’t design for, and the very thing that made it work becomes the thing that breaks it.

Founder version: Word-of-mouth is your growth engine, so you pour fuel on demand. Demand outruns onboarding. New customers get a worse first experience than the early ones did — and word-of-mouth runs on delight, so the engine that got you here starts spreading disappointment instead. You didn’t lose the loop; you overran its limit.

What to watch for: any engine you’re scaling hard. Ask what constraint it depends on (quality, attention, a single team, trust) and whether that constraint scales with it. Engines have limits; find them before they find you.

The move: read the board before you touch a piece

You don’t need a whiteboard full of arrows. You need three questions before any “obvious” fix:

  1. What loop am I in? Is this fix relieving a symptom, moving a cost, or overrunning a limit — or actually changing the structure?
  2. Where does the cost go if this works? Follow the effect two steps out, into the metrics and teams you’re not looking at.
  3. What was holding the old behavior in place? If you don’t remove that, the system will pull back to where it was — and treat your effort as fuel.

Answer those and you’ve done the thing most founders skip: you looked at the whole board instead of the piece that’s shouting loudest.

Why this is a practice, not a trick

Nobody sees every loop the first time. Seeing structure is a muscle — the more real decisions you run through those three questions, the faster the patterns jump out, until “wait, where does the cost go?” becomes reflex instead of hindsight.

That’s the whole idea behind what we build at Gaia Gauge: small, repeatable ways to keep a systems lens on without it feeling academic. If this way of seeing is useful, here are three low-friction ways to keep the muscle warm:

The next time a fix backfires, don’t ask “what did I do wrong?” Ask “what loop was I in?” That question is the entire skill.