Gavel Gavel
Experts Guides Skills
Guides
AI advisor

How to make better decisions as a founder

You make better decisions as a founder by changing the process, not by thinking harder. Four moves do most of the work: name whether the decision is reversible or irreversible and match your speed to it, reach for a named framework instead of your gut, force the choice down to the one constraint that actually matters, and actively seek where credible operators disagree so you are choosing, not guessing. Below is each step, grounded in how operators like Alex Hormozi, Garry Tan, and Sean Ellis actually decide.

Why this matters. Most founder paralysis is not a knowledge gap, it is a process gap. The founders who move well are not smarter; they have a repeatable way to tell a fast call from a slow one and a framework to run each through.

Ask Gavel, free

Bring the decision you are stuck on. Gavel names the framework, shows both sides, and applies it to you.

5 or 6

hair-on-fire users are enough to settle a reversible decision. Garry Tan says launch the jankiest version to them and let their behavior, not your gut, make the call.

Garry Tan Garry Tan, YC Office Hours

The short answer

How to decide better, in five steps

A process you can run on any founder decision, from a feature cut to a pivot.

  1. 1

    Sort by reversibility

    Decide fast on calls you can undo; slow down only for the ones you cannot take back.

  2. 2

    Reach for a named framework

    Run the decision through a framework an operator actually used, not your gut feeling in the moment.

  3. 3

    Find the one constraint

    Name the single bottleneck that limits growth and put the decision against it, instead of optimizing everything at once.

  4. 4

    Seek the disagreement

    Find two credible operators who split on the question and choose against your own situation, rather than accepting one confident answer.

  5. 5

    Probe reversible calls

    For anything you can undo, ship a small test and let real behavior decide, instead of debating in a doc.

Speed is not the goal and neither is rigor. Matching the two to the decision in front of you is.

The cited playbook

The cited decision moves

Each move is how an operator who has made hard calls actually decides, linked to the source and built to run on the decision in front of you.

  1. 1

    Force the decision down to one bet

    Alex Hormozi frames the Core Four as a forcing function: if you could commit to only one channel for the next ninety days, which is it, and the answer is a closed door on the other three. Apply the same discipline to any decision, name the single thing you are committing to and consciously say no to the rest, because effort spread evenly across options moves none of them.

    Alex Hormozi · Hormozi's Core Four
  2. 2

    Score the options with a framework, not your gut

    When you are stuck between bets, force a number on each. Sean Ellis's ICE framework scores Impact times Confidence times Ease, and the value is not the ranking itself, it is that scoring surfaces the assumptions you were hiding, so a low-confidence bet stops masquerading as an obvious yes.

    Sean Ellis · Hacking Growth (ICE)
  3. 3

    On reversible calls, ship a probe and let reality decide

    Garry Tan's rule is to launch the jankiest version that still provides value and iterate on real reaction, because the first launch is a probe, not a permanent record. For any decision you can reverse, get it in front of five or six people with a hair-on-fire problem and let their behavior settle it, instead of deciding in a meeting.

    Garry Tan · Garry Tan, YC Office Hours
  4. 4

    For the irreversible calls, use the framework the decision deserves

    Some decisions, like how you position the company, you cannot cheaply undo, so slow down and use a real method. April Dunford's is to decide from the customer's real alternative, what they would use if you did not exist, rather than from a feature list. Match the weight of the method to the weight of the decision.

    April Dunford · Dunford on competitive alternatives

Where experts disagree

Where operators disagree: decide fast, or measure first?

Garry Tan

says ship the jankiest version now and let real reaction decide, because fear of launching kills more startups than competition. The first launch is a probe, so move fast on anything you can reverse.

Sean Ellis

says measure before you commit: 40% of active users answering very disappointed is the product-market-fit line, and scaling on a vibe instead of that number is how founders pour fuel on a fire that is not lit.

Reversible decisions reward Tan's speed; irreversible ones, where you point the whole company, reward Ellis's threshold. ChatGPT smooths this into be data-driven but move fast. Gavel shows you both and which one your decision calls for.

FAQ

Founder decision-making questions, answered

How do founders make better decisions?

By fixing the process: sort the decision by whether it is reversible or not and match your speed to that, run it through a named framework instead of your gut, force it down to the one constraint that matters, and actively seek where credible operators disagree so you are choosing between real options rather than accepting one confident answer.

Should I decide fast or gather more data?

It depends on whether the decision is reversible. For calls you can undo, Garry Tan's advice is to ship a probe and let reality decide fast. For calls you cannot take back, Sean Ellis's advice is to measure first, for example the 40% very-disappointed line before you scale. Match speed to reversibility.

What decision framework should a founder use?

There is no single one; use the framework that fits the decision. Score competing bets with Sean Ellis's ICE, find the growth-limiting constraint the way Alex Hormozi does, position from the real alternative with April Dunford, and treat reversible launches as probes with Garry Tan. The skill is matching method to decision.

Can ChatGPT help me make better decisions?

It can lay out options and explain a framework, but it tends to agree with whatever you propose, gives you the internet's average, and forgets your context between sessions. For a real decision, you want a cited operator method, both sides of the genuine fork, and an answer applied to your stage, which is what a grounded tool provides.

Bring your actual numbers. Get a cited answer you can defend.

Try Gavel free

Related guides

AI advisors

AI business advisor

An AI business advisor should cite named operators, show where they disagree, and remember yo...

AI advisor

AI co-founder

An AI co-founder should bring cited operator frameworks to your hardest calls, remember your ...

Product

Prioritization framework

RICE and ICE score initiatives so you can rank them. Here's how each works, and why operators...

Keep exploring

Experts & frameworks

Every operator Gavel cites, with timestamped sources.

Is ChatGPT good for…

The decision-by-decision take on where generic AI falls short.

Playbooks

Multi-expert guides for pricing, PMF, first customers.