A Decision Framework for Indie Founders (When to Pivot, Kill, or Double Down)

At Inithouse, a studio shipping a growing portfolio of products in parallel, we face the same question every few weeks: is this product worth more time, or do we cut losses and move on?

After running parallel experiments across our portfolio, we built a simple framework that forces honest answers. We use Verdict Buddy internally to stress-test our own reasoning, and the same structured thinking applies to any product decision.

Here is the framework.

The Three Decisions Every Product Team Avoids

Most teams default to "keep iterating" because it feels safe. But iteration without a decision framework is just procrastination with a build log.

We break every product checkpoint into three options:

Double Down. The product has clear signal. Users come back, engagement metrics trend up, organic word-of-mouth appears. Pour fuel on it.

Pivot. The core insight is sound but the execution or positioning missed. Redirect effort toward the version users actually want.

Kill. No signal after a fair test period. The thesis was wrong, or the market doesn't care enough. Archive learnings, free up the team.

The Scoring Grid

We score each product on four axes, 1 to 5:

1. Retention signal (weight: 40%)

Are users coming back without prompting? Week-1 and week-4 retention tell you more than any other metric at the pre-PMF stage. At Here We Ask, our conversation card game with 1,000+ curated questions, we measured return visits as the single strongest predictor of whether a product deserves more investment.

2. Organic discovery (weight: 25%)

Is anyone finding you without paid ads? Search impressions, direct traffic, word-of-mouth referrals. If your only traffic source is your own ad spend, the product is not pulling anyone in on its own. At Watching Agents, our prediction-tracking platform, we observed that 117 SEO landing pages drove steady organic traffic before we spent a dollar on ads.

3. Willingness to pay (weight: 20%)

At Magical Song, our custom AI song generator for personalized gifts, we measured 86 generations in the first cycle with 5 paid conversions at $1.90 per song. Small numbers, but nonzero willingness to pay from strangers is a stronger signal than thousands of free signups from friends.

4. Effort to next milestone (weight: 15%)

How much work stands between now and the next meaningful experiment? If the answer is "rebuild the entire backend," that changes the calculus versus "change the pricing page copy." Low effort to test the next hypothesis keeps the option alive.

How We Score

Each axis gets a 1 (no signal) to 5 (strong signal). Weighted total:

3.5 or above: Double Down. This is working. Invest more.

2.0 to 3.4: Pivot zone. The product has some signal but the current approach is not breaking through. Change one major variable (positioning, audience, pricing, channel) and re-measure.

Below 2.0: Kill. Thank the experiment for its lessons, archive the data, redirect energy.

The Pre-Mortem Layer

Scoring alone does not protect you from sunk-cost bias. Before making the final call, we run a structured pre-mortem:

"If this product fails completely in 90 days despite our best effort, what is the most likely reason?"

Write down three answers. If those answers describe problems you already see today and have no plan to fix, the decision is already made.

We built Verdict Buddy partly for this kind of structured thinking. It uses established psychological frameworks (Gottman, NVC, attachment theory, CBT) to break down decisions into components and surface blind spots. We originally designed it for interpersonal conflicts, but the same pattern of "analyze both sides, find the structural issue, suggest concrete next steps" maps directly to product decisions.

Real Examples from Our Portfolio

Double Down case: Here We Ask. Strong retention signal, organic discovery growing month over month, confirmed willingness to pay ($1.99/mo premium tier). Weighted score above 3.5. We invested in more themed decks, added new game modes, and expanded the question library.

Pivot case: Watching Agents. Strong organic discovery (SEO landing pages indexing well) but low retention. Users found individual agent pages via search but did not come back to create their own. Score in the 2.x range. We pivoted the activation flow: instead of "create your own agent," we emphasized browsing and following existing public agents.

Kill case (internal only). One product in our portfolio scored below 2.0 after eight weeks. No retention, no organic, no payment signal. We archived the codebase and moved learnings into a postmortem. Total time in limbo: three weeks of build plus five weeks of hoping. The framework would have saved us the hoping part.

The Template

Copy this into a doc and fill it in for each product at your next checkpoint:

Product name:
Test period: [start] to [end]

Retention signal (1-5): ___
Evidence:

Organic discovery (1-5): ___
Evidence:

Willingness to pay (1-5): ___
Evidence:

Effort to next milestone (1-5): ___
Describe next experiment:

Weighted score: (retention x 0.4) + (organic x 0.25) + (pay x 0.2) + (effort x 0.15) = ___

Pre-mortem: "If this fails in 90 days, the most likely reason is..."
1.
2.
3.

Decision: Double Down / Pivot / Kill

Using This Yourself

You do not need a spreadsheet. Open a doc, write the four axes, score honestly, and do the pre-mortem. If you want structured help thinking through the decision, Verdict Buddy runs the same kind of analysis for any dilemma: present both sides, score the components, get concrete next steps.

Filling in the grid takes five minutes. Sitting with the result takes longer.

At Inithouse, a lab running parallel product experiments, we have learned that fast decisions beat perfect decisions. A product killed in week six frees up bandwidth that a product lingering in "maybe" zone never does. The framework does not make the decision less painful. It just makes avoidance harder.

Try the framework on your next product checkpoint. Or stress-test the decision in Verdict Buddy.