Signal, Not Gut: How a Solo Founder Reads Data (North Star Metric)

The moment you swap gut for data, real business begins.

3 weeks of gut, 3 months saved by data

In 2023, a solo SaaS founder launched "an automated reporting tool for marketing teams." Gut was good. "Surely this is the feature customers want most." She read every piece of feedback for 3 weeks and kept adding features that matched her own idea. 3 weeks later, users fell from 100 to 97.

Another solo founder classified every piece of feedback from the first 20 users. "What's most mentioned?" — by data. The most frequent complaint wasn't the feature she imagined — it was "just make this one thing fast." She built that one thing in 3 days, and the next week users went from 20 to 45.

What's the difference? First was gut; second read signal. This post is about reading signals.

Pick one core metric: the North Star Metric

"Measure everything" is a lie. If a solo founder watches dozens of metrics at once, they end up watching none. Instead, define your North Star (NSM).

NSM is "the single metric that best represents your business's success." For SaaS: MAU. For B2B tools: team invites. For subscriptions: conversion to paid.

3 conditions for NSM

1. Leading indicator: money coming in is a lagging signal. NSM should indicate future money. For SaaS, monthly logins are a faster signal than monthly revenue.
2. Actionable: if NSM is "team invites," "make the invite feature more visible?" should immediately follow.
3. Measurable: "user satisfaction" is abstract; "NPS ≥ 70" is concrete. Must be a number trackable daily.

Before picking NSM, set 3 hypotheses a day for a week. "If users log in 10 times a day, will paid conversion be higher?" The metric you're certain says "yes" is your NSM.

AARRR framework, solo-founder edition

AARRR (Acquisition → Activation → Revenue → Retention → Referral) by marketing investors is too complex for solo founders. Compress to 3 stages: "come in, use it, come back."

Stage 1. Acquisition ("come in"), first 100 users

Metric: new signups. Solo founders often mistake here. "They signed up, done." Wrong. What matters is "did they actually do something within 7 days of signup?" If 30 of 100 signups come back on day 1, your product is "interesting."

Stage 2. Activation & Usage ("use it"), core feature use

Metric: MAU and core-feature-use rate. Is the share of users who tried the most important feature at least once over 50%? Over 70% means your product is "usable."

"Lots of users but somehow no revenue" is common with solo founders. Usually you think "price is the issue," but the problem is usually that users haven't tried the core feature. Many free signups, but only 5% have "actually felt the value."

Stage 3. Retention & Revenue ("come back"), continued use + paid

Metric: Day-7 / Day-30 retention, paid-conversion rate. Day-7 retention = "% of signups who return after 7 days." B2B SaaS sees 40%+ as healthy. Successful Korean solo SaaS often holds 40–60%, then 20–30% at the paid-conversion point.

AI-classified feedback: read signals with Claude

Feedback isn't a metric, but patterns in feedback are data. Reading 100 pieces by hand works; with 1,000, not so much. Claude steps in.

Workflow

1. Collect feedback in a spreadsheet (email, Slack, surveys, comments)
2. Ask Claude to classify
3. Claude returns JSON or a table
4. Spot "the most repeated problem" in 10 seconds

Below is 100 pieces of user feedback. (paste)

Classify each as one of:
- Feature request
- Bug report
- Usability complaint
- Praise
- Price complaint

Then list the 3 most-common words per category.

Real case: a solo founder classified 50 feedback items and found "slow" bug reports were 34%. A complaint they'd have intuitively ignored — but the data said "nearly 1 in 3 complains." 3 days on performance optimization later, weekly logins rose 25%.

Claude's auto-classification accuracy is 85–95%. Not perfect, but much faster than manual and better than gut bias.

Signal vs noise, reading real signals

The most important question: "Is this change real, or coincidence?"

Users went up by 5. Real improvement, or luck? Statisticians call this "significance."

Rule 1. The 30-person rule

Sample under 30, ignore small changes. 2 of 15 users tried a new feature ("11% increase!") doesn't mean a real 11% improvement. Probably coincidence.

Rule 2. Minimum 7 days, 2 weeks if possible

Deployed Monday, "everyone's using the new feature!" on Tuesday — don't believe it. You need a week of data. User patterns differ by day of week.

Rule 3. Create a comparison

Shipped feature A? Show it to some users first (A/B test). Compare "saw feature vs didn't" — that's how you see real signal, not gut.

Signal checklist

• Data from 30+ people?
• Measured at least 7 days?
• Compared yesterday vs today, or group A vs B?
• Can you name at least 1 cause for the change?
• Confident the change will persist?

Metric priority matrix

Step 1. Pick your 3 metrics

• NSM: e.g., monthly active users
• Short-term signal: e.g., new signups
• Wound to heal: e.g., low D7 retention

Step 2. Priority matrix

"Login 2x faster" (high cost, high impact) → do now. "Make background blue" (low cost, low impact) → idea notes.

Step 3. Feedback classification prompt

Below is user feedback from [SaaS name]. (paste)

Classify each as JSON:
{
  "feedback": "...",
  "category": "feature / bug / usability / praise / price",
  "sentiment": "positive / neutral / negative",
  "urgency": "high / medium / low"
}

Wrapping up

The moment you swap gut for data, you finally become a real operator. Not luck — someone who makes decisions.

Next post: with that data, choose one of 3 forks. Keep building (Persevere), change direction (Pivot), or stop (Kill). The bravest moment.

References

1. North Star Metric: why and how to find good ones — MarketFitLab
2. AARRR funnel: retention, monetization, referral — Ascent Korea
3. Claude Code: best practices for agentic coding
4. Korean B2B SaaS startups: status and success strategies — KDI
5. P-Value: role in startup success evaluation — FasterCapital

Previous: How to Honestly Get Your First 100 Users
Next: Pivot, Persevere, Kill — the courage to kill makes the next MVP

Minchul Kim, CEO of Freeive, Fail School