Founder field note · Annual review

Growing Pantry Ledger to $12.4k MRR—and what came after

By Kiran Bell · 31 December 2025 · 18 min read
Monthly recurring revenueJan–Dec 2025
$12.4k MRR
JanAprJulOctDec

2025 was the year Pantry Ledger stopped feeling like a side project and started behaving like a small, demanding company.

I began the year with an uneven product, a patient group of customers, and a list of assumptions that turned out to be mostly wrong. We ended with steadier revenue, a rebuilt core experience, and much clearer ideas about what not to pursue.

This review covers the shipping decisions, conversion experiments, and operating changes that moved the business—not every feature or bug fix.

Why the year began with a rewrite

The original app was fast to build but difficult to extend. Every nutrition view carried its own logic, sync failures were hard to diagnose, and small interface changes touched too many screens.

We paused feature work for six weeks and moved the shared model into one tested layer. It was not glamorous, but it let the rest of the year compound.

The rewrite mattered because it reduced the cost of the next twenty decisions—not because customers wanted a rewrite.

A clearer daily log

We replaced the old timeline with a compact daily ledger. Meals, targets, and recurring foods now share one visual rhythm, and common actions stay close to the information they change.

Tuesday•••
Protein
84g
Fiber
27g
Energy
1,842
Oats, pear, walnuts421
Lentil bowl538
Tofu noodles612
Evening snack271

Shipping changes that moved retention

Faster food search

Recent ingredients and personal recipes moved ahead of the generic database, cutting repeat logging time noticeably.

Reliable recurring meals

People could save a meal once, schedule it, and edit a single day without breaking the series.

Import that showed its work

Recipe import began displaying uncertain ingredients instead of silently guessing.

A calmer weekly view

Seven days of totals became readable without turning food into a dashboard of warnings.

What we learned about conversion

The best-performing paywall did not add urgency. It explained the difference between casual logging and the workflows paid members already used: saved meals, weekly trends, custom nutrients, and reliable exports.

7.8%trial-to-paid
3.1%monthly churn
68%annual-plan share

Annual plans grew after we added a plain explanation of renewal timing. Hiding the detail had produced more support, not more trust.

The platforms we chose not to chase

We explored a watch app, a desktop companion, and an expanded recipe marketplace. Each had an audience. None was as valuable as making the core log dependable.

The strongest growth came from doing fewer surfaces well: phone, tablet, web review, and export.

Marketing became a practice

For most of the year, marketing happened when I remembered to do it. The shift was to a small repeatable cadence: one useful article, one customer lesson, and one product note each month.

Search traffic grew slowly. Referrals grew faster. The most effective pages answered narrow questions people were already asking rather than explaining the entire product.

Support made the roadmap better

I began tagging conversations by the job a customer was trying to complete. A request for “more folders” sometimes meant faster search; a request for “AI meal plans” often meant help choosing from foods they already ate.

Listen for the job

The proposed feature is evidence, not the requirement.

Write the decision down

A short reason prevents the same debate from restarting.

Close the loop

Tell people when their report changed the product.

Protect the core

Every new capability should make daily logging clearer.

Operating alone without operating chaotically

A weekly planning block replaced the endless task list. I chose one product outcome, one growth experiment, and one maintenance goal. Everything else waited unless it was urgent or tiny.

Meetings remained rare, but written updates became more precise. Contractors could see the constraint, the decision, and the acceptance criteria without a long call.

The emotional part is still part of the work

Recurring revenue makes progress visible, but it can also turn every cancellation into a verdict. I learned to review signals in groups: four weeks of churn, a cohort of onboarding behavior, a cluster of support conversations.

Distance did not make the work less personal. It made decisions less reactive.

What comes next

In 2026 we are investing in better planning, more trustworthy nutrition sources, and a small set of collaboration tools for households and coaches.

The goal is not to make Pantry Ledger bigger everywhere. It is to make the useful loop—plan, log, notice, adjust—feel lighter every time.

Follow the next year in public

Monthly numbers, product decisions, and the lessons behind them.

Read the founder notes