Pleasant Valley Labs

Pleasant Valley Labs/Work/FenceRow

How we built FenceRow

FenceRow shows farmers what every field has to make before they plant it. How we built it, the rules it follows, and what we learned shipping paid software in under three weeks.

The problem

Before a farmer plants a field, they need two numbers: the yield it has to produce and the price the crop has to sell for to cover what went into it. Most farmers work those out on a spreadsheet, a notepad, or a calculator at the kitchen table. The numbers change every season, and the work from last year rarely carries forward.

The free tools that do exist usually come from companies that sell seed, inputs, or grain marketing. For them, the breakeven is a way to start a sales conversation.

We wanted a tool that does one job well, keeps the farmer's records year over year, and never shows a number it can't explain.

What we built

FenceRow tracks breakeven yield and breakeven price for each field a farmer works, season after season. The farmer enters their own costs. FenceRow pulls current prices and yields from USDA's National Agricultural Statistics Service and does the math.

It's live at fencerow.farm.

A sample corn plan in FenceRow, from setup to breakeven.

The rules we built it around

We wrote five rules before writing any code, and the calculation engine itself enforces them, so no screen can skip them:

  1. No made-up numbers. If USDA has no price or yield for a crop and state, FenceRow says so. It never fills the gap with an estimate.
  2. Blank is not zero. A cost the farmer hasn't entered is shown as missing, never quietly counted as $0.
  3. Every number shows its work. One click from any result to the costs and formula behind it.
  4. History doesn't change. Once a result exists, editing the inputs creates a new version. A number a farmer showed their lender stays the number they showed.
  5. Rotation advice shows its limits. Guidance is given as a cited range, with its source.

Every one of those rules came from the same question: would a farmer trust this number enough to take it to the bank?

Pricing that fits a farm year

Most software gives you a 14 or 30 day trial. Farming doesn't run on 30 days. Costs come in over the winter, prices matter in spring, and the real review happens after harvest. A trial would expire before the app had a chance to prove itself.

So FenceRow is free for up to three fields, with every feature included, for as long as the farmer wants. Past three fields it's $49 a year, or $5 a month. We priced it annually first because farm cash flow is annual.

How it's put together

  • Web app: Next.js, with Google sign-in
  • Calculation engine and API: Python and FastAPI, with a test suite that pins the math to a known-good reference case
  • Database: Postgres
  • Payments: Stripe, handling yearly and monthly plans, renewals, and failed-payment retries
  • Data: USDA NASS, cached and labeled so a farmer always knows how fresh a number is

Timeline

  • Day 1: spec locked, calculation engine ported and tested, app deployed to a test address
  • Week 3: real payments live, domain moved to fencerow.farm, public launch

What it shows about how we work

FenceRow is a small product with strict rules, and that's the point. We'd rather build the thing that does one job and can be trusted than the thing that does ten jobs and needs a disclaimer. That's the same approach we take with client work: figure out which numbers people will make decisions on, then make sure those numbers are right and explainable.

Have something you want built, connected, or automated? Tell us about it.