The systems were never built to agree

Late July in the California Central Valley means one thing: harvest is bearing down, and nobody wants to think about year-end yet. Fair. But this is the season the calls start, because the same picture shows up inside almost every operation we walk into. The plan lives in one system, usually an agronomy platform. The applications live in another, usually a tracker spreadsheet somebody maintains by hand. The invoices live in a third, the ERP. Ask a simple question, how much have we applied against the plan to date, and the answer takes three spreadsheets and a phone call.

Nobody built it that way on purpose. The agronomy platform was bought to write plans. The ERP was bought to cut POs and post invoices. The spreadsheets grew up in the gaps because somebody needed an answer by Friday. Each system is doing exactly the job it was bought for. They were just never built to agree with each other.

The weekly number is the bottleneck

In season, the whole operation runs on one number a week: what goes out, by field and set. Producing that number means reconciling the plan against what has already gone out, and what has already gone out lives in two places that disagree. So a fifteen-minute question turns into a Thursday-night exercise, and when the number runs late, the guys in the field are stuck guessing: am I ahead or behind, and what goes out this week? Hesitation like that in July can show up in the tissues by September. That is how one gap between systems gets felt every week of the season.

Here is a perfect example, the generic version of a scene we have watched play out more than once. Mid-May, a ranch manager needs the week's number for a 300-acre block. The plan says 18 gallons an acre of potassium by the end of the month. The tracker shows 11 gone out. The invoices total 14. Three systems, three answers, one truck waiting. He does what any sensible person does: calls whoever keeps the newest spreadsheet, waits, and the crew starts a day late. Multiply that by every block, every week, and you have found the operation's real capacity limit. It was never the people. It was the reconciliation.

Then year-end lands on top of it. Proposed vs. applied vs. billed, trued up across every field, becomes a horrendous exercise when each of those three words lives in a different system. We have watched capable operations spend weeks on it. Not because anyone is sloppy. Because agreement was never designed in, so it all happens by hand, at the worst possible time of year.

Nobody is lying

That is the part worth saying plainly. When the plan says 400 pounds, the tracker says 340, and the billing says 370, nobody is lying. Every system is accurately reporting the slice of reality it was pointed at. The GAO looked at this across the country and found only ~27% of US operations use precision practices analytically, with data fragmentation and interoperability named among the core obstacles. Growers did not refuse the tools, right? The tools refused each other.

Getting the systems to agree is the whole job. Not more dashboards, and not perfect sensors, because there is no such thing. One place where the plan, the applications, and the billing read the same, so the weekly number takes minutes and year-end stops being a season of its own.

Planning season opens in November. The cheapest time to fix disagreement between systems is before the next plan gets written into them. If your July looks like the one above, that is worth a conversation while there is still time to act on it. No pitch, just the conversation.