Operate · plan tracking & multi-site benchmark

The plan becomes the thing the week is measured against

illustrative — awaits registration data

Two mechanisms, both fed by data the locations already record: plan-vs-actual per week, and a multi-site benchmark on the same metrics across all five locations. Numbers below are illustrative until the registration streams connect.

This week vs plan — Location M illustrative

Kappen planted

Plan 6

on plan

Kappen harvested

Plan 6 · contract floor 2

contracts covered

Stems delivered

Plan 246,000

238,700

−3% · uitval above table

Realized price

Plan €0.55 /stem

€0.58

+5% vs estimate curve

Cycle age of oldest kap

Standard 72 d this season

74 d

watch — occupancy risk

Re-planning triggers rolling horizon

The weekly MILP re-solves whenever actuals drift from the ledger. Within the cutting lead time (10 weeks rolling, 16 weeks for a variety switch), the plan is committed; beyond it, everything is re-decidable.

Cycle running +2d over standard

re-solve occupancy

Price realization ±10% for 3 weeks

re-fit price curve

Uitval above seasonal table

update success rates

Contract volume change

new harvest floor → re-solve

Data needed to go live intake

Everything on this page runs on streams already recorded on site — see the Calibrate intake list. Connecting weekly exports of planting/harvest registration and Floriday transactions makes this page real, then repeats per location.

Model

Intake

Live

Multi-site benchmark — all five locations, one table

The ask: compare the locations on the same metrics. Location M's column comes from the validated model; every other column fills in by data intake alone (one entry in the case file — no new pages, no new software).

From one location to five

Benchmarking is the on-ramp for the other four

Each location's intake immediately pays back twice: its own plan-vs-actual tracking, and a new column here that explains where the hectares differ.

See the intake list →