Digitised US farm financials at enterprise resolution, generated as the byproduct of doing the books. Because every transaction is coded to a farm enterprise (corn, soybeans, cattle, hay) within Schedule F categories, the platform continuously produces real-time profit and loss per enterprise, cost of production per unit, daily cash position, balance sheets and FSA documentation — plus the raw layer: bank transactions syncing daily and receipt images captured in the field and auto-attached. Aggregated across thousands of farms, this is a live cross-section of US farm cost structures — input costs, margins and cashflow by enterprise, region and year — assembled from actual bookkeeping rather than surveys, which is precisely what official cost-of-production statistics estimate by sample and recall.
Equities, commodities in US. Related tickers DE, ADM, BG, CHS, SMG, CW, CTVA, NTR, MOS.
Multimodal but predominantly structured — categorised transaction records, enterprise-level profit-and-loss and cashflow tables, Schedule F report outputs, and receipt and document images. Accounting-grade data, machine-clean at source.
Very dense per farm and compounding yearly: each operation generates hundreds to thousands of categorised transactions and receipt images per year, each enterprise-coded, with daily bank-feed updates and derived profit-and-loss and cashflow snapshots. Thousands of farms over multiple years plausibly reaches tens of millions of labelled transaction records — the compounding nature matters, because each additional year of history converts a cost snapshot into a longitudinal cost panel across price cycles, which is where the pricing power sits.
Bookkeeping is the label factory, and it runs continuously rather than at survey intervals. Every transaction arrives labelled to an enterprise (corn, soybeans, cattle, hay) within standardised Schedule F categories, so cost structures are comparable across farms by construction — the normalisation that survey panels struggle with hardest. Computing cost of production per unit labels each farm-enterprise-year with an efficiency score, and because the mechanic generates it automatically, the label exists whether or not the farmer ever asks. Daily bank syncs label timing — when a farm actually paid for fertilizer, which reveals input purchasing seasonality as behavioural data. Auto-attached receipt images carry extracted line-item content: what was bought, at what price, from which supplier. CPA collaboration adds a professional-review layer that functions as periodic validation of the coding.
Ambrook — Digitised US farm financials at enterprise resolution, generated as the byproduct of doing the books. Because every transaction is coded to a farm enterprise (corn, soybeans, cattle, hay) within Schedule F categories, the platform continuously produces real-time profit and loss per enterprise, cost of production per unit, daily cash position, balance sheets and FSA documentation — plus the raw layer: bank transactions syncing daily and receipt images captured in the field and auto-attached. Aggregated across thousands of farms, this is a live cross-section of US farm cost structures…
Not yet on file: Sample · Data dictionary · Data frequency · Reporting lag · Point-in-time · License