PaddIQ/Help & How-to Guide
Open the app →

How to use PaddIQ

Every page, what it's for, and how to get the most out of it. If you only read one section, read Getting started — the rest follows from it.

🚜Getting started — the workflow

PaddIQ runs on one idea: every operation on every paddock is a Field Record, and everything else — the map, the fert plan, the P&L — is computed from those records times your prices. Nothing is typed twice, so nothing disagrees.

How to use it
  1. Set up your farm: Import → Farm Setup, or draw paddocks straight onto the Farm Map (Add paddock/boundary).
  2. ✨ Ask PaddIQ (the round button, bottom right): ask about paddocks, records, nutrient balance, stock, rain or margins in plain words. It reads the same numbers the screens show — every figure comes from your data, never a guess — and it can stage record changes ('add Metsulfuron to the planned knockdown on every wheat paddock on lupin stubble'): you see the list and press Approve before anything is written. Your question and the data it reads go to Anthropic's API; each account has a monthly allowance shown at the foot of the panel (document and photo reading draw on the same allowance, and there is a limit for everyone together each month), and questions are logged so the answers can be improved. If an answer is wrong or comes back empty, press the ! beside it to send that question to the PaddIQ admin.
  3. New account? Paddocks come first: on the Farm Map press Import from John Deere and every paddock is created from its field (named Farm:Field) with its boundary and area, and Deere's farms become your farm list in Settings → Farm & Paddocks with each paddock under its farm, or upload a boundary file or draw them. The Getting-started guide's first button does this for you. Operations and map layers need paddocks to land on.
  4. Get your operations in: connect John Deere (Import → John Deere) for a live sync, drop Excel/CSV exports or PDF field reports on Import → Manual / PDF (John Deere Operations Center reports today; other exporters can be added as sources), or add records by hand in Field Records.
  5. Set your prices: Pricing tab — chemicals, fertilisers, seed, grain prices, overheads, fuel. Everything you imported is waiting there at $0 until you price it.
  6. Check Financials: profit per paddock appears as soon as records + prices exist.
  7. Plan next season: Planning → Rotation, Nutrition, Cropping Cards, Inputs & Margins.
Tips & tricks
  • The order matters less than you'd think — you can import three old seasons first and price them later. Income and costs show $0 until prices exist, never made-up numbers.
  • Anything marked with an amber ⚠ or $0 is PaddIQ telling you it needs a price or a setting — it never silently invents one.
  • On a phone the main pages live in the bottom bar (Map / Records / Weather / Planning); More opens everything else.
  • The Getting started strip in the sidebar counts six setup steps (paddocks, region, prices, machinery, operations, plan) for the active season and links to the next one. It disappears on its own once all six are done; a step ticks as soon as the page it points to has any real data.

🗺️Home — Farm Map

Your whole farm on satellite imagery, colour-coded by crop, with overlays for profit/loss, season progress, nutrients, soil sites and NDVI.

How to use it
  1. Add paddock/boundary → draw point-by-point on the map, or upload a shapefile/KML. Un-arable areas (rocks, dams, timber) can be cut out so areas stay honest.
  2. Uploading boundaries is checked first. A yield map, as-applied map or other machine export is recognised and refused with a note saying what the file looks like — they are all just a .zip from John Deere, and one dropped in here used to replace every paddock boundary on the farm. If you already have boundaries, you are asked whether to Add and update (keeps any paddock the file does not mention) or Replace all.
  3. Constraints (Layers → Constraints): a zones-style map of where inside each paddock the trouble is. Four seasons of peak growth, every pixel scored against its OWN paddock that year, then averaged — so a corner behind in every year, wet and dry alike, is the ground rather than the weather. It reads one farm at a time at about 12 m a pixel, because a pixel any coarser averages a salt patch together with the good country beside it until neither shows. Slower than the backtest; several minutes is normal, and it is kept on file afterwards — reopen the tab and it draws straight away, with Read the imagery again when you want it rebuilt.
  4. The Constraints list gives hectares as well as a percentage, because gypsum, lime and ripping are quoted per hectare. It finds underperformance, never a cause — sodicity, salt, compaction, non-wetting sand, an old sheep camp and a tree line all look the same from orbit. Walk the patch, and take a soil test where it is worst.
  5. The 🗺 Layers button holds everything the map shows: paddock colouring (Crop, Profit/Loss, Season Progress, Nutrients), the basemap (Satellite / Street / Hybrid) and overlays (NDVI, Soil sites). The panel stays open while you flick between views.
  6. Round buttons on the map's edge: Measure (📏) and My location (📍) — on a phone they also include Soil sites and NDVI.
  7. 💰 Gross margin map (Layers → Overlays): every paddock with a John Deere yield map is coloured cell by cell — yield × grain price (less levies/freight) minus this season's direct costs, with as-applied maps spreading their record's cost by the rate applied at each spot and everything else flat. A paddock with costs but no yield map (fallow, sprayed out, hand-entered harvest) is painted flat at its Financials income less costs, so a paddock that lost money still shows red; paddocks whose income is still a projection are left off. White is break-even, red loses, green earns. Two scales: Farm $ (loss and profit sides each stretched to the farm's own spread) or Within paddock (white = that paddock's mean, so its good and poor ground show even when the whole paddock is in profit). Hover any spot for the sum behind it — yield × price, each mapped pass at its rate, the flat costs, and the margin; tap to pin it on a phone. Tap a paddock for its mean, range and the hectares below break-even. The map's mean reconciles to the paddock's Financials margin.
  8. 🧪 Nutrients colour mode IS the applied-nutrient map: kg/ha of N, P, K, S, Zn, Mn or Cu applied at every spot this season, built the first time you pick the mode (paddocks with no as-applied map keep their flat season-total colour); the chips pick the nutrient. Same colour scale as NDVI (white → green → blue → red = light → heavy); farm-wide scale by default, and tapping a paddock stretches that paddock to its own range so its pattern shows (the legend says which scale is showing). Each fertiliser pass is converted with the same product nutrient contents the nutrient balance uses, and spread by the rate its John Deere map recorded at that spot (a pass with no map is flat; outside a map's coverage that pass added nothing). Hover for the passes behind a spot with the kg each contributed; tap a paddock for mean, range and total kg per nutrient and the list of passes (▮ mapped, ▯ flat and why). Herbicide passes add nothing and aren't listed.
  9. 📈 Multi-year margin (Layers → Overlays): every season's margin map averaged cell by cell, in nominal dollars with each real season counting once. Chips pick All years, Cereal, Canola or Lupin/legume — a paddock only shows in a group for the years it grew that crop, and is blank in a group it never grew. Hover a spot for the years behind it (2023 Wheat $412 · 2024 Canola $610 · … → average); tap a paddock for its average per crop group. Building it reads every year's layers, so the first switch-on takes a few seconds.
  10. A dark pill with a spinner appears above the tab bar whenever something is working in the background — imagery loading, a John Deere layer coming down, a margin map building, a season syncing. It says what is happening and survives moving between pages, so a long job is never a blank screen.
  11. Send feedback (sidebar) takes a suggestion or a bug with a screenshot, and now a sample data file — a soil lab’s export, a yield file, a PDF in a format PaddIQ cannot read yet. Up to five files, 25 MB each; only you and PaddIQ’s admin can open them, and they are used to make that format work.
  12. Poor signal: PaddIQ keeps its own code on the device, so the app opens and works on a weak connection or none at all — everything you type is saved to the phone first and syncs when the signal returns. What still needs the network is anything that comes FROM the network: satellite imagery, John Deere, weather and the assistant. If the app ever will not open at all, it is usually a sign-in problem rather than the app itself.
  13. The date scroller sits above the round buttons on a phone and steps through every clear satellite pass — drag it or tap a dot. It follows whichever layer is showing: the whole farm when nothing is selected, and that paddock's own dates once you tap one. Tapping a paddock keeps the map clear; the ℹ button below the scroller brings up that paddock's details and puts them away again.
  14. On a phone the top bar is contextual (Ops-Center-style): tap a paddock and it becomes that paddock's layer picker — any pass, harvest map, or its latest NDVI (freshest image under 5% cloud, dated); yield layers add measure chips (Yield / Moisture / …) under the bar. Turn Soil sites on and the bar swaps to the analyte picker. ✕ returns to season + view.
  15. Measure tool (📏): tap points on the map for running distance; three or more points also show the enclosed area in hectares. Right-click (or ↩ Undo) removes the last point.
  16. Click any paddock for its details: crop, variety, area, season numbers, nutrients applied / planned (N, P, K, S; hover the heading for which products counted, and an amber line names any fertiliser that matched no analysis in your product library), this season's jobs with each John Deere map beside its record (a map whose file names no product is named from that day's record and says so), and the rotation history read from the seeding records of the last three seasons.
  17. The ⋯ menu holds the Setup guide, boundary export (Shapefile / KML / GeoJSON — for an agronomist, contractor or another platform) and Clear all boundaries.
Tips & tricks
  • NDVI (Layers → Overlays) loads the latest satellite image for the whole farm on the high-contrast red → blue scale, stretched to 2–98% of the farm. Click a paddock and the layer clips to it and re-stretches to that paddock's own range so in-crop differences show; click more paddocks to add them to the set (a block reads on one scale), click one again to drop it (the legend reads the range and the paddocks; clear the selection for farm-wide). With NDVI off, clicking the selected paddock again deselects it. ‹ › step through earlier images. For the full imagery toolkit go to Planning → VRT/Soil → Imagery.
  • Paddock areas update when you redraw a boundary, but existing Field Records keep the area they were created with — history doesn't rewrite itself.

🚜John Deere connection

Links your John Deere account so PaddIQ pulls boundaries and operations straight from Operations Center — no more USB or file exports.

How to use it
  1. Import → John Deere → Connect John Deere. You sign in on John Deere's own page and approve access — PaddIQ never sees your password.
  2. Click Enable access next to your organisation, tick it on John Deere's page, and you're connected. This is a one-time setup per device.
  3. ⟳ Sync John Deere is the one button that does the lot: refreshes farm data, imports new operations (clean ones silently; anything needing your call stages behind the Review banner), and tops up missing map layers. The specialist buttons (Load farm data, wizard import, layer backfill, Redraw in HD) live under Advanced.
  4. The wizard import (Advanced) opens EVERY new operation in the Reconciliation Wizard — product matching, rig costs and duplicate protection all apply.
  5. On the Farm Map, ➕ Add paddock/boundary → Import from John Deere pulls all boundaries and auto-matches them to your paddocks; anything unmatched appears in the assign list for a manual match.
  6. It also syncs itself: on opening the app (at most twice a day) new Ops Center operations import automatically when everything about them is certain; the banner on Import shows what happened, with a Review button when some need your call and a Sync now button to force a check. Review opens the wizard on just the flagged ops — 'Show everything' widens it to every new op from that sync. The sync only writes to your current paddocks: operations on fields John Deere still holds but that are archived in PaddIQ, or not in your paddock list at all (a block you no longer farm, a neighbour's job), are held back and counted on the banner. A paddock that was merged or reshaped still takes operations from seasons before the merge, under its own name — that is its history, and the merged paddock reads it. The wizard holds the same fields back, with an Include them button when the work was yours; a field not yet in your paddock list is added to it when its operations are imported, so they show in Field Records. On a brand-new farm with no paddocks yet, nothing is held back and nothing is imported on its own: the first sync sends every operation to the wizard, and the paddocks are listed as you import them. Undo on the banner takes back everything the last sync wrote on its own: records it added are deleted and plans it marked done go back to plans.
  7. 🚩 Mark a point (the round button column) — one button for flags and hazards alike. Drop a flag or note for yourself, mark rocks, rock piles and stumps, or mark what a pilot needs: powerlines and single-wire lines, poles, towers, houses, sheds, silos, treelines, fences, roads and stock routes, dams, watercourses, airstrips, loading sites, beehives, livestock, sensitive areas and no-spray buffer zones. Anything shown with ⚠ goes to the spray contractor in an export; the rest stay yours. “Something else” lets you name your own. A point takes one tap, a line several along it, an area several around it, then Done. Touch anything you have marked and it tells you what it is, the note you attached and the day you marked it; tap it to edit or remove it. The switch at the top of the picker hides everything on the map at once, “Drop where I am” puts a marker on your GPS position without tapping, and List shows everything you have marked — drag a row's grip to reorder it, tap it to edit, or remove it. Dragging a grip in the type list puts the ones you use most at the top, the swatch beside it changes that type's colour everywhere — map, printed page and legend — and the switch on the right turns that type off on the map without deleting anything. Point markers fly a flag whose pole base sits on the spot, so the flag never covers what it marks; an unconfirmed one flies a dashed flag. When you export a farm map you choose which kinds of marker go on it, so a spray contractor gets the hazards without your rock piles and flags. Give a height where you know it, and tick it off once you have confirmed it on the ground — the export says which ones are unconfirmed.
  8. ✈️ Export farm map (⋯ menu): tick the farms you want, pick a format, and it combines them into one map. The Shapefile bundle is what an aerial operator's guidance system wants — boundaries, hazard lines and hazard points as separate layers (a shapefile can only hold one shape type), plus a KML for Google Earth and a read-me. All in WGS84 lat/long. KML on its own is the easiest thing to send someone who just wants to look.
  9. Map layers: Planning → Zones → Get John Deere layers pulls a paddock's yield and as-applied maps (fetched once, instant afterwards); Field Records → Map view shows them per paddock in season order.
  10. The import wizard opens on ⚠ Needs input — the operations it cannot commit without a decision from you. The All pill shows everything including the ones that need nothing, and the other pills narrow to low confidence, fuzzy product names or date issues.
  11. First season, in order: bring the operations in (the John Deere import is also what finds your machines), build your rigs from those machines, then run the import wizard once more. Records freeze their machinery cost on the day they are created, so passes imported before the rigs existed came in at no cost and only a re-run fixes them. Planning → Overview shows these three as a strip that ticks off and then disappears.
  12. Suggestions come from PaddIQ's shared library: common chemicals, fertilisers and crop varieties. They appear as you type and the John Deere importer matches names against them, but nothing from that list enters your own library, prices or records until you pick it. Anything you add yourself stays yours, and is offered to PaddIQ as a candidate for the shared list — if enough farms add the same product, it becomes a suggestion for everyone.
  13. A paddock showing no crop after an import: John Deere does not always name the crop on a seeding operation, and the app otherwise reads the crop from the variety — which needs that variety in your library. The paddock panel asks once, with the variety it was sown with; picking the crop sets the season AND adds the variety to your library, so every other paddock sown with it fills in by itself. Imports also teach the library whenever Deere does name the crop.
  14. Hover any open map layer (a seeding, spray, spread or yield map) and the value under the cursor reads out — seed rate, applied rate, yield — in the layer's own unit, taken from the recorded data rather than sampled back out of the colours. Click to hold the reading, click again or pan to clear it.
  15. Yield detail (Import → John Deere): Standard is the default — the same yield columns at one reading per second, about 5 MB a paddock. Full detail keeps every sensor reading (40–80 MB a paddock, 1.5–2.5 GB a farm-year) and is worth it only for fine zone work. A first fetch on a big farm takes a while either way; the progress line shows layers done, MB pulled and elapsed time, and the run survives moving between pages.
Tips & tricks
  • The connection lives on the device you connected — connect once per computer/phone you use.
  • Long fetches show a progress bar with a per-paddock roll call (D1 8/8 ✓ …) and a small spinner pill bottom-right — you can keep using the app while it runs; the pill means it's still working.
  • Re-running either import is safe: unchanged boundaries are skipped and the wizard recognises operations it has already written.
  • Boundary imports update paddock areas (with a confirm step) — past seasons keep the areas their records were created with.

📥Import — Manual / PDF files and John Deere

Turns Operations Center exports (PDF field reports or Excel/CSV) into Field Records. Four file types: Seeding, Applications, Harvest, Tillage. The Harvest tab also takes ANY results spreadsheet (your own yield/protein/grade sheet, a CBH download): Import a results sheet shows every column and row for checking — which paddock each row matched, the yield it will use (a t/ha column, else correction tonnes ÷ area, else total ÷ area) — then loads protein, grade, moisture, oil, screenings and hay onto the paddocks; yields already on file are kept unless you tick Replace. Correction tonnes are weighed tonnes — scales or docket weight — and are kept on each paddock's harvest record along with where they came from, because they are the only figure that can check the header: its yield map and the John Deere yield are both the same monitor. A column called just "net weight" is kept but not labelled scales or docket; header tonnes are never treated as correction tonnes. Protein, grade and moisture now reach the recommender even on a paddock with no yield recorded yet. Correction tonnes (on the Harvest tab, under the total): enter each crop's season weighed tonnes — scales or docket — for when loads were not split by paddock, and PaddIQ sets it against what the paddock records add up to. It says whether that is a clean header check (every paddock's yield came from the monitor), only indicative (a mix), or no check at all (the paddock yields were already worked out from weighed tonnes). Within 5% is green, 5–15% amber, beyond that red. The gap is shown, never applied. Headers against each other (just below it): where two headers cut the same crop in one paddock, John Deere keeps each one’s own average. One paddock can’t separate calibration from which end each header cut, but across many paddocks the ground averages out and calibration doesn’t — so each pair gets a 95% range from its shared paddocks and a verdict: together (under 7% either way), worth checking, calibration difference, or can’t tell yet. It says the headers disagree, never which is right — only correction tonnes do that. For seasons imported before this, re-run Import → John Deere: it adds each header’s figure to the harvests already on file without changing any yield. If you've connected your John Deere account, the John Deere tab imports the same data with no files at all.

How to use it
  1. The John Deere wizard always writes to the season the data was fetched for, shown in its title — never to whatever another open window has selected. If they differ it says so at the top. And no operation dated two or more years from the season can be written into it at all: it’s refused with a note, and nothing for that paddock is written.
  2. In the John Deere wizard, picking an operation type on an undecided card fills in every other undecided card in the same run with the same John Deere work name — pick Spray once for "Post emerg broadleaf" and the rest follow. Cards already sorted, or set by hand, are never changed. Paddocks you have already confirmed stay as written; later paddocks arrive filled in. The choice is also remembered for future imports.
  3. Drop each file on its card. PDFs and spreadsheets both work — the card shows what it found (paddocks, applications, yields).
  4. Choose your path: the Reconciliation Wizard walks paddock-by-paddock and lets you match passes to plans, pick cost sources (rig or fleet) and fix product matching. Quick Import writes everything in one go.
  5. In the wizard, unmatched products import under their JD name and are added to Pricing at $0 — nothing is ever dropped. Use “+ Add as new product” or map them to library products you already have.
  6. Seed varieties register themselves in your library and appear in Pricing → Seed Prices automatically.
  7. Re-importing the same file is safe: existing records are recognised and skipped, replaced passes are paddock-scoped.
Tips & tricks
  • Fert applied the same day as seeding (MAP down the tube) bundles into the seeding record — one pass, one machinery cost.
  • Each card has Delete imported — removes every record that file type brought in for the season (manual records untouched). Cleanest way to start a season over.
  • Machine costs are snapshotted at import. Fixed a machine's costs afterwards? Delete imported → re-import, and records re-snapshot at the corrected rates.
  • Paddocks in the file that don't exist yet are created automatically under the right farm.
  • Delete season data sits closed at the bottom of the Manual / PDF tab. Open it for per-season buttons: planned only, all records, or everything (records, prices, targets, rainfall) from this device and the cloud. Every button confirms, and Delete everything makes you type the year.

📋Field Records

The single source of truth: every seeding, spray, fert pass and harvest on every paddock, with a cost per hectare frozen onto each record.

How to use it
  1. A seeding pass that sowed several varieties keeps each one at its own rate and area — the card shows every variety, and seed is costed variety by variety at its own price. On the record, “+ another variety” adds a row. Passes imported before September 2026 came in as one joined name at an averaged rate; they show split into their varieties with the average rate and a note, until you type the real rates or re-run Import → John Deere for that season — the import now adds each variety’s own rate and area, and each header’s yield, to records already on file, without creating or duplicating anything. The background sync does the same for the current season.
  2. Each pass shows what it puts on as N P K S pills (kg/ha from its products at the recorded rates — the same conversion the nutrient balance uses); the paddock card’s pills are the season’s total against the plan. Withholding periods: every spray and fert record shows a pill read from the APVMA label of each product — “WHP harvest 14 d · clear 26 Sep”, and grazing where the label gives one — for the paddock’s crop, from the record’s date. Tap it for the label’s own words. A harvest record shows whether the paddock was clear on the day, or which product still had days to run; the Harvest tab shows the same on each paddock. Tap any pill for the label’s own words, the clear date and a link to the label on the APVMA register. Settings → Label markers turns the pills off. Every clear reads “clear · check products” on purpose: it is read from the products and dates as recorded, so a wrong entry makes it wrong — it is a prompt to check the label, not a certification. A product with no APVMA link, or a label wording that could not be read for that crop, shows “WHP: check label” — the app never invents a number. Pricing → Chemicals lists at the top any product not linked to an APVMA registration, and any linked provisionally (~): open one and Confirm keeps the match; “No APVMA registration” marks a name that is not on the register — an unregistered adjuvant, say — so it stops asking you to check a label that does not exist. Link products there (most link on their own: by name, by product family, or for generics like “Glyphosate 450” to a registered product carrying that active at that strength — those show a ~ on the pill until you confirm them; “label ↗” opens the product’s page and labels on the APVMA register). Under each chemical, Alternatives lists the other brands of the same thing you use — Hyperspray under Hasten, Genfarm Imazapic and Implode under Imazapic — each with its own APVMA link; an import that meets one of those names writes YOUR product on the record and remembers the brand, so the season’s totals stay in one line and the label read is the brand’s own. “Same actives, on this farm” suggests candidates, and the import wizard fills the list itself: map “Implode” to Imazapic once and Implode is added under it, with its own registration looked up. Want brands kept separate? Add them as products and list nothing. Fertilisers and ameliorants have the same list (Genfarm urea under Urea) with “same analysis” suggestions instead — they are not APVMA-registered, so no label link. Source: APVMA PubCRIS (data.gov.au), CC BY 3.0 AU; the registered label is the legal document.
  3. Plantback: each paddock card and the paddock side panel on Planning → Overview check every herbicide applied in the last 24 months against the plantback table for the crop the paddock is planned or sown to — “plantback Imazapic: 3 mo or 25 mm” until the label allows it, or “plantback clear”. Months and rain are what is STILL needed: the label’s requirement minus the time and the gauge rain since the pass. The check is on the conventional rule; where the label exempts a tolerant type (Clearfield, TT …) the line says “a CL variety may be exempt — check variety and crop type” rather than deciding for you.
  4. A record’s rig & cost picker always shows what is actually costing it. A harvest costed by the older Harvest Fleet setup shows as “Harvest Fleet (auto) · $/ha (as recorded)” rather than “Pick a rig or fleet”. That $/ha was frozen when the record was made and isn’t changed unless you pick something else. To have new harvest records cost from a fleet by name, set up a Harvest fleet in Machinery → Fleets and mark it the default.
  5. Each paddock card shows N, P and K on its front: kg/ha applied so far over kg/ha planned, e.g. N 53/63. Green means that nutrient's plan is met, amber means it's short, grey means none is planned and only the applied figure shows. Both numbers come from the same sum over your actual and planned records, scaled for part-paddock passes — hover for the detail. A figure that looks about double what you put on is worth checking for a pass recorded twice.
  6. Two views — ☰ List (the classic per-paddock timeline) and 🗺 Map: click a paddock on the map for its season's operations in date order; entries with a John Deere layer show the key metric (t/ha, kg or L applied) and paint it on the map when clicked.
  7. Expand a paddock to see its season timeline. Each row: date, type, products + rates, status (planned/actual) and $/ha. On the phone each row is the products on their own line with the date, status and area under it and the cost at the right; tapping a row opens the editor in a bottom sheet (swipe down or Cancel to close), and + Add record opens the same sheet.
  8. + Add record → pick a type, date, products (from your library, with an add-new option), rig/fleet for the cost, operator and notes.
  9. The variety picker filters to the paddock's crop; typing a new name offers “+ Add to library”.
  10. Search box filters across products, notes and operators. Type + farm filters up top.
Tips & tricks
  • Records need a cost source (rig, fleet, or a CONTRACTOR from Pricing → Contractor rates) — picking a contractor writes the job and its cost into the Contractor bucket automatically. Works for aerial spraying, cartage, anything with a rate.
  • Harvest records show the paddock's receival levy + freight line (rate × tonnes). Grain staying on farm? Hit '✕ No freight for this paddock' and it isn't charged.
  • “+1 more” on a seeding row means fert was bundled in the same pass — expand the record to see it.
  • Costs are snapshots (price and area at the time of the record). Changing prices or boundaries later never rewrites existing records.
  • A paddock's header crop comes from its actual seeding record first — what's in the ground beats what was planned.

💰Pricing

One library for every price: chemicals, fertilisers, fuel & oil (diesel, any other fuels you add, bulk engine/hyd/gear oils), ameliorants, seed, grain, overheads, leases and contractor rates. Season-scoped — each year keeps its own prices.

How to use it
  1. Work through the tabs. Products imported from JD are already here at $0 — just fill in the numbers.
  2. Seed Prices fills itself with every variety you've seeded or planned this season.
  3. Grain Prices drive income: each crop needs at least one price with a real number, or income shows $0.
  4. Fill prices from history copies a previous season's prices in — then adjust what moved.
  5. Filing cabinet (also in Inventory → Coming in — same cabinet): every invoice that enters the app, grouped by supplier then by what it mostly was (herbicides, fertilisers…), dated, with the original file. Type 'glyph' in the search to see every invoice with glyphosate on it. An invoice priced here is filed 'prices only'; receiving it in Inventory upgrades the same card. An account statement with tonnes and amounts (CSBP) prices each row too and files one card per document number — upload the real invoice later and it replaces that card by number, no portal needed.
  6. Fuel & Oil: log every delivery (or 🧾 upload the fuel invoice — each diesel/petrol/oil line is matched to its log for review) and 'use weighted average' sets the price from real purchases.
  7. Grain Prices: one table of crop × grade × channel. Per crop the highest Farm Gate price wins, else the highest Market price — that drives Financials income, harvest load values, card gross margins and the N economics. 📷 Snap price sheet reads a LoadNet / DailyGrain / broker screenshot into a review list; nothing saves until you Apply.
  8. Import price list / invoice takes a supplier CSV/XLSX, or an invoice PDF/photo (same AI reader as Inventory), and matches products by name — matched products get the invoice's per-unit price, unknown ones are added. For an invoice you can Save prices only, or Save + add to Inventory, which opens the same invoice in Inventory's receive sheet to book the stock in.
Tips & tricks
  • Enter prices as $/kg, $/L, $/tonne or per-pack — the table converts and shows the per-unit rate.
  • A product's Type controls which tab it lives on. “— set type —” in amber means PaddIQ couldn't guess — pick one.
  • Log purchases on a product to track actual invoices against the season price.
  • Levies & Freight tab: set each crop's receival levy $/t plus one freight/GFR $/t (site-based). Every harvested tonne is charged (levy + freight) against its paddock — the Receival Levy bucket in Financials.
  • Products, varieties, grain price rows and their Types are permanent once used: they appear every season at $0 until priced (grain rows keep their receival levy); deleting removes from this season and stops carrying forward (past seasons keep their data).
  • Contractor tab holds rates only — jobs are created from Field Records or Cropping Cards by picking the contractor as the pass's cost source: choose “👷 Contractor…” in the machinery dropdown, then which rate. The cost bills through the Contractor bucket, never machinery.
  • Prices apply from now on. Existing records keep their snapshot — that's deliberate (your March urea pass shouldn't change cost because June urea is dearer).

📅Planning — Overview, Rotation, Inputs & Margins, Cropping Cards

Season planning end-to-end: what goes where (Rotation), what it needs (Nutrition), what it needs, costs and returns (Inputs & Margins), and reusable crop programs (Cropping Cards).

How to use it
  1. Overview is the landing view: every paddock's crop, target yield, planned $/ha and status chips (Sown / Sprayed / Spread / Harvested). Click through to any paddock's detail panel.
  2. Rotation: the year-by-year crop matrix. Set this season's crop and variety per paddock; past years show actual yields.
  3. Varieties and timings are picked from your library everywhere (Rotation, Cropping Cards, Field Records) — choose “＋ Add …” at the bottom of the dropdown to type a new one; it back-adds to the library and Pricing → Seeds.
  4. Target yields: set them in Rotation/Overview — until you do, PaddIQ suggests one from that paddock's own history (3-yr same-crop average), then the farm average, then a conservative crop default. The badge tells you which.
  5. Cropping Cards: build a full-season program (seeding pass, sprays, topdress) once, then apply it to any set of paddocks — pick them from the list, “Select all <crop>”, or 🗺 Pick on map (click paddocks on the farm map) — it writes planned Field Records.
  6. Card passes collapse to one line each (op · timing · products · $/ha) — click a line to edit it, ▴ Done to close, ⧉ Duplicate to copy a pass. Each product line shows its own $/ha from your Pricing-tab prices.
  7. Cards never double up with the nutrition plan: a card product matching a planned fert row (same product + timing) merges it into the card's pass. Tick “Use recommender's rates for matching products” and the plan's per-paddock rate wins; unticked, the card's rate wins — merged either way. DTT-timed plan ferts ALWAYS fold into the seeding pass, even products the card doesn't list (all DTT goes down with the drill).
  8. Planned rows in Field Records have quick actions right on the row: “✓ actual” converts the plan to a done pass, “✕” deletes it — no need to open the editor for plans that never got a pass.
  9. Inputs & Margins: three views. Margin what-if — yields start at the area-weighted average of your paddock targets, grain prices at this season’s bids (or the latest season that has them), inputs at seed + fert + chem + machinery per crop from the plan; type over any cell. Order list is the fully costed plan by paddock and by product (seed is costed from rate × Pricing $/kg until a seeding record exists); products the Nutrition recommender suggests that you have not yet accepted appear greyed and tagged recommended, with their own subtotal, so the buy list is complete before every paddock is accepted — open Nutrition once so the snapshot exists. Plan vs History compares against one year’s actuals, or Same-crop average: each paddock against its own history of the crop it is planned to (blank where it never grew it). Every view has Export CSV.
Tips & tricks
  • A season you are planning gets one planned Seeding record per paddock with a crop (variety and seed rate from Rotation) the first time its fert plan is saved — open it in Field Records to change the rate, and every planned line carries the product's catalogue unit (Flexi N in L/ha, Agflow Xtra in kg/ha). Run all repairs renames aliases like plain 'Agflow' to the catalogue name and fixes planned units. Once harvest records exist, Inputs & Margins income switches from target-yield estimates to actual yields automatically.
  • A paddock you seeded but never planned still shows up everywhere with its real crop — actual records always win.
  • Build next year's card from a past season: it collapses that year's actual passes into a representative program.

🧪Planning — Nutrition

Per-paddock nutrient accounting (applied vs planned, kg elemental/ha) plus an economics-aware fertiliser recommender built on your soil tests, field records and prices.

How to use it
  1. Grain protein is turned into nitrogen removed on the basis the receival lab reports it: wheat is N × 5.7 at 11% moisture, barley is N × 6.25 on a dry basis (GTA standards), brought back to the tonnes delivered using your recorded moisture (11% if none). Barley used to be treated like wheat, which put its protein-based N removal about 23% high. Barley’s N-status label and residual-N credit still use wheat’s protein thresholds — barley-specific ones are on the list.
  2. Organic-matter nitrogen follows the season’s rain: growing-season rain × OC% × 0.15, on each paddock’s own gauge, in both the recommender and the Nutrient Balance. Until September 2026 the recommender used a flat rate per soil type, which gave the same credit every year — in a dry year like 2023 it counted about 70% more nitrogen than was released, so it recommended too little. Expect dry-year N recommendations to rise by a few kg/ha and wet-year ones to fall. The OC mineralisation rate in Assumptions is now only the fallback for paddocks with no rainfall history. The 0.15 is a Victorian figure, not yet calibrated for WA.
  3. Fert Plan: the grid shows each paddock's N/P/K/S units — big number applied, small number planned. Click a paddock for its full program.
  4. Add products to a paddock's plan with timing (seeding vs topdress); the nutrient breakdown updates live.
  5. The 🧭 Recommendation wizard (Fert Plan, top right) walks every paddock one by one and shows its workings in order: what's in the soil, what last year's crop removed, where that leaves each nutrient vs its critical level (build/maintain/mine), what the target yield needs, and what to put on. Accept & next writes each plan as planned Field Records. A preflight checklist under the button shows what it still needs before it can run.
  6. The N·P·K·S cards at the top of each wizard step re-total live as you edit: they show what the current program supplies vs the rec (with % coverage). Every program row has an editable rate AND timing, a ✕ to drop it, and the green row at the bottom adds ANY product from your library — pick gypsum for S and the cards show exactly what it supplies before you add it.
  7. Legumes (lupins, peas, chickpeas…) get N rec 0 by design — they fix their own nitrogen. The wizard says so on the N row rather than leaving a bare zero.
  8. Nutrient Balance: soil-bank view — what tests say, what crops removed, what you've added.
  9. Nutrition setup has moved to Settings → Nutrition Setup: preferred products, removal rates, product nutrients, your equipment (what the seeder can carry) and the assumptions the recommender runs on. It’s the same screen as before — anything you set there applies here straight away.
Tips & tricks
  • N recommendations FILL YOUR TARGET YIELD: demand at target minus soil supply (test N + OC mineralisation + legume/protein credits) plus stubble tie-up. The economic-optimum response curve (grain price vs N price) is shown as ADVISORY — it runs on industry priors, not your farm's results, until enough seasons accumulate for PaddIQ to fit it from your own rate/yield/rainfall history.
  • The optimum chart (open a paddock) draws the paddock's own yield-vs-N curve for an average finish, and behind it the same model run against every recorded finish at the paddock's rain gauge (or the farm's station): the shaded band is where 20–80% of finishes land, the dotted line the median, the amber marker the rate with the best EXPECTED margin across all finishes, and the row under the axis the share of finishes in which each rate out-earns nil N. It splices each historical year's remaining months onto this season's rain to date and prices haying-off from that year's Aug–Sep, so a dry finish counts against the top rates. It is not a soil-water simulation (Yield Prophet); it is the app's own fitted model run once per year of record. Rainfall: SILO, Queensland Government (Long Paddock), CC BY 4.0.
  • The data-confidence badge (High/Med/Low) tells you how much guessing went in: old soil test, missing yield, borrowed analysis all lower it.
  • Import Soil Tests takes the lab’s CSV or Excel export, a PDF report, or a photo of it, from any lab. CSV and Excel are read on your device; a PDF or photo is read by PaddIQ’s AI. Either way you get a review sheet first: every column in the file, its unit, its first value, and what PaddIQ will read it as — fix any column from the dropdown and tick Remember so that lab is right next time. Units are converted on the way in (ppm, cmol/kg, mS/cm all become PaddIQ’s units). A file that cannot be read is explained on screen and its column names (never the values) are reported to PaddIQ automatically so the reader can be taught that layout.
  • Products with no nutrient analysis count zero units and get flagged — fill their analysis in the product library.
  • Legume history earns an N credit automatically when last year's paddock grew lupins/peas — scaled by that crop's actual yield.

🛰️Planning — VRT / Soil (Soil Map, Imagery, Zones)

The precision-ag studio: soil sampling on the map, free Sentinel-2 NDVI back to 2017, and a zone builder that turns layers into variable-rate prescriptions.

How to use it
  1. Grain samples by zone (Zone Studio → zone card → + sample): at harvest, pull a bag of grain at one of a zone’s numbered sampling dots and record its protein, moisture and screenings. It’s the only way to get protein zone by zone without a protein sensor on the header — receival protein is one number for the whole paddock. Each sample is saved against the dot’s location, so if the zones are ever redrawn it lands in whichever zone now covers that ground. The card shows each zone’s average protein and how many samples it’s from.
  2. Soil Map: every sample site plotted and colour-coded by analyte (Colwell P, pH, K…). Click a paddock to filter to its sites; pick the year or Latest.
  3. Soil Tests / Tissue Tests: the full lab tables. Import lab CSVs — sites, depths and years are preserved; each year archives separately.
  4. Imagery: pick paddocks → load a single date, build the whole season's biomass curve, or compare 5 years of peaks. Save any layer to the paddock's library (Save to Zones).
  5. Four indices, picked with the chips beside the colour ramp and carried on the saved layer: NDVI (green biomass — flattens once the canopy closes), NDRE (red edge — chlorophyll and canopy nitrogen, keeps working after canopy closure; the mid–late season index for N), MSAVI2 (soil-adjusted — early cover on pale sand where NDVI mostly reads the soil) and NDMI (shortwave infrared — canopy water content, moisture stress and drying-off). NDRE and NDMI use 20 m bands, so they are a little coarser than NDVI. The season curve stays NDVI.
  6. Zones: each paddock's saved-layer library. Zone studio generates 2–5 zones from any weighted blend of layers, or draw them by hand with full vertex editing.
  7. Layers across devices: layer definitions sync with your account; the pixel files are copied to your cloud when fetched (John Deere, open layers, their measure files) and any other device pulls a file the first time it needs it — the Farm Map margin and nutrient views, the Zone studio and the Zones library all do this. This also happens automatically: about 20 seconds after you open the app, any layer file on this device that hasn't been sent yet goes up in the background (the ☁ button on the Zones tab does the same on demand). A file over 4 MB compressed is skipped and needs fetching on each device. Deleting a layer removes it from every device: the deletion syncs with the layer list, so another device's copy can't bring it back. If a layer was deleted on one device but its file is still on another, that device's Zones tab shows it under 'deleted on another device' with a Restore everywhere button. Soil colour and radiometrics are one per paddock; duplicates left by older versions are merged into one copy the next time the Zones tab opens, keeping its files.
  8. 🌏 Get open layers (per paddock, Zones tab): two free national layers fetched straight from Geoscience Australia for zoning — Soil colour (Barest Earth: every Sentinel-2 pixel at its barest 2017–2020, 10 m, so colour = soil; brightness separates pale sand from darker loam, redness picks iron/red soils) and Radiometrics (airborne gamma K / Th / U, ~100 m — coarse but real). Both CC BY 4.0, attributed on the layer. There is one of each per paddock (per source for soil colour); re-fetching replaces it, and duplicates left by earlier versions tidy themselves the next time the Zones tab opens. John Deere exports also carry the receiver's Elevation — new syncs save it as a measure, the library shows one ⛰ Elevation row per paddock (from the newest layer that has it, harvest preferred), and the studio offers it plus a Relative elevation surface (low spots − / rises +). Every layer preview draws the paddock boundary over the snapshot. An as-applied layer's headline is the applied rate; when Deere logged none it falls back to the target rate, never fuel. All of these appear as tick-able sources in the Zone studio alongside yield, biomass and as-applied maps.
  9. Computed sources in the Zone studio (no fetch needed): 📈 Yield productivity — each season's yield map divided by that season's paddock mean, then averaged, so 1.3 means 30% above the paddock every year; 📉 Yield stability — how much that ratio swings between seasons (low = steady ground, high = the season decides; needs 3 seasons to trust); 💰 Gross margin for the current season and the multi-year average — the same grids as the Farm Map overlay. Zoning on productivity + stability is the published way to find ground that reliably earns inputs versus ground that reliably doesn't.
  10. Zone studio v2: sources are grouped (yield & economics, biomass, soil & terrain, radiometrics, as-applied), each group folded until you open it (a group with a ticked layer opens itself); as-applied folds year → pass kind (seeding, fertiliser, spray). Layers carry the same names as the Zones library. The left column scrolls on its own. A Layer slider top-right of the map sets how strongly the ticked layer shows over the satellite image; each kind of source paints in its own colour family (yield red → green, biomass straw → green, moisture blues, soil browns, radiometrics purples, elevation green → brown → pale), with the family's swatch beside the slider. Hover a zone on the map (or its row in the list) and it lights up with its name and hectares while the others fade. The Biomass group is always listed; with nothing saved for the paddock it tells you to save a layer on Imagery first. After Generate, a Breakpoints bar sits along the bottom of the map: the cut points between zones on the ticked layer's own scale (t/ha for a yield map, NDVI, or 0–1 for a blend) with each zone's share of the paddock. Drag a handle and the zones re-cut on release; Auto hands the cuts back to the method. A Rates block under the zone list takes one product and a rate per zone (kg/ha, t/ha for lime and gypsum, or L/ha — tonnes is a display choice, the record still carries kg/ha) and shows hectares, product needed per zone, the paddock total in kg, t or L, its cost at your Pricing price, and the flat rate it equals. Attach to picks where the rates go: a new planned record of the purpose's type, or one of the paddock's existing planned records — a seeding record varies its seed rate. Either way the layer saves and the record carries the flat rate with each zone's %, which is what the Rx export and the plan read (seed rates export as a SEED column). The sketch also saves with the layer, so linking that layer from a record's 🗺 VRT chip in Field Records prefills the zone % when the product or seed matches. Harvest and as-applied grids are gap-filled across the passes before zoning, so a 1 Hz yield map gives whole zones, not ribbons. Older 1 Hz harvest maps that showed stripes between passes are re-drawn gap-filled the first time they open. ✨ Suggest lets the data pick the number of zones (k-means 2–5 scored by silhouette; fewer wins unless more is clearly better). 🧭 Stability classes → zones (3+ yield seasons) makes the four Blackmore classes directly: high/low productivity × stable/unstable. Set your machine width: generating removes every zone finger or sliver narrower than that width (the ground joins the neighbouring zone) so nothing on the map is thinner than the bar or boom that applies it, and the studio still flags a zone whose average width is under one pass and any layer with more than four rates. Smooth zone edges rounds the raster staircase twice over. 📋 Zone card shows, per zone, the mean of every ticked source, yield productivity and stability, the soil-test sites that fall in it (several sites are averaged when they agree, outliers named) and three numbered sampling points sitting deepest inside the zone — sample there, topsoil and subsoil. Tag the layer's purpose (Nitrogen, Bar fert, Lime/gypsum, Seed) on save. ↶ Undo (or Ctrl+Z) steps back through every zone change — generate, re-cut, draw, vertex edit, simplify, add, remove. To draw into a particular zone, press Draw polygon then click that zone's row in the list before drawing. A class puts every pixel of the same value in one zone wherever it sits, so two different soils that happen to read alike land together — 🎯 Move parts fixes that: click the row of the zone they should join (Add zone first if they need their own rate), then click each separate piece on the map and it moves across whole. Blending in Soil redness or Radiometric K with the first layer usually separates pale sand from red loam before you get to that. On the Farm Map, a margin overlay's paddock panel has 🗺 Make zones from this, which opens the studio with that map pre-ticked.
  11. Link a zone layer to a planned fert record (the 🗺 VRT chip in Field Records) with per-zone % of the flat rate.
  12. 🎯 Rx (Prescriptions): every VRT-linked plan for the season. ⬇ Export Rx downloads the John Deere USB zip (unzip at the USB root, it holds the Rx folder). 🚜 Send to Ops Center uploads the same file straight into your Operations Center Files — Foundational tier, no Ultimate — then in Ops Center open Files and press Send to machine; the file is read back after upload so 'sent' is never claimed blind. Tick several plans and Send all ticked does them one after another. Plans from every season are listed, grouped by year then operation type; inside a group rows sort not-sent first, then plans edited after they were sent (amber, 'changed since sent'), then sent (green); the All / Not sent / Sent chips filter. 🗑 deletes a prescription: one the Zone studio created goes with its planned record; one linked from Field Records keeps the record and just loses the zone link. Nothing already in Ops Center is touched. Send again uploads a fresh copy — delete the old one in Ops Center Files. Set 'rate outside prescription' on the display to the plan's flat rate.
  13. 💰 Margin by zone: preview a zone layer and press Margin by zone — one row per zone with hectares, yield, gross margin $/ha, the gap to the paddock mean, the total, and the hectares below break-even, for one season or the multi-year average (All / Cereal / Canola / Legume). Same engine and figures as the Farm Map margin overlay; 'Outside zones' is ground the layer doesn't cover. Use it to decide which zone deserves more or less input before you set the per-zone %.
Tips & tricks
  • “Stretch to crop” rescales colours to the 2–98% range inside your paddocks — differences stay visible even when NDVI saturates late season.
  • Peak biomass = each pixel at its season maximum. Multi-year averaged peaks are the most stable base for zoning.
  • Layer definitions sync with your account; the preview images rebuild from fresh satellite data on any device — nothing is lost.
  • Grey bars under the season curve are cloud cover — points on those dates are less trustworthy.
  • Generated zones too dense to edit? ✨ Simplify edges in the zone studio reduces every zone to its corner points (run it again for a stronger pass), then Edit vertices gives you corners + midpoints to drag.

📈Financials

The season P&L: gross profit, input costs, yield and income for the farm, each block and each paddock — recomputed live from Field Records × Pricing.

How to use it
  1. Harvested grain is priced by the grade it went. Each harvest record’s tonnes are valued at that grade’s price in Pricing, so a paddock recorded as ASW1 is booked at the ASW1 price, not the crop’s best one. Grade names match however they’re typed (ASW 1 = ASW1). Where there’s no grade, or the grade has no price in Pricing, the crop’s top price is used and the income line says so — open a paddock’s income to see which. Projected crops have no grade yet and use the crop’s top price. The per-crop price shown is the realised one its income was built from, and the margin map and farm history use the same prices.
  2. Top cards: whole-season totals. Farm comparison bars: $/ha profit per block.
  3. Filter by farm, crop or variety. Sort and group the paddock table however you read it.
  4. Expand a paddock for the category cards — seed, fert, chem, machinery & fuel, contractor, harvest income — each drilling down to the actual passes behind the number.
  5. Picking a product anywhere (cards, tank mixes, records, nutrition) uses one picker: search by name or active, or drill Chemicals › Adjuvant › Hasten. Not in your library? Pick it from the built-in catalogue or ＋ add it — it lands in Pricing at $0 under the right Type.
  6. Export Season Plan / Export Paddocks for spreadsheets that leave the app.
  7. Sensitivity: Price Assumptions sliders re-cost every crop's breakeven — Diesel $/L, Urea and Compound $/t, Chemicals in $/ha (the farm's area-weighted chemical spend is the base; every crop scales by the same ratio), Flexi N per 1000 L, Overhead $/ha. Each crop's card has its own Yield slider (50–150% of its average) that moves the highlighted row and the Margin box without touching the other crops. Reset all puts everything back.
Tips & tricks
  • Income = tonnes × your grain price. A crop showing $0 income almost always means no grain price is set for it in Pricing.
  • Current-season paddocks without harvest project income from target yields (marked as projected). Past seasons use actuals only.
  • If a number looks wrong, expand the drill-down — it shows the maths (rate × price × area) for every line.

🚜Machinery

Your fleet's real cost per hectare: machines, implements, rigs (tractor + implement combos) and fleets (harvest gangs, truck teams), with ownership, fuel, labour and R&M built in.

How to use it
  1. Machines: add each tractor, header, sprayer with purchase price, life, salvage, insurance — money fields start at $0, fill in your real numbers.
  2. Connected to John Deere? The blue JD Machines Discovered panel lists everything in your Operations Center equipment list plus any machine seen in imported operations. ＋ Add all from John Deere creates them in one go, typed by Deere's own classification (tractor, combine, sprayer, air cart, bar, head → machine or implement); anything of a Deere type PaddIQ does not map stays in the panel for you to add by hand with the right type. Purchase price and work rates start at the type defaults — fill them in after.
  3. Implements: bars, carts, trailers — they attach to tractors as Rigs.
  4. Rigs: tractor + implement = the cost source Field Records snapshot for seeding/spraying/spreading.
  5. Under a tractor's Operations table, '＋ split / duty…' splits a category into one row per implement (each bar with its own ha, ha/hr and L/ha — the rig for that implement reads its row) or adds an hours-only support duty like filling the seeder, whose $/yr is spread onto that category's $/ha.
  6. 🔧 Maintenance: the workshop view — every machine's services, repairs, parts and hours, sorted tractors → headers → trucks → sprayers, newest first; tap a machine for its full log, or export the whole log as CSV. Trucks/road vehicles can switch their service meter from hours to km (⇄ toggle next to the interval).
  7. 📷 Scan a part / fix: photograph the part box (or the finished job) inside a machine's log — AI reads the part name and numbers off the label and pre-fills the record with today's date, the machine's engine hours and the photo attached. Add the cost, hit Save.
  8. Fleets: multi-machine teams (harvest: header + chaser + trucks). Workers are area-weighted, support units flat-added.
  9. Maintenance log per machine: each record has a job description ("replaced rear spreader bearings"), the part, its part numbers, invoice and photos — cost feeds R&M. List bulk-oil litres on a service (your own oil types from Pricing → Fuel & Oil) and they're costed automatically; each record freezes the price list it was saved with.
  10. Cost auto-fill: type or scan a part number and the cost box fills itself from the last time that part was logged on any machine, else from a filed service invoice that lists it (Inventory → Coming in), else from the same part name — the green line under the box says which. Type over it if the price has moved. John Deere's parts catalogue isn't reachable through the Operations Center API, so there is no live dealer price.
Tips & tricks
  • JD imports carry actual fuel per operation — those beat the estimates whenever they exist.
  • Contractor-mode ops (flat $/ha or $/yr) cost at the contract rate with no labour on top.
  • One physical pass never charges twice: fert applied through the seeder in the seeding pass carries no separate machinery cost.

📦Inventory

What is in the shed, what is on its way, and where each drum or tonne physically sits. Three tabs: Stock (every product and how much you hold), Coming in (orders not yet delivered, then every invoice filed), and Places (your sheds, tanks and silos and what is in each). Stock goes OUT by itself: saving an actual Field Record draws the product down — you never type usage twice.

How to use it
  1. Stock: every product with its on-hand figure. Tap one to see where it is coming from (Unassigned first, then each place), the drums it is in, average cost, stock lots, and a plain-English movements log — 'Used 40 L · Home:01 · 6 May', 'Received 2 × 1000 L', 'Corrected −40 L · stocktake', 'Moved 1000 L to Tank 2'.
  2. Stock comes in three ways: Add stock (typed in), Upload invoice (each line becomes a stock lot, the invoice is filed, and its price feeds the Pricing purchase log once), or ticking an order in on Coming in. Every one goes through the same door, so cost is carried whenever it is known.
  3. Correct the count is the only manual out: a stocktake difference, a spill, or stock lent to a neighbour. It needs a reason. Do not use it for applications — those come off automatically from Field Records.
  4. Needs attention (a chip on Stock, with the count on the tab) shows negative stock, what the season plan still needs (need / on hand / on order / to buy), and products that have run dry. Negative usually means old spray records posted before inventory started — Zero it to write that off; the spray records stay. Better still, set Count stock from (More) to the day you started counting: usage before that date is never posted, so a Rebuild from records can't dig holes for past seasons. Rebuild also clears earlier write-offs (they answered the usage being re-posted), so after a rebuild only genuine shortfalls remain negative.
  5. Coming in: add a supplier, log what you have ordered (quantity, unit, expected date, and the agreed price if you know it). When the truck arrives tick Received → into a place; if the order had no price the stock lands at no cost and the toast says so — set it in Pricing. Uploading the delivery's invoice instead receives the stock AND closes the matching order.
  6. Places: one card per shed, tank or silo. The emblem fills from the bottom as the place fills — yellow for seed silos, blue for liquid tanks, red for chem sheds, brown for fert sheds — with a fill bar when a capacity is set. A silo also carries the variety it holds and the season it was entered for (set both in its sheet); the variety shows on the card (stock in another unit is converted with the product's density; anything with no density is listed as 'not counted'). Every card has two doors: What's in it (type what the silo, tank or shed holds right now — 20 t, 15,000 L — and the ledger is corrected to match with one dated row) and Add stock here (a delivery with cost and invoice, received straight into that place). Tap a place to set which farms it serves, move stock (one container at a time when sizes are known, or 'Move all' which opens the move sheet with the whole amount filled in), send it back to Unassigned, or write it off. If a Stock chip like Seed comes up empty, In the shed is hiding products with no stock yet — Show all lists them, or count them into a place.
  7. Unassigned is stock with no shed yet. Its one rule: it is used first, before any shed's stock — sort it into a place when you know where it went. 'Draws from' on a product is Automatic (Unassigned first, then its own place) unless you pin a place.
  8. Stock opens on In the shed — only products with stock on hand or any movement; All products shows the whole library (a search or Needs attention looks at everything). The three everyday actions sit on top (Add stock, Upload invoice, Correct the count); ⋯ More holds Add product, Tidy list, Fill details from John Deere and Rebuild from records.
  9. Tidy list: the Stock list starts with every product in your library. Press Tidy list, tick the products you actually keep, then Keep only ticked — the rest are hidden from Inventory (they stay in Pricing and keep any stock). Show hidden brings them back; tick and Unhide to restore one.
Tips & tricks
  • Red buttons delete or take stock out — 'Write off (takes it out of stock)', 'Zero this', 'Delete order', 'Remove this location'. Removing a place never loses stock: everything that pointed at it moves to Unassigned first.
  • Rebuild from records (red) deletes every automatic usage movement across all seasons and re-posts them from your Field Records. It does not change purchases — only which paddock and shed each use is attributed to. Do it after fixing rigs or sheds, and read the count in the confirm box first.
  • A ≈ on a place figure means that use was not tied to a shed, so it is spread across the places holding the product. Received amounts and moves are always exact, and the places always add up to the farm total.
  • The stocktake reminder (EOFY and new year) is one line — Done records it, Dismiss hides it for the season. Stock carries across seasons, so the half-yearly walk through the sheds is the truth check.
  • The duplicate-invoice warning fires on invoice number, or supplier + date + total — not on file name, so two phone photos called IMG_0001 do not false-alarm.

🌧️Season Tracker

Rainfall, weather and water-limited yield for your nearest SILO station — no paid weather subscription.

How to use it
  1. Pick your town, Fetch Data. Monthly rain lands against long-term averages, with GSR months marked.
  2. Weather view: the Right-now card gives a big GOOD/MARGINAL/AVOID spray verdict with wind, Delta-T and inversion; the green pills list your best upcoming spray windows; the compact chart underneath is the hour-by-hour model detail. The 16-day outlook sits second; model comparison last.
  3. Gauge Override: type your own rain-gauge reading for any month — it feeds GSR and the yield estimate while the station data stays untouched.
  4. Rain gauges (Rainfall): add a gauge per location, drop its pin on the map and tick the farms it serves — Home on the Kalannie gauge, Dodds and Painters on the Painters gauge. Type each month's reading; blank months fill from SILO at the pin. Every farm on a gauge then reads its own rain for growing-season total, the water-limited yield ceiling and N mineralisation, and the Overview shows GSR farm by farm. Farms on no gauge keep using the station table.
  5. Yield Estimate: French–Schultz style water-limited yield per crop, adjustable WUE and soil evaporation.
  6. Satellite Yield: a SECOND, independent estimate that works from satellite imagery instead of rainfall — it measures how much sunlight each paddock canopy actually caught, converts that to biomass, then to grain. It knows nothing about the rainfall ceiling, so where the two agree that is real corroboration rather than one number shown twice.
  7. Start with the backtest button. It runs the model over harvests you have already cut and prints how far out it was, separating BIAS (a consistent offset, which is a fixable parameter) from TYPICAL MISS (scatter, which is noise). Read that before you read this year estimate, then use Refit harvest index to pull the model onto your own numbers.
  8. This season gives a RANGE, never a single number: the canopy is measured but the finish has not happened, so every paddock is run once per historical finish percentile and once off the 16-day forecast. Tap any paddock row to see how each finish plays out.
  9. Red-edge trial (the switch above the backtest buttons): NDVI stops telling a good crop from a great one once the canopy closes — its response to intercepted light is measured to fall away past about 0.7, which is exactly when a good crop is setting its yield. The red edge keeps responding. Switching the trial on reads a second band on every pass, so every sweep roughly doubles, which is why it is off by default.
  10. It is not a straight swap. The red edge is weaker over bare and part-covered ground, and it tracks chlorophyll — so a full canopy that is nitrogen-short reads low on it while still catching all the light. PaddIQ keeps NDVI below 0.70 and hands over to the red edge above it, ramped so the two never meet as a step, with the bridge between them fitted on your own paired passes rather than borrowed from overseas maize trials.
  11. The red edge is only ever allowed to push the estimate UP, never down. It was brought in for one job: above canopy closure NDVI stops responding and under-reads the light being caught, so in that range the truth is at or above what NDVI says. A red-edge reading that comes back lower is the known confound — red-edge indices track chlorophyll, which is why they are used to map crop nitrogen, so a closed canopy that is nitrogen-short reads low while still intercepting every photon. That is a real signal in the wrong term: nitrogen belongs to conversion efficiency, not to light interception.
  12. The trial keeps both runs on file and compares them season by season. The test is not whether the average improved — it is whether the gain lands on the seasons whose canopies actually closed. A change that improves every season equally is a rescale, and moving a constant is the cheaper way to do that. PaddIQ says which of those it is.
  13. Harvest index moves with the finish — the same biomass yields far less into a hay-off than into a soft spring. Paddocks too small to measure at the imagery resolution are listed as skipped rather than given a confident number.
  14. The satellite passes are stored separately from the scoring, because they have different lifetimes: what Sentinel-2 recorded over a paddock in August 2024 is fixed forever, while the score changes every time you correct a harvest, refit a harvest index or reclassify a paddock as hay. So those changes now re-score in seconds and read no imagery at all. Stored passes are only thrown away by the three things that genuinely change them — a redrawn boundary, a change of resolution, or the season still being under way, where new passes keep arriving every few days.
  15. Switching the red-edge trial on does not re-read NDVI either. The NDVI in a stored season is exactly the NDVI the trial would fetch, so only the missing band is read — the trial costs one extra band, not two.
  16. A finished season stays on file once it has been read, so the backtest opens instantly instead of sweeping the imagery again. Correct a harvest record, redraw a boundary or refit a parameter and it says the stored run is out of date — use Read the imagery again to rebuild it.
  17. Rows worth a look: after the backtest PaddIQ flags any paddock-season that does not sit with the rest of its crop — either it carries most of that crop's whole error on its own, or it is a long way from the others. It flags and stops there. Salty country, a frost hollow and a genuinely exceptional paddock all look the same from here, and two of those are exactly what the model should be learning from.
  18. Tap a row to open it and rule on it. Set a row aside only when the RECORDED figure is not what the paddock grew — only part of it was cut, it went for hay, the yield was booked against the wrong paddock, the area is wrong, or a one-off event like frost. A genuinely poor paddock should stay in: that is the model being told something true.
  19. Nothing is ever deleted. A row you set aside keeps its place in the table, greyed, with its reason on it, and it comes straight back out if you change your mind. It drops out of the score, out of the learning and out of the harvest-index refit at once — no re-reading of imagery, because the pixels never changed.
  20. This matters more than it sounds. One wrong harvest record does not stay in one place: a single part-harvested lupin paddock was inflating the lupin constant, that farm's lupin term and the paddock's own factor at the same time — three corrections all chasing the same phantom.
  21. Whose error is it: after the backtest, every miss is split four ways at once — the season, the crop, that farm's version of that crop, and the paddock — all fitted together rather than one after the other. They compete for the same error, which is what makes it work: when a crop constant tries to swallow one farm's offset, the paddocks growing that crop elsewhere pull it back, so the offset lands on the farm where it belongs. Fit them in sequence instead and whichever goes first takes the lot.
  22. Why the split is worth it: the crop level is a constant that ships to every farm, while the farm and paddock levels stay yours. If one farm grows nearly all of your barley, letting its difficult country be recorded as 'barley converts worse' would misprice barley everywhere else you grow it — in the wrong direction. Where a crop is grown almost entirely on one farm the two genuinely cannot be told apart, and PaddIQ says so instead of picking.
  23. The farm level is always farm-AND-crop, never a bare farm offset: a place can grow poor barley on salty country and perfectly ordinary wheat on the rest of it.
  24. The held-out test also reports WHICH level earned the gain — farm-and-crop alone, paddock alone, or both. That is a different question from whether there was a gain, and it decides what to do next: a farm-and-crop term stands on a dozen or more paddock-seasons while a paddock term can stand on one. If the farm level is doing all the work, the paddock level is decoration until more seasons are scored.
  25. That learning is tested before it is used: each season is hidden in turn, the levels are learned from the other two, and they only get applied if they improved the year the model had never seen. If they did not, PaddIQ leaves them off and tells you. You can switch it either way yourself, and your choice sticks.
  26. Refit harvest index, farms held apart: this moves the crop constant using the crop level above, not raw average error. It refuses any crop whose record comes almost entirely from one farm. The farm and paddock levels are never folded in, so the refit and the learning switch can never double-count each other.
  27. It says HOW MUCH a paddock is behind and how consistently, never WHY. Salt, sodicity, compaction, a frost hollow and a paddock that is always sown last look identical from orbit — a paddock sitting well under year after year is worth walking over with a soil test.
  28. Flowering and maturity are worked out from THERMAL TIME — accumulated degree-days from your own seeding date against real daily temperature — not from a fixed calendar. A warm season flowers earlier and a late-sown crop gets a shorter grain fill, and both change which rain counts as finishing rain. Without a seeding record it falls back to typical WA dates and says so.
  29. BUT temperature is only part of flowering — vernalisation and photoperiod matter as much, so the estimate reads LATE for early sowings. Measured on Scepter at Kalannie it catches about half the real sowing-date effect. So use Flowering dates (on Satellite Yield, or Library in admin) to record when a variety actually flowered: a date you saw replaces the estimate outright.
  30. Record TWO sowing dates for a variety, not two records of the same one. The gap between them is what tells PaddIQ how much sowing date moves flowering on your place — one record fixes a level but cannot show a slope. Outside the range you have recorded, the figure is held at the nearest one rather than projected, and says so.
  31. Admin entries go in the master list every account reads; a grower recording their own variety overrides it. Adding the maturity date (when it went off green) is the only way grain fill ever gets measured — nothing has measured it yet.
  32. In admin these live under Model parameters, not the product library — flowering and grain fill decide which rain counts as post-anthesis and therefore the harvest index, so they are numbers the yield model runs on rather than a catalogue somebody maintains.
  33. Farm History: yields vs rainfall across every season you've imported.
Tips & tricks
  • The yield estimate's N factor reads your actual applied N from Field Records — fert decisions show up in the forecast.
  • 10yr History pulls a decade of station data for context in one click.

⚙️Settings & Data Health

Preferences, defaults, backup/restore, and the data-health toolkit that keeps a season's records clean.

How to use it
  1. Crop colours also lists any crop name in your records that PaddIQ doesn’t recognise — say John Deere’s “Wheat white”, or a new crop you’ve imported. An unrecognised name has no colour, no grain price and no removal rate until you say what it is: pick “the same as Wheat” (your records keep the name they came in with, and the choice applies everywhere a crop is read, every season) or “its own crop”, which adds it to the colour list. Choices are listed underneath and can be undone. White wheat is recognised as Wheat already — noodle and soft classes are grades, priced by grade.
  2. Repairs run automatically after every John Deere import and sync: records filed in the wrong season are moved to the season their date belongs to, part-paddock passes get their covered area, records get their machine, fuel and hours, harvests get each header’s yield, seeding passes get each variety’s rate, and the data-health repairs (units, paddock names and so on) run for every season. None changes a frozen cost. Settings → Data management shows when they last ran. Removing duplicates and leftover planned records still waits for you — a genuine second pass looks exactly like a duplicate, so deleting unattended could remove real work — but the sync counts them so you know when to look.
  3. Farm & Paddocks is the master registry; Assumptions & Calibration holds every number the nutrition models run on (water- and nitrogen-use efficiency, your SILO station) with where each came from; Default Values keeps the overhead $/ha that Financials and the exported plan use.
  4. Team & Permissions: add a person by email, pick a role, then click each tab to cycle Hidden → View → Edit. Adding someone is an invitation: they see “You’ve been invited to join <your farm>” at their next sign-in and choose Join, Not now or Decline — nothing about their account changes until they say yes, and a person with a farm of their own keeps it. Their email must be confirmed. It is live — they sign in with their own account, see your farm's data, and a change you make reaches them within a minute, or the moment they next open the app. Removing someone signs them out. Adding someone emails them an invitation (a set-your-password link); Invite on a row re-sends it. A member with Edit on a tab saves onto your farm — records are merged by id, so two people can work the same season; you see their work at your next sign-in or Sync.
  5. Backup & Restore: full JSON export/import of everything.
  6. Data Management also holds Reset account (empties every season in the cloud and this browser but keeps your sign-in) and Delete account (removes the sign-in, the cloud data, team memberships, saved layers, invoice originals, feedback, the assistant's question log and its meter — everything but the record that you accepted the terms; type DELETE to confirm; download a backup first).
  7. Data Health → Run all repairs: one button runs every safe, idempotent fix — duplicate imports, orphaned paddock keys, mis-labelled mL units, double-counted seeder costs, and stock lots received in tonnes (or litres for a kg product) that posted at face value — re-expressed in the product's base unit.
  8. The duplicate scanner lists suspected doubles for review before anything is deleted.
Tips & tricks
  • Run all repairs after big imports — it's safe to run any time; repairs only touch records that are provably wrong.
  • Hard refresh (Ctrl+Shift+R) after an update if a page looks stale — your browser may be running old code.
Stuck on something this page doesn't cover? Email admin@paddiqag.com — guide updated as the product changes.