Door-to-Door Soil Science

Nishi vs the field — every Nishi cell is measured against real organ source at emit time; each gap names the watch contract that will close it.

Nishi reusable house-by-house soil analysis with deficiency-gated recommendations vs university extension labs and mail-in consumer kits and soil-health labs and in-field sensor platforms -- Aug-2026 field

Layer 1 · Executive

Where we are. MEASURED 2026-08-19: no Nishi soil instrument exists. In tree and measured: the lab-science service facade (lab_dispatch -- grounded chemistry and physics verbs on the shared service base), the GTM channel facade, the product decomposer. The field is mirrored end to end: 15-dollar three-day extension Mehlich-3, mail-in kits, Haney and CASH soil-health panels, and farm-scale reusable spectroscopy handhelds.

Where we need to go. OPERATOR ASK (2026-08-19): soil-science analysis better than the ag colleges, reusable if possible, easy to go house by house, and feedback that allows the upsell. The thesis: do not out-lab the labs -- out-TRUST the incumbents. Reusable instrument at the door, calibration against extension-lab splits (the ag college becomes our referee, not our rival), plain-language honest reports, and a recommendation engine that mechanically refuses unearned upsells.

The unit. 1 u = one measured session-leg (organ plus gate plus live flip), the standing calibration.

Cost to the analysis and report engine: 4 u. SS1 panel definition, SS3 prescription tables, SS4 report emitter, SS5 upsell refusal gate -- all organ work on the labsci base.

Cost to a calibrated field instrument decision: 2 u more plus instrument. SS2 runs the probe buy-vs-build: evaluate a bought handheld against extension splits before any thought of building sensor hardware. 6 u cumulative.

Cost to a route-ready service pilot: 3 u more. SS6 household route plane with consent rows, SS7 pilot gate over a declared street of houses with paired lab splits. 9 u cumulative.

Research bar. MU extension Mehlich-3 package, the price-performance bar is measured on the 2026 extension lab benchmark for a complete fertility panel. Theirs: 15 dollars a sample, about three days [@mumehlich]. Ours: our per-house reading must land within the calibration bounds of that panel on paired splits, at the door in minutes -- and the service prices the FEEDBACK, not the chemistry.

Business case

Axis	Position	Figures	Decision
Investment	Organ legs ride the labsci base; the instrument decision is a BUY evaluation first -- farm handhelds exist today [handhelds] and extension splits cost 15 dollars each as ground truth [mumehlich]	estimated 9 u to a route-ready pilot; instrument capital gated on the SS2 accept rule	fund the engine legs now; buy one handheld for evaluation; never build sensor hardware before BUY fails a pre-declared bound
Market	Every house with a lawn is a candidate; the incumbent door-to-door model proves the demand shape at national scale [trugreen]; extension pricing anchors willingness to pay for chemistry [mumehlich] [unhprices] while kits anchor the consumer ceiling [mysoil]	the service prices the standing FEEDBACK relationship, not the assay	sell the season relationship: reading plus report plus only-what-you-need plan, re-measured each visit
Opportunity cost	The labsci facade gains a second grounded domain either way; SS4 and SS5 (honest report, refusal-gated selling) are reusable across every door-to-door service the estate ever runs	the trust machinery is the durable asset even if soil is not the wedge	charge SS4 and SS5 to the service-substrate program
Go to market	House-by-house with the instant reading IS the demo -- the incumbent leads with a sales analysis [trugreen], we lead with a measurement; soil-health literacy is rising on the Haney and CASH vocabulary [haneycmp] [cornellcash]	the honest report is the referral engine	start one street pilot (SS7), publish the calibration record, let the paired-split honesty be the pitch
Buy vs build	Chemistry: BUY (extension labs at 15 dollars are unbeatable reference chemistry [mumehlich]). Instrument: BUY first from the handheld class [handhelds], BUILD only on a failed pre-declared bound. Trust engine: BUILD -- the refusal gate does not exist to buy	the wedge is the trust engine, not the sensor	compose the labs, evaluate the sensors, build the honesty

6 of 14 capabilities measured|1 of them measured exceeds|8 open|coverage 428/1000|adoption 1 full / 5 partial

Layer 2 · Roadmap

Do this next — computed by the ranker, never chosen by a seat

Order from nx_compare_rank (nx_dr_ocm: (deficit + cost-of-delay + option + enables) x sponsor x self-sufficiency x momentum / cost). FINISH rows are rungs whose symbol is present but whose organ is short of full adoption: the cheapest closures on this board, listed before any new work. Stamp: # asof=1787883550 domain=soilscience target_version=0.1 rungs=7 done=0 open=7 finish=0 ranker=nx_dr_ocm

#	Stage	Rung	Priority	Derivation
#1	0.1	Homeowner report emitter (SS4) ss_report_emit	9000	v=18 m=1 c=2
#2	0.1	Panel definition and reference protocol (SS1) ss_panel	5500	v=11 m=1 c=2
#3	0.1	Prescription engine (SS3) ss_prescribe	5000	v=10 m=1 c=2
#4	0.1	Upsell refusal gate (SS5) ss_upsell_gate	1000	v=2 m=1 c=2
#5	later	Probe buy-vs-build verdict (SS2) ss_probe_hw	750	v=3 m=1 c=4
#6	later	Household route plane (SS6) ss_household_plane	100	v=1 m=1 c=10
#7	later	Street pilot gate (SS7) ss_pilot_gate	0	v=0 m=1 c=20

Critical path — contract, done-rule, executor, cost

Rung	Closes with	Definition of done (pre-declared)	Executor	Est.
Panel definition and reference protocol (SS1)	ss_panel	panel rows as DATA with the Mehlich-3 extension panel as reference method [@mumehlich]; gate refuses a panel row without a reference mapping	Organ	1 u
Probe buy-vs-build verdict (SS2) after SS1	ss_probe_hw	evaluate one bought reusable handheld from the field class [@handhelds] [@stenon] on paired extension splits; pre-declared accept rule: correlation bounds per analyte published BEFORE the trial; BUILD is considered only if BUY fails the bounds or the unit economics	Organ plus instrument	2 u
Prescription engine (SS3) after SS1	ss_prescribe	agronomy tables as DATA (turf class, region, target); neg-control: an in-spec soil MUST produce an empty amendment plan	Organ	1 u
Homeowner report emitter (SS4) after SS3	ss_report_emit	plain-language per-house report with confidence bounds per number and measurement class per figure; emitted by the organ, never typed	Organ	1 u
Upsell refusal gate (SS5) after SS3	ss_upsell_gate	every recommended line item carries its measured deficiency; the gate neg-control (deficiency-free lawn) MUST produce an empty upsell; a filled one is the gate FIRING [@trugreen]	Organ	1 u
Household route plane (SS6)	ss_household_plane	household rows with consent, visit history and paired-split schedule in the sovereign store; re-put per visit as heartbeat	Organ	1 u
Street pilot gate (SS7) after SS2,SS4,SS5,SS6	ss_pilot_gate	a declared set of houses, every reading banked, every Nth paired with an extension split [@mumehlich], calibration drift published; REFUSES on partial coverage	Organ	2 u

Milestones

Milestone	Rungs	Cumulative
M0 · Analysis and report engine proven	SS1,SS3,SS4,SS5	4 u
M1 · Instrument decision made on evidence	SS2	6 u
M2 · Route-ready pilot proven	SS7	9 u

Layer 3 · Engineering

How this is scored. Every Nishi mark is measured: the generator reads the real organ source on disk and requires the implementing symbol to exist (no self-grading). A watching tag names the organ and symbol contracted to close a gap — the mark flips itself on the next compare beat when that workstream ships, and the comparewatch- plane row flips with it. The flip is necessary, not sufficient: it proves the symbol exists, never that the capability is good. The bar is the rung's pre-declared done-rule, proven by its gate — a symbol shipped without the behaviour behind it is a defect, and the flip is exactly what makes that defect visible instead of quiet. Competitor marks record documented capability presence — presence, not depth or scale. Adoption is measured too: every measured row carries where its organ stands on the estate's ladder (source → built → promoted → registered → invoked; libraries by importer reach minus validation importers; gates by the execution surfaces that run them). A row is fully adopted only at the top of its ladder; anything short is tagged partial with the exact remedy, so a build nobody promoted can no longer read as shipped. Census stamps: importers asof 1787849099, gate census asof 1787855507 (unix seconds; -1 = census absent).

Capability matrix — measured against source

◉ leads / measured exceed● present◐ partial○ absent · click any capability for its evidence

Capability	Nishi	Extension lab	Mail-in kit	Soil-health lab	Field sensor
PANEL
Routine fertility panel, pH, P, K, Ca, Mg and micros Open — watching `runtime/nx_soilsense.nx` : `ss_panel`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. MU extension runs the Mehlich-3 complete panel at 15 dollars in about three days [mumehlich]; the mail-in class reads pH plus 13 nutrients [mysoil]. Contract = our panel definition as DATA with the extension panel as the reference method watching ss_panel	○	◉	●	●	●
SPEED
Reading at the door in minutes Open — watching `runtime/nx_soilsense.nx` : `ss_field_reading`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. Stenon FarmLab reads the core panel in seconds by combined NIR, UV-Vis and impedance spectroscopy [stenon]; labs take days and mail kits about a week [mysoil]. Contract = the door-side reading path, whatever sensor mix wins the buy-vs-build row watching ss_field_reading	○	○	○	○	◉
REUSE
Reusable instrument, no consumable per sample Open — watching `runtime/nx_soilsense.nx` : `ss_probe_hw`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. The handheld sensor class is reusable by construction [handhelds]; every lab and kit burns a consumable per sample. The operator ask is exactly this: reusable, house by house watching ss_probe_hw	○	○	○	○	◉
TRUTH
Calibration against reference-lab ground truth Open — watching `runtime/nx_soilsense.nx` : `ss_calib_model`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. Sensors trade accuracy for speed and carry cloud-model calibration debt [handhelds]; the labs ARE the reference. Contract = paired-sample calibration: every Nth house also ships a split to the extension lab [mumehlich] and the model recalibrates on returned truth -- the grade-1 referee built into the route watching ss_calib_model	○	◉	◐	◉	◐
BIO
Soil-health biology, respiration and organic carbon Open — watching `runtime/nx_soilsense.nx` : `ss_bio_assay`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. The Haney panel adds microbial respiration and water-extractable organic C and N [haneycmp]; Cornell CASH scores twelve attributes into a rated assessment [cornellcash]. Contract = the biology assay we can run reusably, or an honest permanent ABSENT on this row watching ss_bio_assay	○	◐	○	◉	○
RX
Prescription engine, amendment plan from readings Open — watching `runtime/nx_soilsense.nx` : `ss_prescribe`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. Extension recommendations are the trusted format; kits emit generic plans. Contract = agronomy tables as DATA (turf class, region, target) driving the per-house plan -- thresholds in data rows, never hardcoded watching ss_prescribe	○	●	●	●	◐
REPORT
Plain-language homeowner report with honest bounds Open — watching `runtime/nx_soilsense.nx` : `ss_report_emit`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. Contract = the emitted per-house report: what was measured, the confidence bounds per number, and what NOT to buy -- every figure carries its measurement class watching ss_report_emit	○	◐	●	◐	◐
SELL
Upsell gated on measured deficiency Open — watching `runtime/nx_soilsense.nx` : `ss_upsell_gate`, re-measured on every compare beat. Ship that symbol and this mark flips itself; the `comparewatch-` plane row flips with it. The incumbent residential model runs a lawn analysis at signup and carries a documented aggressive-upsell reputation [trugreen]. Our inversion IS the product: the recommendation engine refuses any line item without a measured deficiency behind it, proven by a gate whose neg-control deficiency-free lawn MUST produce an empty upsell watching ss_upsell_gate	○	○	○	○	○
SERVICE
Lab-science service facade, the substrate this rides Measured exceed: `lab_dispatch` in `runtime/nx_labsci_svc.nx`, verified at emit. The estate lab-science facade already dispatches grounded chemistry and physics verbs as JSON endpoints on the shared service base -- soil chemistry lands as new verbs on a proven substrate, not a new service Adoption: `REGISTERED-DARK` — PARTIAL: callable, authorised, no MCP invocation on record (a direct fork logs the runner, so this is not proof it never ran); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked). adoption REGISTERED-DARK	◉	○	○	○	○
BUSINESS
Go-to-market channel engine, facade live in an adjacent domain Measured: `gtm_channel_best` exists in `runtime/nx_gtm_svc.nx`, verified at emit. The GTM facade ranks channels per product today in the supplement domain -- soil-service channel rows are a data admission, not new code Adoption: `REGISTERED-DARK` — PARTIAL: callable, authorised, no MCP invocation on record (a direct fork logs the runner, so this is not proof it never ran); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked). adoption REGISTERED-DARK	●	○	○	○	○
Outcome-to-capability decomposition Measured: `pd_parse` exists in `runtime/_hdl_build/nx_product_decompose.nx`, verified at emit. The product decomposer maps the service outcome onto required capabilities and unproven-opportunity gaps over the design-to-market planes Adoption: `REGISTERED-DARK` — PARTIAL: callable, authorised, no MCP invocation on record (a direct fork logs the runner, so this is not proof it never ran); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked). adoption REGISTERED-DARK	●	○	○	○	○
SERVICE
Grounded chemistry verbs on the same base (water activity, stability) Measured: `lab_water_activity` exists in `runtime/nx_labsci_svc.nx`, verified at emit. The facade already computes water activity and shelf stability from grounded models -- soil moisture and amendment chemistry land beside verbs of the same shape Adoption: `REGISTERED-DARK` — PARTIAL: callable, authorised, no MCP invocation on record (a direct fork logs the runner, so this is not proof it never ran); no execution surface runs it either (clock, cron, daemon, roster, actlog and surfaced forks checked). adoption REGISTERED-DARK	●	○	○	○	○
BUSINESS
Landed-cost engine for the instrument import decision Measured: `lc_calc` exists in `runtime/_hdl_build/nx_landed_cost.nx`, verified at emit. The field-sensor class is import-priced hardware; the SS2 buy decision runs its landed cost through the sovereign corridor engine before purchase Adoption: `LIVE` — fully adopted (top of its ladder).	●	○	○	○	○
EVIDENCE
Sovereign mirrored and pinned reference register for this field Measured: `knowledge/fetched/cmp_soilscience` exists in `knowledge/compare/soilscience.refs`, verified at emit. Every rival claim on this page resolves to a mirror in the estate evidence store, content-pinned by hash and re-measured on every emit Adoption: `ABSENT` — PARTIAL: no source and no binary: this is a build contract, not a partial deploy. adoption ABSENT	●	○	○	○	○

Risk register

Risk	Likelihood x impact	Mitigation
The handheld reads badly on residential soils and the pilot publishes wrong numbers	probable x high	SS2 accept rule pre-declared; every report carries bounds; paired splits every Nth house keep the referee in the loop; a failed bound stops the lane at M1, not after a season of bad calls
A watch contract flips on symbol existence, so a stub reads green	possible x high	the flip is the receipt, the rung gate is the proof
The honest-refusal model sells less per visit than the incumbent upsell machine	probable x medium	that is the bet, stated plainly: retention and referral against per-visit extraction [@trugreen]; the pilot gate banks per-house revenue so the answer is measured, not argued

On these two registers. Rows are declared in the domain's plan file and carry the debt id, which is the join key back to the sovereign debt plane — that plane, not this page, is the authority on state. Reconciling them automatically (the regen reading the plane and refreshing these rows) is a named, owed rung; until it lands, treat an id here as a pointer to look up, not a status to trust.

Honest verdict. The bar to beat is not the ag college -- it is the ag college PRICE: Missouri extension now runs a Mehlich-3 complete panel (P, K, Ca, Mg, S and micros) at 15 dollars a sample with results back in about three days [@mumehlich], and the extension pricing ladder is published [@unhprices]. Consumer mail-in kits read pH plus 13 plant-available nutrients about a week out [@mysoil]. Soil-HEALTH labs add the biology: microbial respiration and water-extractable organic carbon on the Haney panel [@haneycmp], and the Cornell CASH framework scores twelve biological, physical and chemical attributes [@cornellcash]. The reusable in-field class already exists at farm scale -- Stenon FarmLab reads N, P, K, Mg, SOC, pH, texture and moisture in seconds by combined optical and impedance spectroscopy [@stenon] -- but the handheld class is subscription-priced for agriculture and carries cloud-calibration debt [@handhelds]. And the incumbent door-to-door residential model is a sales machine: a lawn analysis at signup and a documented aggressive-upsell reputation [@trugreen]. The Nishi wedge inverts it: a REUSABLE instrument at the door, minutes-per-house readings calibrated against extension-lab splits, a plain-language honest report -- and an upsell engine that mechanically REFUSES to recommend anything the measured soil does not need. Trust is the moat; the refusal gate is the proof.

Person · product · place — not yet measured for this domain

Every compare carries this layer. Declare knowledge/compare/soilscience.ppp (rows surface|nishi or c1..c4|label|url|connect naming OUR live surface and each rival's front door), run nx_ppp_probe domain soilscience, and this section fills itself on the next beat: the same ruler on both sides — privacy and CX (third-party hosts, tracker classes, cookies, security headers), design and longevity (design hygiene, computed WCAG contrast, render-blocking resources, unsized media, script weight, theme and motion queries), findability (landmarks, skip link, on-site search, breadcrumb, headings, internal links).

References

Beyond a link list. Every reference below resolves twice — the publisher's copy and, where banked, the estate's own non-rottable library mirror with a content pin — and carries its evidence class plus the exact claim on this page it grounds. Keyed marks like [key] in the matrix notes jump here. A dash means honestly absent, never assumed.

[mumehlich] University of Missouri Extension. MU Soil and Plant Testing Lab moves to cheaper, quicker testing (Feb 2026) -- the lab transitions to the Mehlich-3 multi-nutrient extractant: the complete soil test package priced at 15 dollars a sample with results typically back in three days. publisher · read in our library knowledge/fetched/cmp_soilscience_mumehlich.html · pin h6b23bce48a2b441b44b519f73c398b589126fe70ee460dce613cb4242bc4195a · accessed 2026-08-19 · vendor-docGrounds: The routine fertility panel row and the calibration row: the 2026 price-performance bar and the reference method our paired-sample calibration ships splits to.
[unhprices] University of New Hampshire Extension. Pricing list for all soil test options -- the published extension pricing ladder for homeowner and grower soil analyses. publisher · read in our library knowledge/fetched/cmp_soilscience_unhprices.html · pin h41273c32cbcd523c147bd8334a2b6a24277cc4f53568939966064c63de1b2b24 · accessed 2026-08-19 · vendor-docGrounds: The verdict pricing context and the market biz row: the extension ladder anchors homeowner willingness to pay.
[mysoil] MySoil. Soil test kit product page -- mail-in kit reading pH plus 13 plant-available nutrients with online results about a week out. publisher · read in our library knowledge/fetched/cmp_soilscience_mysoil.html · pin h864500323a73dd62edd4597f2b237877ebdb9901b186f3b1f31c979f53543696 · accessed 2026-08-19 · vendor-docGrounds: The mail-in kit column: the consumer format on panel breadth and turnaround.
[cornellcash] Cornell Soil Health Laboratory. Individual soil analyses (CASH framework) -- the comprehensive assessment measures and rates biological, physical and chemical indicators including soil respiration into a scored assessment. publisher · read in our library knowledge/fetched/cmp_soilscience_cornellcash.html · pin h13ff6c2d478ef52151b13d43d8a61a2454e19da9dd01a9b7058e95567ca1b758 · accessed 2026-08-19 · vendor-docGrounds: The soil-health biology row: the CASH indicator set is the academic bar for the bio assay decision.
[haneycmp] Bayer ForGround. Comparison of soil health tests -- the Haney test integrates chemical and biological measures: soil respiration, water-extractable organic carbon and nitrogen, alongside conventional nutrient measures. publisher · read in our library knowledge/fetched/cmp_soilscience_haneycmp.html · pin hda2a7127283550b663267a2a07bf68e8778301282223bfc4ca24e8a9391bb44f · accessed 2026-08-19 · vendor-docGrounds: The soil-health biology row: what the Haney panel measures and why respiration is the biology headline.
[stenon] Stenon. Real-time soil analysis technology -- the FarmLab handheld reads N, P, K, Mg, SOC, pH, texture and moisture in seconds combining optical spectroscopy (NIR, UV-Vis) and electrical impedance spectroscopy, thousands of data points per reading. publisher · read in our library knowledge/fetched/cmp_soilscience_stenon.html · pin h907d62b25511b68c3414c147b9c44ecf436c42bb21db25e4ff22b424eb9e4439 · accessed 2026-08-19 · vendor-docGrounds: The reading-at-the-door row and the probe buy-vs-build row: proof the reusable minutes-per-house format exists at farm scale.
[handhelds] AgTecher. Best handheld soil testing devices -- seven in-field solutions compared (Stenon FarmLab, AgroCares scanner, ChrysaLabs probe and peers): reusable instruments, subscription pricing aimed at agriculture, cloud-model calibration and connectivity constraints. publisher · read in our library knowledge/fetched/cmp_soilscience_handhelds.html · pin h580b15cbbbd2ceb61862a28d32c1dd47c635cdcf97e2d2aa3e29d052aefaddfe · accessed 2026-08-19 · mirrored-indexGrounds: The reusable-instrument and calibration rows plus the probe buy-vs-build biz row: the field sensor class, its pricing shape and its calibration debt.
[trugreen] This Old House. TruGreen lawn care review (2026) -- the incumbent residential model: a lawn analysis at service start, plan-based treatments, and a documented reputation for aggressive upselling of additional services. publisher · read in our library knowledge/fetched/cmp_soilscience_trugreen.html · pin h16418a9f37d180d9bee1d76ab2417ba9fc4acd10d28b70b3f3e51ffeb82547a8 · accessed 2026-08-19 · mirrored-indexGrounds: The upsell-gate exceed row and the go-to-market biz row: the door-to-door demand proof and the trust gap our refusal gate inverts.

generated by nx_swcompare_matrix (sovereign NishiLang organ) from knowledge/compare/soilscience.matrix · every Nishi cell verified against organ source at emit time · watch cells re-measured on every compare beat · zero JS, zero trackers