Validation and pilot plan — first 24 months (East Tennessee two-site)¶
Purpose¶
Define what must be proven cheaply before major infrastructure or enterprise commitments, for a family-run, two-site operation (~5 ac hub, ~120 ac production, ~35 min separation). Minimize spend; use pilots to cut labor, logistics, and automation uncertainty—not to justify shopping.
Aligns with: 10-year roadmap (Phases 0–2 overlap the first 24 months), Validation backlog and decision gates (V1–V12, G1–G3), Validation — before major spend, Remediation backlog (water, processing, labor surge).
Fillable checklists: Pilot and recon checklists — first 24 months.
Execution dossier (Phase 0–1 — use first): Execution dossier hub · First 90 days · First 12 months · First 24 months · Master checklist · Decision memo.
Local evidence (named counties) — V1 soil/slope: Campbell vault WSS AOI is steep/gravelly dominant (Demory site intelligence); clip to ~120 ac operating polygon and walk—AOI total 558.6 ac in capture ≠ planning acreage. Claxton / Anderson: no WSS AOI in corpus—run WSS on homestead footprint. V7 lease: 2024 NASS Anderson $28.50 vs Campbell $27.50 / ac non-irrigated cropland estimates—starting comps only (Claxton, comparison).
What changed: Validation tasks gain county anchors; pass/fail still parcel-level.
Connectivity validation (Starlink / WAN — first 24 months)
Canonical strategy & diagrams: Connectivity strategy — Claxton home base and Demory farm site · Execution topology package — two-site smart farm (Mermaid) · Reference architecture — 5 ac + 120 ac · Network segmentation, site-to-site connectivity, and internet exposure · Remote access and operational security model.
Sources (hardware facts, not field coverage): Electrical, networking, and Starlink — inbox batch (2026-04-23); supplementary Mermaid: Two-site smart farm — network topology and WAN/edge reference.
| Execution deliverable | What “pass” looks like | Typical timing (vs T0) | Gates |
|---|---|---|---|
| Site survey / obstruction / placement | Sky view + mount candidates logged (photo/map) for each terminal or CPE site; re-check after leaf-on if trees matter | Start T0+14–30d; revisit ≤ 12 mo | Feeds V10; blocks “ we mounted it wrong ” surprises |
| Power and grounding | Dedicated circuit / UPS plan for WAN CPE and edge router; surge/ground at demarc reviewed (checklist or electrician ) before WAN is treated as production | Before telemetry / remote admin is operationally relied upon | Ties manual fallback power domain; see Manual fallback and degraded modes, Runbook — power loss at remote site |
| WAN reliability testing | Seasonal notebook or sheet: outages, rain fade, handoff LTE↔LEO (if dual path ) at SITE_HOME and (if deployed) SITE_FARM pins in pilot checklists |
Ongoing from first production use; minimum one 90-day summary in year 1 | V10 pass requires honest log, not screenshots alone |
| Remote access hardening before operational use | Admin path documented (VPN / outbound tunnel / scoped rendezvous ) per Remote access and operational security model; no new production reliance on dashboards until inventory complete |
Before expanding pilot past one cluster | Works with MV- on* Automation stop rules |
Rules — what may depend on Starlink (or any primary LEO/satellite WAN)
| Allowed WAN-dependent (non-safety) | Notes |
|---|---|
| Dashboards, convenience alerts, cloud MQTT egress, remote peek for trip batching | Degraded SOP required when WAN fades (Automation degraded modes SOP) |
| Optional cloud history | Queue / buffer at edge where possible |
Rules — what must remain functional without WAN
| Must work offline / without Internet | Notes |
|---|---|
| Physical welfare checks (water, escape ) | G8 manual baseline — MQTT silence ≠ “ fine ” |
| Authoritative books export discipline | Cloud optional; offline copy policy per SoR |
| Safe defaults on actuators / interlocks | No cloud required for safe fail state |
| Local sensor → gateway path | May buffer uploads; field mesh can outlive WAN |
Cost tracking and stop rules
| Practice | Stop / freeze trigger |
|---|---|
| Separate ledger lines: ISP, LEO subscription, cell (pilot), spares | Quarterly review: if recurring WAN + cell stack exceeds documented value proxy (trip $/labor saved or stated ops benefit ) for two consecutive reviews → pause new WAN-dependent scope until fixed, downsized, or pivot (see CS- on Automation stop rules )* |
| No second farm terminal / redundant WAN before Phase 1 proof | Per Connectivity strategy, Capital plan |
| Gate | Starlink / LEO WAN angle |
|---|---|
| V10 (cell/backhaul at critical points) | If Starlink or LTE is the planned uplink, drive-test and log outage windows (weather, obstructions ) — not just bars. Pass = documented “works/doesn’t” per pin + degraded procedure when WAN fades + rows above in motion. |
| G8 / manual baseline | Satellite WAN must not shortcut physical water checks—MQTT silence during fade is normal risk. |
Must not silently become: a passed V10 based only on vendor maps without seasonal logging at SITE_HOME and (if used) SITE_FARM ; or production telemetry trust before remote access hardening checklist is done.
What changed because of Starlink analysis: V10 evidence can explicitly cite LEO uplink where chosen; connectivity validation is now a first-class execution track (survey, power, WAN log, security, cost stops ) — not only a matrix row.
Off-grid SITE_FARM (Demory) — extra gates (mesh-first, solar + battery)
When the production parcel is modeled off-grid-first, connectivity validation also covers energy and local-only paths—not only WAN behavior:
| Gate / deliverable | What “pass” looks like | Ties to |
|---|---|---|
| Dcrit / Pcrit narrative | Written split of must-stay-up loads (gateway head, pump controls, critical sensing) vs sheddable (extra CPE, non-critical RF) | Off-grid power strategy — Demory |
| DR-1 / CS-5 | Before second RF class, HaLow segment, or farm Starlink: networking load bracketed vs battery autonomy—or explicit waiver | Off-grid operational decision rules, Automation stop rules |
| MV-8 | Local-only drill: mesh/LPWAN + gateway without relying on WAN; SOC-stressed behavior documented | Off-grid degraded modes — Demory, Automation stop rules |
| Pilot scope | One gateway + one RF class until G8 / V10 honest—DR-5 | Same pages + Capital plan off-grid sequencing |
Must not: treat Starlink at the farm as proof of safe or authoritative ops when local power or gateway is the actual bottleneck.
Calendar windows (T0 = execution start date)¶
T0 is the day the family declares execution start (fill in your tracker). All windows below are relative to T0. Gates do not slip when the calendar does—evidence must pass before spend.
| Window | Roadmap phase (approx.) | Focus |
|---|---|---|
| T0 → T0+90d | Phase 0 + start Phase 1 | Books, V8 path, V1/V3 started, V11 log, G8 manual baseline started; no major CAPEX |
| T0+3mo → T0+12mo | Phase 1 | V2/V4/V5/V6 evidence, fence repair tiers, one telemetry pilot cluster max |
| T0+12mo → T0+18mo | Phase 1 | G1 evidence packet complete or explicit waivers; pivot to lease if grazing path fails |
| T0+18mo → T0+24mo | Phase 1 → possible early Phase 2 | G1 PASS → permission for fence/water/herd MVP per validation backlog—not automatic by month 24 |
Quarter overlay (optional—align reviews to real calendar)
| Quarter after T0 | Review emphasis |
|---|---|
| Q1 (days 1–90) | Exit checklist in First 90 days |
| Q2–Q4 (months 4–12) | First 12 months |
| Year 2 (months 13–24) | First 24 months + G1 decision |
1. Validation matrix¶
Legend — cost band: $0 = fuel/time only; $ = incidental fees; $$ = professional/study fees (set your own thresholds in private budget).
| Assumption | Why it matters | How to test it | Cost band | Time | Success criteria | Unlocks (decision) | Task ID |
|---|---|---|---|---|---|---|---|
| Intended use of 120 ac matches deed/lease | Wrong paper kills CAPEX | County zoning / use call; attorney if unclear; written summary | $–$$ | 2–8 wk | No fatal conflict or written path to compliance | Structures, on-farm sales options | V8 |
| Year-round physical access | Winter mud and ice change economics | Drive all lanes in bad weather; photo log; neighbor / road owner conversation | $0–$ | 1–2 mo | Honest go / conditional / no with mitigations listed | Batching rules; equipment siting | — |
| Water rights + quantity | Stocking and well spend depend on it | Well report or pump test; attorney on surface / riparian if needed; seasonal observation | $$ | 2–6 mo | Documented flow band + drought stress narrative | Major water CAPEX; herd ceiling | V2, V4 |
| Soil / slope / wet limitations | Rotation and equipment paths | NRCS Web Soil Survey + walk + photos; optional forage consult sample | $0–$$ | 4–12 wk | Map + “ no-go ” areas named | Fence routing; pasture plan | V1 |
| Fence condition reality | Escape and predator risk | Tagged walk with photos + rough length by segment | $0 | 2–4 wk | Repair tier list (now / soon / with herd ) | Fence CAPEX order | V3 |
| Conservative carrying capacity | Overstocking = welfare + forage failure | Extension worksheet or short consult; document assumptions | $–$$ | 4–8 wk | Upper bound you will not exceed day one | Stocking plan; lease vs owned split | V4 |
| Market price reality | Revenue model sanity | Auction sale bills 12 wk; direct probe only after unit economics skeleton | $0–$ | 8–12 wk | Price range per unit written | Marketing choices; breakeven math | V5 |
| Processing / haul path | Shrink and slot risk | Calls to processors + distance / window notes; haul quote if needed | $0–$ | 4–8 wk | All-in per head placeholder filled (range) | Owned herd scale (with G1 ) | V6 |
| Lease rate reality | Lease vs graze tradeoff | Comps (legal sources ) + conversations | $0–$ | 4–12 wk | Comparable $/ac band | Lease pilot terms | V7 |
| Insurance adequacy | One loss can wipe years | Farm + umbrella quotes; binder review checklist | $–$$ | 2–6 wk | Named perils understood; gaps closed or accepted | Phase 2 equipment / scale with eyes open | V9 |
| Cell / backhaul at critical points | Telemetry and safety comms | Drive test + map screenshots at tanks / gates ; plus Connectivity validation (obstruction, power, WAN log, remote access ) | $0 | 2–4 wk + ongoing | Honest “ works / doesn’t ” per pin with seasonal WAN notes | Radio plan (LoRa vs LTE vs mesh ) ; WAN cost discipline | V10 |
| Family labor + surge | Grid lies without stress week | 8+ week time log + one calving / hay / escape window | $0 | 8–16 wk | Surge hours ≤ policy cap or hire / neighbor plan exists | Automation scope; enterprise breadth | V11 |
| Debt / FSA (if used ) | Covenant risk | Lender / FSA conversation; write limits | $0 | 4–8 wk | Max payment and trigger rules in writing | Financed fence / water / equipment | V12 |
| Telemetry ≠ welfare substitute | Automation fantasy | 30d manual baseline for water check before remote actuation | $0 | 6–8 wk | Paper or simple log matches reality | Pilot sensors only after G8 -style proof | G8 |
2. Pilot program plan (by stream)¶
2.1 Land intelligence and mapping¶
| Element | Minimal pilot | Duration | Cost band | Success signal | Links |
|---|---|---|---|---|---|
| Operating map authority | One file or layer named canonical (even sketch ) | Mo 0–3 | $0 | Everyone uses same IDs | Asset registry |
| Soil + wet ground | NRCS + winter walk | Mo 0–9 | $0–$$ | No surprise mud traps on lanes | V1 |
| Paddock / fence segments | Numbered photos + length estimates | Mo 3–12 | $0 | Repair list prioritized | V3 |
Checklist: Pilot and recon checklists.
2.2 Infrastructure reconnaissance¶
| Element | Minimal pilot | Duration | Cost band | Success signal |
|---|---|---|---|---|
| Water points | Flow bucket / meter read + seasonal notes | Mo 0–12 | $0–$ | Sustainable band for planned stock class |
| Power at pump / gate | Label panels + generator plan on paper | Mo 3–6 | $0 | Safe manual off known to two people |
| Road / drainage | Rain event drive (safe ) | Mo 6–18 | $0 | Access policy updated honestly |
Checklist: Pilot and recon checklists.
2.3 First business experiments¶
| Experiment | Keep it small | Cost band | Success / kill |
|---|---|---|---|
| Market observation | Auction visits + notes only | $0–$ | Price band written or stop chasing wrong species |
| Direct / relationship probe | One buyer conversation / small trial lot if ready | $–$$ | Unit margin positive or pivot |
| Lease pilot (if path ) | Short term / small acres | $ | Cash flow and labor fit or exit |
Rule: No species or scale commitment until V5 + V6 have first pass evidence.
2.4 First telemetry experiments¶
| Experiment | Scope | Cost band | Kill criteria |
|---|---|---|---|
| One green path home ↔ field | Single uplink + one sensor class (e.g. tank ) | $–$$ | No stable comms → fix comms before more sensors |
| Manual before dashboard | Stick / schedule log | $0 | Triage hours > trip savings → stop adding alerts |
| Connectivity validation track | Connectivity validation (obstruction, power, WAN log, remote access, cost CS- )* | $0–$$ | CS-1 / CS-2 → freeze WAN-dependent scope per Automation stop rules |
Links: Instrumentation priority matrix, Instrumentation ROI model, Telemetry system of record.
2.5 First labor / workflow tests¶
| Test | Method | Duration | Success |
|---|---|---|---|
| Weekly rhythm | Written DOW + 8 week adherence | 8+ wk | Realistic or revised |
| Surge window | Include one stress period in V11 | 1–2 wk | Coverage matrix updated with hire / neighbor row if gap |
| Two-site batching | Log trips / month + purpose | 4–8 wk | Batch rules that family can keep |
Links: Labor model, Family labor coverage matrix, Two-site operations model.
3. Decision-gate framework (24-month context)¶
3.1 Phase 1 → Phase 2 (roadmap: “first real operating business”)¶
Intended meaning: Permission to commit to fence/water MVP, owned or contract livestock at planned ceiling, and operating cadence — not maximum scale.
Triggers (all required unless explicitly waived in writing)
| # | Criterion | Evidence |
|---|---|---|
| 1 | G1 satisfied per validation backlog (V1–V4 minimum + legal use) |
Dated pass + links |
| 2 | V6 processing / haul path credible for chosen animal | Call log + range all-in cost |
| 3 | V11 labor truth includes surge or mitigation plan | Log + matrix |
| 4 | Working capital policy exists (even if placeholder ) | Capital plan subsection |
Waive / pivot: If grazing path fails but lease works, Phase 2 can be “ lease-led ” — same gate discipline, different CAPEX list.
3.2 Phase 2 → Phase 3 (within or just after 24 months — only if exceptionally fast)¶
Triggers
| # | Criterion | Evidence |
|---|---|---|
| 1 | G2 (validation page ) — positive operating contribution year OR explicit subsidy budget with end date | P&L category match to accounting baseline |
| 2 | Trailing margin not negative two years running (or documented why ) | Remediation backlog §13 scale freeze |
| 3 | Second stream only if labor test green | V11 + coverage |
Most operations will still be in Phase 2 at month 24 — that is normal.
3.3 What should block scaling (anytime)¶
| Blocker | Until |
|---|---|
| Negative trailing contribution margin | Scale freeze + cut variable costs or herd |
| Water / rights not documented | No additional stock |
| Processing path unknown | No owned herd expansion |
| Insurance gaps unaccepted | No new financed equipment (see V9 ) |
| Telemetry triage > benefit | No fleet rollout |
| Labor surge uncovered | No second enterprise |
4. Cheap proofs before expensive commitments¶
| Cheap proof | Expensive commitment it defers |
|---|---|
| Zoning letter / use clarity | Permanent structures, on-farm retail build-out |
| Well report + seasonal observation | Drilling / major pond / long pipeline |
| Fence walk + segment repair list | Full perimeter rewire at wrong priority |
| Sale bills + one processor path | Breeding herd / scale purchase |
| Manual water check log 30d | Remote actuators on water |
| One sensor pilot + trip log | Multi-zone camera + cloud stack |
| Custom hire quote vs own (hours ) | Loader / hay gear finance |
| 8-week labor log + surge | Second job on farm / agritourism |