Halter sells execution, not analytics — 73 evidence rows show adoption momentum and one verified switch, but almost no churn signal
Asked:
“Research Halter as a named competitor and adjacent pastoral-management platform relevant to AIMER” — covering customers and decision context, jobs, current hires and alternatives, push/pull/habit/anxiety forces, switching events, commercial mechanics, and implications for AIMER, in one SQL-style evidence table distinguishing DIRECT, PATTERN, and STRUCTURAL evidence, labelling vendor claims, keeping negative evidence, and inventing nothing.
73 evidence rows assembled from public sources — independent reporting, peer-reviewed studies, company and investor material — across six research sections: 62 DIRECT and 11 STRUCTURAL rows, 53 high- and 20 medium-confidence, spanning New Zealand, Australia, the United States and Canada. No PATTERN-tier rows were established.
Where the evidence lands
Every one of the 73 rows drawn as a mark on its dimension lane, grouped by research section. Green is DIRECT evidence, amber is STRUCTURAL (product fact or analyst classification, not an observed customer event); hollow marks are medium confidence. Hover any mark for the claim, source and limitation.
DIRECT — observed event or first-hand claimSTRUCTURAL — product/market fact, no customer eventhollow = medium confidence▲ vendor-supplied claim
Coverage by section, tier and confidence
Evidence rows and distinct publishers per section, from the precomputed coverage aggregate — so the reader can judge where the base is deep and where it is thin.
Halter’s pull is execution: virtual boundaries, remote stock movement and grazing control replace physical fencing, motorbike mustering and labour — overlapping AIMER only at the grazing-planning interface, while a pasture-measurement layer over Halter execution is a STRUCTURAL complement with no direct integration evidence on record. halterhq.com
Adoption evidence concentrates where fencing, labour or terrain are costly: a 1,475-cow Canterbury dairy farm reporting 6% more pasture eaten (vendor-distributed ROI study), a 225,000-acre Colorado ranch on the satellite system, and public-land routes via a $2.7m federal land-manager partnership. deseret.com
The one independent friction benchmark is rugged Eagle County: six ranchers, 1,500–3,000 collared cattle, where older virtual-fencing systems had collars falling off, UI issues, spotty coverage and time lags — a warning for that class of terrain, not a generalisation to every Halter deployment. deseret.com
Priced disclosure exists but is partial: a US subscription of $72 per cow per year with free collars and $4,500 wifi towers is on record, yet contract length, churn, total cost of ownership and API/data-portability terms remain undisclosed — and the table holds exactly one named switching event against zero verified rejections, cancellations or reductions. deseret.com
The seven questions, answered by the rows
1 · Who buys, approves, uses, blocksNamed actors are owner-operators and ranch managers (Maronan Farm, High Lonesome Ranch, Alan Williams, James Rodgers & Anna Shaw), regulators as approvers/blockers (SA government green light; Western Australia where only Gallagher’s eShepherd was legal in 2022), and institutions as payers. No rows evidence lenders, processors, cooperatives or channel partners in the buying decision — a gap, not an absence in reality.
2 · Jobs and observable proofThe rows substantiate functional jobs clearly — move and manage cattle by app instead of physical methods, virtually fence without barriers, allocate pasture — and one emotional job (“a great deal of assurance” from remote monitoring). Four clean distinct jobs per category are not supported: the shortfall is recorded as negative evidence rather than filled, and emotional/social signals sit distributed across adoption-force and customer rows.
3 · Current hires and alternatives18 alternative rows span advisers, set stocking, farm software (MINDA LIVE), grazing plans, institutional support, boluses and cow wearables ($40–$197/cow/year), paper records, plate-metering and visual pasture estimation, conventional fencing, spreadsheets, and rival virtual fencing (Gallagher eShepherd, Nofence). Most better/worse/cheaper comparisons are “not established” in the sources and are left blank.
4 · Forces on adoptionThe force ledger is asymmetric: 10 high-confidence DIRECT anxiety rows (welfare and regulatory scrutiny, collar retention, UI, coverage, time lags, training, financing, vendor dependence) against 4 high-confidence pull rows, 4 medium habit rows and a single medium push row. Anxiety is the best-documented force in the public record.
5 · Switching eventsExactly one named adopted event is recorded, with its outcome unreported. No verified rejection, cancellation, reduction, renewal or switch-away event exists in the assembled evidence — the largest negative-evidence gap in the table.
6 · Commercial mechanics21 rows cover funding ($220m Series E at a $2b valuation, ~$400m total), pricing, financing, implementation, partnerships, IP (first VF technology marketed for intensive pastoral dairy), expansion across NZ, Australia, the US and Canada, Beef Pro launch, and performance claims — several vendor-supplied and flagged as such.
7 · Implications for AIMER12 implication rows separate DIRECT AIMER facts from STRUCTURAL analyst classifications: position AIMER as the measurement-and-recommendation layer, prioritise documented data portability and APIs, offer a transparent lower-capital adoption path, and treat implementation proof — not claims — as the defensible layer.
The evidence table
All 73 rows, all 21 requested columns. Search any text; filter by section, tier or confidence. Blank cells are genuinely blank in the sources and are left unresolved. Rows tinted amber carry vendor-supplied claims per their limitation note.
Method: 73 evidence rows, one per materially distinct public claim or event about Halter, with 21 columns as requested; coverage totals from a precomputed aggregate of 40 section × dimension × tier × confidence cells. Tiers: DIRECT (observed event or first-hand claim) and STRUCTURAL (product or market fact without a customer event); no PATTERN rows were established. Confidence is high or medium. Source dates run to August 2026 where given; many event dates are undisclosed and left blank. Nothing was inferred to fill blanks; confidential pricing terms, contract length, churn, full total cost, unit economics and API detail remain undisclosed in the sources.