Revenue Architecture for Aquaculture Software in 2027 (Per-Kg Economics, Feed Channel, AI Fish Vision)

Two ways to build the go-to-market: direct-first versus channel-first

Every vendor selling into salmon, shrimp, or tilapia production eventually confronts the same fork, and the choice determines the shape of the entire revenue organization for the next three to five years. The first option is a direct-first architecture: you hire inside AEs for the small-operator segment, field AEs plus solutions consultants for mid-market, and a named-account enterprise team that works the twenty or so producers who control the majority of global farmed salmon volume. Marketing generates demand through industry press, trade shows, and technical content aimed at heads of farming and veterinarians. You own the customer relationship end to end, you keep the full contract value, and your product roadmap answers to producers rather than to a partner's bundling strategy.

The second option is a channel-first architecture built on the feed manufacturers and equipment vendors who already sit inside every farm's operating rhythm. Cargill's EWOS business, Skretting under Nutreco, and BioMar collectively supply a very large share of the world's aquafeed, and their technical sales teams visit sites on a cadence no software vendor can match. On the equipment side, AKVA group, ScaleAQ, and Innovasea sell the pens, barges, feeding systems, cameras, and sensors that your software has to read from anyway. A channel-first motion means you build integrations first, sign reseller or referral agreements, comp a channel team on partner-sourced pipeline, and accept a lower net contract value in exchange for a warm path into accounts that would otherwise take eighteen months of cold outreach.

Neither is purely better, and the honest observation from adjacent verticals — precision agriculture, dairy herd management, poultry house automation — is that the strongest revenue architectures run both, weighted by segment. Direct dominates SMB, where the operator is the buyer and the deal closes in two to six months. Channel dominates mid-market and enterprise, where roughly 30% of pipeline is influenced by a feed or equipment relationship and where the technical credibility of a nutritionist who has walked the site carries more weight than any demo. The failure mode is picking one, staffing for it exclusively, and then discovering at $15M ARR that half your addressable pipeline is unreachable.

There is a third hybrid worth naming because it is increasingly common in 2027: hardware-bundled software, where the equipment vendor sells your platform as a line item on the barge or camera purchase order. This converts a software buying decision into a capital-expenditure decision, which for a producer running a multi-year site expansion is often an easier internal approval. The trade-off is that your renewal is now coupled to a hardware refresh cycle you do not control, and your pricing power is capped by the partner's margin expectations.

How to decide between direct-first and channel-first

The decision is not a matter of preference; it falls out of four measurable inputs. First, segment concentration. If more than 60% of your target ARR sits in accounts with more than twenty sites, you are effectively selling to a list of named producers — Mowi, Cermaq, SalMar, Lerøy, Bakkafrost, Grieg Seafood, AquaChile, Multi X, plus the shrimp integrators like CP Foods and Thai Union — and channel warmth matters enormously because those accounts have 8–14 named stakeholders and a 5–12 month cycle. If your ARR is spread across hundreds of small operators, channel economics do not work; the partner's cost to serve exceeds their margin.

Second, integration depth. If your product reads live sensor data from feed barges, underwater cameras, or oxygen probes, you already have an engineering dependency on AKVA, ScaleAQ, or Innovasea hardware. That dependency is leverage in a partnership negotiation and a reason to formalize the relationship rather than maintain brittle unofficial integrations.

Third, whether you can prove per-kg economics unassisted. A direct motion requires your own team to walk a CFO through feed-conversion-ratio improvement and mortality reduction. If you do not have a nutritionist or production specialist on staff, the channel partner's technical team becomes your credibility proxy — which works, but it means the partner owns the value narrative.

Fourth, capital runway. A direct enterprise team costs roughly $1.5M–$2.5M fully loaded before it produces a closed deal, given enterprise AE OTE in the $400k–$580k range plus solutions consultants at $185k–$255k. Channel-first defers that cost but caps your growth rate to the partner's attention.

The diagram compresses a decision most teams make implicitly and badly. Note that every branch terminates at the same instruction: instrument per-kg economics regardless of motion. That is not a rhetorical flourish. A channel partner will not carry your value story indefinitely, and a direct team without production data loses to competitors who bring a number.

The concrete numbers behind each segment and each motion

Start with segment ACV bands, because everything downstream — quota, headcount, coverage — derives from them. Small operators running one to three sites under roughly 500 tons of standing biomass land in a $22,000–$98,000 band. The module mix is deliberately thin: site management, feed logging, mortality records, basic environmental monitoring. Win rates run 22–28% because the competitive field is fragmented and the buyer is a single owner-operator who decides fast. Cycles run 60–180 days.

Mid-market operators — four to twenty sites, roughly 501 to 15,000 tons — land at $140,000–$840,000. The module mix expands to feed optimization, biomass tracking, sea lice detection, and integration with barge feeding systems. Stakeholder count jumps to five or six: CEO, COO, head of farming, veterinarian, IT. Win rates compress to 18–25% and cycles stretch to 90–210 days. Coverage should run 4.4x.

Enterprise producers and vertical integrators — twenty-one sites and up, into the hundreds of thousands of tons — land anywhere from $840,000 to $14M. This is multi-country, multi-species, custom ML, and certification automation against ASC, BAP, and GlobalG.A.P. standards. Win rates fall to 13–19%, cycles run 150–360 days, coverage should be 5.2x, and the buying committee reaches 8–14 named people including procurement and a head of sustainability who did not exist as a title in this industry a decade ago.

Pricing in 2027 typically resolves to a per-site annual fee with a biomass tier. Small operators pay roughly $8,400–$28,000 per site per year. Mid-market pays $22,000–$84,000 per site with tier overage. Enterprise pays a discounted $14,000–$48,000 per site at volume but layers biomass overage on top, which is where the real expansion lives. Modules price separately: AI fish vision at roughly $24,000–$98,000 per site per year, feed optimization at $24,000–$140,000, sea lice and health monitoring at $22,000–$84,000, and sustainability certification automation at $24,000–$98,000. Implementation runs $8,000 to $140,000 depending on integration surface.

Compensation follows the segment split. SMB AEs carry $880k–$1.4M in new ARR against a $145k–$195k OTE at a 50/50 split. Mid-market AEs carry $2.4M–$3.4M against $235k–$320k at 45/55. Enterprise AEs carry $4.4M–$6.8M against $400k–$580k at 45/55, with multi-year deals vesting roughly 55/30/15 across years one through three and a $90k–$140k draw to survive the ramp. A feed-and-equipment channel manager runs $240k–$340k at 55/45. Solutions consultants and per-kg economics specialists — often literal aquaculture nutritionists — run $185k–$255k at 70/30. The two overlay roles that matter most in 2027 are a sustainability and welfare certification specialist at $185k–$245k and an AI fish vision specialist at $215k–$295k. CSMs run $115k–$155k at 70/30 against $340k–$480k in expansion quota plus retention gates.

Retention is where the two motions diverge most sharply. Small-operator NRR sits at 102–110%, mid-market at 108–118%, enterprise at 115–130%. The gap is entirely explained by expansion vectors: enterprise accounts grow site count, cross biomass tiers, and attach AI modules, while a three-site operator has nowhere to grow. Channel-sourced accounts tend to renew well but expand slowly, because the partner controls the account conversation. Direct-sourced enterprise accounts expand fastest and are the reason forecast methodology should shift to roughly 70% expansion and 30% new logo once you are past two hundred enterprise customers.

Value realization anchors all of it. Producers do not measure software on uptime or user adoption; they measure it on per-kg production cost. Strong feed optimization is worth an FCR improvement in the 0.05–0.12 range, which at typical feed cost translates to something on the order of $200–$500 per ton of biomass. Strong health monitoring and sea lice detection can move mortality by three to eight percentage points across a production cycle. Vendors who instrument and attribute those numbers win enterprise deals at roughly twice the rate of vendors who lead with feature lists. That ratio is the single most important number in this entire architecture.

Adjacent motions worth borrowing from: RAS, shrimp, and neighboring protein verticals

Land-based recirculating aquaculture systems deserve separate treatment because their revenue architecture behaves less like farming software and more like industrial process control. Facilities operated by Atlantic Sapphire, Nordic Aquafarms, Salmon Evolution, and Proximar Seafood run closed-loop water treatment where oxygen, ammonia, nitrite, and CO2 management are continuous control problems rather than periodic observations. Software intensity per ton is far higher than at a sea site, the buyer skews toward a plant engineer rather than a head of farming, and the sales cycle often attaches to a facility construction timeline — meaning your deal is gated by a capital project schedule you can neither accelerate nor forecast normally. The practical implication for revenue architecture: RAS deals belong in a separate forecast category with construction-milestone-based commit dates, not in the standard enterprise weighted pipeline, or they will wreck your quarterly accuracy.

Shrimp is a different animal again. Pond-based production in Southeast Asia and Latin America runs shorter cycles, far more units, and thinner per-unit margins, which pushes toward lower ACV and higher volume — closer to an SMB motion executed at scale. Vendors like Eruvaka and XpertSea have built around water-quality sensing and biomass estimation for exactly this shape. If you are architecting revenue across both salmon and shrimp, resist the temptation to run one comp plan. The cycle lengths and deal sizes are incompatible, and blending them produces AEs who chase the salmon deals and neglect the volume business that actually scales.

The most useful borrowed patterns come from precision agriculture and dairy herd management. Both went through the same evolution: sell hardware, attach software, then discover that the software renewal depends on proving a production-economics delta the buyer can see in their own books. Dairy vendors learned to instrument milk yield per cow and somatic cell count; row-crop vendors learned to instrument yield per acre against input cost. Aquaculture's equivalents are FCR, mortality, and cycle time. The transferable lesson is that the vendors who built the measurement layer into the product — rather than producing it as a quarterly business review artifact — retained better, because the customer saw the number continuously rather than being told it once a quarter by a CSM.

Poultry house automation offers a warning rather than a pattern. That market consolidated hard around integrated producers who eventually built or bought their platforms outright. Several of the largest aquaculture producers already run substantial internal platform teams, and the strategic risk for any vendor whose enterprise ARR concentrates in three or four accounts is that one of them decides to internalize. The defensive move is the same one that works everywhere: own a data asset or a model that gets better with cross-customer scale — sea lice detection accuracy, for instance, improves with training data breadth that no single producer can match internally.

Implementing the architecture: sequencing, instrumentation, and the 2027 AI overlay

Sequencing matters more than any individual decision, because building the pieces in the wrong order produces a team that cannot sell what it has. The order that works is roughly this.

Phase one, before $5M ARR: prove the economics on a handful of sites. Do not hire a channel team. Do not hire enterprise AEs. Take five to ten mid-market accounts, instrument FCR and mortality against a pre-deployment baseline, and run the measurement for at least one full production cycle — which for Atlantic salmon means the better part of a year in seawater, so start immediately. What you are building is not a case study; it is the instrumentation layer that every later motion depends on.

Phase two, $5M–$15M ARR: split the comp plans. The most common structural error at this stage is running SMB and enterprise AEs on one plan. A 60-day cycle and a 360-day cycle cannot share a quota period, a ramp curve, or an accelerator schedule. Split them, give the enterprise team a draw, and move the enterprise forecast to quarterly commit with monthly named-account review. Add solutions consultants here, not earlier — they are wasted on deals that close in ninety days.

Phase three, $15M+ ARR: stand up the channel. This is where the feed-manufacturer and equipment-vendor relationships get formalized with a dedicated channel manager and real comp on partner-sourced pipeline. The attribution question — does the channel manager get credit, the AE, or both — should be settled in writing before the first partner deal closes, because retrofitting it destroys trust in the plan.

Phase four, running concurrently through 2027: the AI overlay. Computer vision applied to individual fish — biomass estimation, sea lice counting, behavioral anomaly detection, welfare scoring — is the largest single expansion lever available, and vendors report it commanding meaningful incremental ARPU on top of the base platform, in the range of a quarter to nearly half again on attached accounts. It requires a specialist overlay role because the sale is genuinely technical: you are asking a veterinarian to trust a model's lice count over a manual sample, and that conversation cannot be delegated to a generalist AE.

Instrumentation requirements for RevOps are specific and non-negotiable. You need per-customer FCR delta against baseline, mortality rate by cycle, production cycle time, module attach rate, channel-influenced pipeline percentage, and certification pipeline against ASC, BAP, and GlobalG.A.P. deadlines. Those six metrics drive the weekly pipeline council, the monthly expansion review, and the quarterly board conversation. Expansion comp should trigger on site count growth plus biomass tier upgrade at ninety days live, AI module activation at ninety days live with an accelerator, and a documented per-kg economics milestone at the six-month mark.

The failure modes map cleanly onto skipped phases. Skip phase one and you sell features into a market that buys production economics, losing at roughly twice the rate. Skip the comp split in phase two and your enterprise AEs quit in month eight when the SMB reps hit accelerators on volume. Skip the channel in phase three and a third of your mid-market and enterprise pipeline stays invisible. Skip the AI overlay and you leave the largest 2027 expansion lever on the table while a competitor attaches it to your installed base during their own renewal conversation.

Related questions

Should aquaculture software price per site or per ton of biomass?

Per site as the base with a biomass tier overlay. Pure per-ton pricing punishes customers for growth and creates renewal friction during strong production years; pure per-site pricing under-monetizes large sites. The hybrid captures expansion without penalizing the outcome your product is supposed to improve.

How long should the enterprise ramp be for an aquaculture AE?

Nine to twelve months, with a draw covering the first six. Enterprise cycles run 150–360 days, so a rep hired in January realistically closes in Q4. A six-month ramp on a 360-day cycle guarantees attrition regardless of rep quality.

Does the feed channel cannibalize direct enterprise deals?

Rarely, if territory rules are explicit. Feed manufacturers influence pipeline rather than resell, so the usual arrangement is partner-sourced credit to the channel manager and full quota retirement to the AE. Conflict arises only when the same account is worked cold and warm simultaneously.

What retention metric matters more, logo or biomass?

Biomass retention, at roughly a 92% target against 95% logo retention. A customer who keeps the contract but decommissions half their sites has churned economically even though the logo survives. Track both; forecast on biomass.

How should RAS facility deals be forecast?

Separately, on construction milestones rather than weighted stage probability. Land-based facility software attaches to a capital project timeline, so standard stage-based forecasting produces systematic slip. Give RAS its own commit category and report it separately to the board.

FAQ

What NRR should an aquaculture software vendor target by segment?

Roughly 102–110% for small single-site and multi-site operators, 108–118% for mid-market, and 115–130% for enterprise. The spread reflects expansion headroom rather than product quality — a three-site operator has almost nowhere to expand, while a hundred-site producer can add sites, cross biomass tiers, and attach three or four modules over a two-year window.

Why is per-kg economics instrumentation described as the single largest structural lever?

Because it is the metric the buyer already uses. Producers manage the business on feed conversion ratio, mortality rate, and production cycle time, and every capital or software decision is evaluated against cost per kilogram of harvested biomass. A vendor who can attribute an FCR improvement in the 0.05–0.12 range, or a mortality reduction of several percentage points, is speaking the buyer's language. A vendor showing dashboards is not, and loses at roughly twice the rate at the enterprise tier.

How much pipeline actually comes through feed and equipment partners?

Around 30% of mid-market and enterprise pipeline is influenced by a feed manufacturer or equipment vendor relationship. Feed suppliers — Cargill's EWOS, Skretting, BioMar — have standing technical relationships with nearly every commercial producer. Equipment vendors like AKVA group, ScaleAQ, and Innovasea control the sensors and feeding systems your software depends on. Neither channel is optional above $15M ARR.

What does the AI fish vision overlay actually sell?

Individual-fish computer vision covering biomass estimation, sea lice counting, welfare and behavior scoring, and early disease indicators. It replaces manual sampling workflows that are labor-intensive, stressful for the fish, and statistically noisy. It carries meaningful incremental ARPU on attached accounts and needs a dedicated specialist because the buyer is often a veterinarian evaluating model accuracy against their own sampling protocol.

Why do land-based RAS facilities need a different revenue motion?

Because the software intensity per ton is dramatically higher and the buying process attaches to a construction schedule. Closed-loop water treatment requires continuous control of oxygen, ammonia, and CO2 rather than periodic monitoring, so the product surface is larger. But the deal closes when the facility commissions, which no amount of sales pressure moves. Forecast these on construction milestones.

At what ARR should a dedicated channel team be hired?

Around $15M ARR, and not before. Below that, the direct team should still be proving per-kg economics on reference accounts — the credibility asset that makes a channel partnership worth signing in the first place. Hiring channel managers before you can prove production impact produces partnerships that generate meetings and no revenue.

Sources

• https://www.fao.org/fishery/en/publications/sofia — FAO, The State of World Fisheries and Aquaculture
• https://www.mowi.com/investors/ — Mowi ASA investor reporting
• https://www.salmar.no/en/investor/ — SalMar ASA investor relations
• https://www.akvagroup.com/investor — AKVA group ASA investor information
• https://asc-aqua.org/ — Aquaculture Stewardship Council certification standards
• https://www.bapcertification.org/ — Best Aquaculture Practices certification program
• https://www.globalgap.org/ — GLOBALG.A.P. aquaculture standard
• https://www.nutreco.com/ — Nutreco (Skretting) corporate reporting
• https://www.biomar.com/ — BioMar Group
• https://www.undercurrentnews.com/ — Undercurrent News seafood industry coverage

Related on PULSE

• [Revenue Architecture for Addiction Treatment + Behavioral Health Residential Software in 2027 (42 CFR Part 2 + ASAM-CONTINUUM CDS + MAT Attach Economics)](/knowledge/ra0174)
• [Revenue Architecture for Fertility + IVF Clinic Software in 2027 (Employer-Benefits Channel, AI Embryo-Grading, RFID Witness Per-Cycle Overage)](/knowledge/ra0171)
• [Revenue Architecture for AP Automation + AR Automation + Spend Management Software in 2027 (Virtual Card Take-Rate Flywheel, AI Native Spend Disruption, ERP Partner Channel Dominance)](/knowledge/ra0170)
• [Revenue Architecture for Senior Living Operations Software in 2027 (Dual Buyer, REIT Channel, AI)](/knowledge/ra0137)
• [Revenue Architecture for AI Note-Takers + Conversation Intelligence in 2027 (Auto-Actions, Channel)](/knowledge/ra0128)
• [Revenue Architecture for AI Voice Platforms in 2027 (CSAT Parity, Agent Replacement, CCaaS Channel)](/knowledge/ra0127)