It is easy to understate the economic value of a first Reference Farm (a reference validation farm) when it is treated as only one first customer. It is equally easy to overstate it by assuming that securing one famous farm will open the market. A reference farm can create revenue, validation, product improvement, sales trust, and operating standards at the same time, but none of these benefits is automatic. Its value depends on what is measured and which rights and deliverables are secured.

Therefore, the economic value of the first farm should be measured as the present value of the future costs that the farm reduces and the reusable assets it creates. This includes not only cash received, but also lower technical-failure risk, a shorter sales cycle for the next customer, standardized installation and operations, and a verifiable dataset. Discounts, free equipment, field response, custom development, data cleaning, and verification costs must also be deducted.

First, correct the misconception: the number of references is not the quality of the evidence

The fact that sensors worked at one farm can show product feasibility. It does not mean the same performance will appear with different barn structures, seasons, herd sizes, or ventilation methods. NASA’s technology-readiness concept distinguishes laboratory validation, demonstration in a relevant environment, and proof in an operational environment. One site’s success is important progress, but it does not validate every market condition.

The word “reference” must also be broken down. The place where equipment is installed, where data is received, where the customer paid, where results may be published, and where the customer will answer a third party’s questions may all be different. A free validation farm may be highly valuable for technical learning without proving price acceptance. Even a paying customer has limited external sales value if the company name, photos, or results cannot be published.

What increases the first farm’s value is not a logo but evidence that can be used. Before signing, agree on data ownership and access, anonymization, the scope of case-study publication, customer interviews and site visits, and the results-review process. Never disclose a farm owner’s personal information or sensitive operating data without permission, and do not turn a reduction figure into marketing copy before adequate measurement and review.

Calculate economic value in five ledgers

The first is direct contribution value. Subtract the direct costs of serving that farm—sensors, gateways, communications, travel, cloud, verification, and customer support—from contract revenue such as the initial installation fee, subscription, and maintenance. Treating the entire contract value as value hides the burden of discounts and field service.

The second is evidence value. Ask what verifiable answers the customer provided in a real environment. If uptime, calibration maintenance, measurement stability under ventilation conditions, baseline feasibility, and before-and-after feed intervention comparisons were documented, they can be reused for technical review and purchasing decisions. As with the GHG Protocol project-accounting principles, boundaries, baselines, monitoring, and reporting rules must be clear before evidence can support carbon claims.

The third is learning value. This includes shorter installation time, discovered wiring and power problems, standardized sensor locations, causes of missing data, and the screens and alerts farm staff actually use. If the next farm needs fewer engineer-days and return visits, convert that into cost savings. But one-off custom development may create complexity debt rather than learning.

The fourth is follow-on-order contribution. Estimate the differences in win probability, contract size, sales duration, and discount rate with and without the reference. number of follow-on opportunities × contribution profit per contract × improvement in win probability is a simple starting formula. Discount the result to when revenue occurs and separate deals that would have closed for other reasons.

The fifth is option value. This is the possible use for methodology development, research cooperation, entry into new species or regions, partner training, and testing a standard interface. Keep uncontracted options as probability-weighted scenarios, separate from committed revenue.

Calculate it with an assumption scenario

The following figures are hypothetical examples for understanding, not actual results of a particular company or farm. If first-farm revenue for one year is KRW 30 million and direct equipment, installation, travel, support, and cloud costs are KRW 22 million, direct contribution is KRW 8 million. If the standard installation procedure created there has a 70% chance of reducing installation cost at the next 10 sites by KRW 800,000 each, the simple expected learning value is 10 × KRW 800,000 × 70% = KRW 5.6 million.

Suppose the reference materials raise the win probability by 8 percentage points for five follow-on opportunities with expected contribution profit of KRW 15 million per contract. Expected follow-on value is 5 × KRW 15 million × 8% = KRW 6 million. However, this 8-percentage-point assumption must be tested against CRM conversion history, buyer interviews, and changes in sales stages. Inserting 30% or 50% without evidence only makes the calculation look precise.

Conversely, if obtaining the reference required a KRW 10 million discount, KRW 7 million for custom features unrelated to the product roadmap, and KRW 3 million for emergency field response, deduct all of them. The simple pre-discount total is direct contribution 8 + learning 5.6 + follow-on 6 − additional burden 20 = −KRW 400,000. The first year may be negative. However, reusable data and standards can later change the value by reducing real costs or creating orders.

The point is not to guess one perfect number. It is to identify the assumptions that move the result most. Results change greatly if the win-probability improvement is 3% or 8%, or if the installation standard applies to 3 sites or 30. Use optimistic, base, and conservative scenarios, and replace assumptions with actuals each quarter.

Operating standards that realize reference value

When selecting a farm, do not choose only the most cooperative site. Assess whether its species, scale, ventilation, connectivity, and operating conditions resemble the target customer; whether a sufficient measurement period is possible; and whether activity data and field events can be recorded. An unusually specialized farm may produce good validation results that are difficult to generalize.

Define success before the contract: installation complete, 90 days of valid data, an emissions estimate within a specified uncertainty range, or actual weekly use of reports by the operator are different outcomes. Include a carbon-reduction rate as a success criterion only when intervention, baseline, comparison group, and confounder controls are ready.

Design the deliverables as a bundle: site drawings and sensor placement, installation checklist, data dictionary, calibration and missing-data log, performance report, customer interview, anonymized case summary, and a standard BOM for the next deployment. Separate access to raw data from access to public materials. WIPO treats technology commercialization as a continuous process spanning protection, business planning, marketing, licensing, product development, and market diffusion; a patent or validation should not be equated with commercialization as a whole.

Finally, define expiry conditions. If equipment, software, or measurement methods change substantially, do not present old results as current performance. Show the measurement period, system version, scope, and limitations in each case, and periodically reconfirm the customer’s consent to publication.

Execution checklist

  • Has the target customer group and operating environment that the first farm should represent been defined?

  • Have free installation, paid contract, public case, and customer referral status been distinguished?

  • Have direct equipment, installation, support, and verification costs been deducted from direct revenue?

  • Have data access, case publication, site visits, and customer interview permissions been agreed?

  • Do the success criteria specify period, boundary, data quality, and evaluation method?

  • Have installation procedures, data dictionary, calibration records, and performance reports been left in reusable form?

  • Are follow-on win probability and sales-cycle improvements updated with actual CRM values?

  • Have one-off custom development and standard product learning been separated?

  • Are committed revenue and option value managed as separate scenarios?

  • Are system version, scope, publication approval, and expiry date managed for each case?

Conclusion: measure the first farm by how much better it makes the next farm

The economic value of a first Reference Farm cannot be answered by one contract amount. It creates value through direct contribution, lower technical and operating uncertainty, reusable evidence, improved follow-on conversion, and future options; it loses value through discounts, custom development, and support burden.

A good first farm does not merely produce a one-page success story. It records the conditions under which the system worked and failed, how much faster the next installation became, and which evidence buyers trusted. When that asset converts into actual follow-on contracts and cost savings, the reference farm has economic value beyond one customer.

Sources