Methodology & evidence

MachineWorth is only useful if you can trust it. So we grade every material claim, show our sources, and say plainly when we don't know. Here's how.

Evidence levels

Level	Means
verified primary	Independently verifiable primary evidence (e.g., a filing, a first-party demonstration we can inspect).
customer confirmed	Confirmed by the buyer/operator, not just the vendor.
manufacturer reported	Stated by the manufacturer; plausible but not independently verified.
demonstrated unverified	Shown in a demo or announcement without sustained, independent confirmation.
analyst estimate	Our estimate or a third-party estimate, clearly labeled as such.
speculative	Forward-looking targets or claims with little hard support.

The data model

What we collect on every machine.

Identity & specs

Manufacturer, model, category, height, weight, payload, speed, runtime, mobility, effectors — each with a spec-evidence level.

Commercial status

Announced → prototype → demonstration → pilot → commercially available → production → discontinued. Status is a claim, and it's graded.

Deployments

Customer, task, units, duration, supervision, and documented outcome — with a source and an evidence level. Empty means unverified, not zero.

Economics

Price (or range), integration, maintenance, energy, useful life, utilization. Most prices aren't public; estimates are labeled.

Evidence

Every material claim: source, source date, evidence level, verification status, and last-reviewed date.

Buyer intent

Non-sensitive signals only — industry, use case, budget range, timeline. Never financial-application data.

The Robot Economic Signal

The signal answers one question: after real costs, does this beat the labor it offsets — with room to spare? It compares the annual labor a robot displaces against its true operating cost, computes payback and five-year ROI, and then stress-tests utilization. It is intentionally hard to earn a "Potentially Attractive". It is an estimate, not advice.

Potentially Attractive

Fast payback, 5-yr ROI > 100%, still positive under a utilization stress test.

Needs More Analysis

Positive but sensitive — the answer depends on keeping the robot busy.

Currently Unattractive

Payback too long or savings not positive at these inputs.

FAQ

Is the automobile analogy real?

We treat 'the Kelley Blue Book / CARFAX for robots' as a hypothesis, not a fact. Robots lack standardization, iterate fast, depend on software, and have unclear resale and useful life. Those differences are the hard part. We only claim what the evidence supports.

How is the Economic Signal computed?

From your inputs: annual labor offset minus real robot operating cost, then payback and five-year ROI. 'Potentially Attractive' requires a fast payback and still-positive savings if utilization drops 20 points — a deliberate margin of safety.

Do financing relationships affect scores?

No. Evidence grades and Economic Signals are computed independently of any advertising, lead, or financing relationship. If that ever changed, we'd disclose it prominently.

What do you refuse to do?

Invent prices or deployments, present a demo as a deployment, or imply certainty we don't have. Absence of data is shown as absence, not filled with a guess.

Found something wrong or out of date? That's expected in a fast-moving field — tell us and we'll re-grade it. Every record carries a last-reviewed date for exactly this reason.