Methodology

How RoboMowerIndex handles mower data

The goal is simple: keep the compare surface useful without faking certainty. We would rather show a blank, a baseline label, or a follow-up note than silently guess a spec and make the page look complete when it is not.

Go back to compareMissing values stay visible by design

How owner ratings work

Current owners, past owners, and regular household users can rate a mower, including early failures and returns. Overall satisfaction is rated independently of the six core categories. Writing and ownership evidence are optional.

Each active, published consumer rating with a confirmed email counts equally. Overall scores are simple arithmetic means, rounded to one decimal. Category scores exclude skipped responses and show their own counts. Samples under five ratings are labeled “Early ratings”; an empty sample is “Not rated yet.” Updating a rating replaces its vote, and withdrawing it removes its vote.

Manufacturer, competitor, dealer, and compensated-review feedback is attributed separately and excluded from consumer averages. Research alone is not eligible. Owner ratings are separate from manufacturer specifications and external review counts.

“Email confirmed” means inbox access was confirmed—not ownership. “Ownership evidence submitted — not checked” means a private file was received. “Ownership evidence checked” means an administrator examined supporting evidence during an investigation. Evidence never increases rating weight and does not establish impartiality.

Optional evidence is a redacted receipt or a mower photo with a reviewer- and model-specific code that expires in seven days. Never include an address, government ID, full serial number, or manufacturer login. Evidence is private, can be replaced or removed, and raw files are scheduled for deletion after 90 days with retries on failure. Minimal status, decisions, and duplicate-detection information remain.

Private flags identify possible duplicate text, reused proof, or coordinated bursts. Flagged reviews stay public and counted until an administrator acts. Reports, star values, new accounts, shared networks, or a failure to answer us do not alone prove fraud. Praise and criticism follow the same standards. Affected reviewers receive a decision and can appeal through their secure rating-management link.

Email confirmation and these safeguards discourage casual spam; they cannot guarantee impartiality or prevent coordinated reviewers with separate accounts from temporarily affecting scores.

Core principles
  • Separate manufacturer claims from later real-world verification.
  • Keep missing data visible instead of guessing.
  • Track provenance so every important field can be checked later.
  • Prefer official support pages and manuals over marketing summaries when they conflict.
What blank fields mean

A blank field means we have not verified that spec yet, not that the mower lacks the capability.

Visible gaps prevent fake precision and make follow-up research easier to prioritize.

Source tiers

Not all sources carry the same weight, so each one is labeled accordingly.

Official sources

Manufacturer pages, manuals, and support documents are the primary baseline whenever they exist.

Secondary sources

Retailers, reviews, and trusted third-party listings help fill gaps, but they do not override better primary documentation.

Community baseline

Reddit and community roundups can seed coverage early, but they stay clearly labeled until a better source confirms them.

Review state labels

Not yet human reviewed

The source exists in the system, but nobody has confirmed it manually yet.

Human reviewed

A person has checked the live source and added reviewer metadata or notes.

Needs follow-up

The source was checked, but the claim still needs more verification or clarification.

Typical workflow

  1. 1

    Seed the model with a source-backed baseline row.

  2. 2

    Attach source records for the claims we currently know.

  3. 3

    Mark review state explicitly instead of pretending every field is verified.

  4. 4

    Leave blanks in place until a stronger source fills them.