Every machine in the WhichBot catalog carries a number out of ten, and that number decides its place in the rankings and the Buy Now page. Numbers you can’t audit are just vibes with fonts — so here is the entire methodology, published: what we weigh, why demos get punished, and exactly what would move any score tomorrow.
The five pillars (and their weights)
1. Does it actually do the job? (35%) — The core task, performed in real homes, judged against the category’s honest state of the art. A vacuum is scored on navigating YOUR furniture chaos, not a showroom grid; a mower on real lawns with real dogs; a humanoid on what owners and credible third parties report it doing unassisted, not what the keynote implied.
2. Can you actually get one? (20%) — Availability is a feature. A shipping machine with proven units in customer homes carries evidence a waitlist can’t; that’s why the Buy Now page exists and why waitlist machines carry a structural score ceiling until real deliveries happen. Vaporware doesn’t get to outrank hardware here, ever.
3. The price-to-capability honesty (20%) — Not “is it cheap” but “does the price buy proportional capability.” This is how a $799 machine can outscore a $1,599 one, how the score-per-dollar picks get computed, and why flagship pricing gets flagship scrutiny.
4. Living with it (15%) — Maintenance ritual, app quality, noise, dock footprint, update cadence, the company’s record on supporting old units. The spec sheet is a promise; year two is the truth.
5. Trajectory (10%) — Firmware momentum, the maker’s shipping history, ecosystem health. A machine improving monthly earns points a stagnant one loses — this is the pillar where the waitlist tracker’s honest-status notes come from.
The anti-hype rules
The pillars have teeth because of three standing rules. Demo footage counts for nothing — edited video is marketing, and marketing is not evidence; only repeatable, third-party-witnessed or owner-reported performance moves pillar one. Waitlist machines cap until units land — we track them enthusiastically at the tracker and score them honestly, which means an 8.5 ceiling no reservation-stage robot has broken yet. No pay-for-placement, structurally — rankings sort by score, the score formula is this page, and no affiliate relationship touches a pillar. If we ever earn a commission on a link, it’s disclosed on the page and invisible to the math.
What moves a score (watch it happen)
Scores are versioned, not carved. Real triggers from the catalog’s short life already: firmware that fixed a mowing bot’s edge behavior (+0.3), a price cut that transformed a vacuum’s pillar-three math (+0.4), shipping delays that dropped a humanoid’s availability pillar, and owner-reported reliability patterns that outweighed our own unit’s luck. When robot pages change scores, the verdict line changes with them — the catalog is a living document because the machines are living products.
Why publish the formula at all
Because the robot industry’s default is the opposite: scores from nowhere, rankings that follow affiliate payouts, and “best of” lists written by people who never held the remote. WhichBot’s bet is that an auditable number beats an impressive one — that buyers deserve to know WHY the category leader leads, and that a methodology page like this one is worth more than a hundred five-star badges. Read the scores, then check them against the formula. That’s the entire product.
Start anywhere: the catalog, the rankings, what actually ships, or the quiz if you want the machine matched to you.