How We Score

Every product on BaniCar gets five scores, each from 1 to 5 in half-point steps. The overall score is the plain average of the five, rounded to one decimal. There is no hidden weighting. If you'd weigh the criteria differently, every line is visible, so you can.

One rule follows from this: a product needs an overall score of 4.0 or higher to appear in a category's Top Picks. Below 4.0 it still gets a page and a fair write-up. It just doesn't get recommended.

Where the numbers come from

Three places, in this order of trust:

  1. The manufacturer's product page or PDF manual. Resolution, peak amps, PSI, lumens, operating temperature, warranty term. This is the primary source for every spec we score.
  2. Standards bodies and regulators. FMVSS 108 for lighting, UL 2743 for jump-starter safety, NHTSA recall records. We use these to check whether a manufacturer's claim means anything.
  3. The Amazon listing. Used only to confirm the ASIN, the variant, and what's in the box. We don't use Amazon ratings, review counts, or prices, and we don't display them.

If a spec appears only in marketing copy ("ultra-bright," "military grade") and never as a number, it earns nothing.

The five criteria

1. Spec strength

What it measures: the core published numbers for the category, compared against other products in the same category and price band. For dash cams that's resolution, sensor, field of view, parking mode, and GPS. For jump starters it's peak amps, watt-hours, rated engine size, and protections. Each category page lists its core specs.

  • 1 — Bottom of the category on the numbers that matter, or key specs not published at all.
  • 3 — Meets the category baseline. Does the job; nothing on the sheet beats its peers.
  • 5 — Leads the category on two or more core specs, with every number published and specific.

2. Build quality

What it measures: what the manufacturer discloses about materials and durability. Metal vs. plastic housing, IP rating, operating temperature range, supercapacitor vs. battery, listed certifications (UL, ETL, CE, FCC). We score disclosure and the numbers, not how it feels in the hand, because we haven't held it.

  • 1 — No materials, no ratings, no temperature range published.
  • 3 — Basic materials stated; at least one rating or range published.
  • 5 — Full disclosure: materials, IP rating, temperature range, and a third-party safety listing we can look up.

3. Fit and compatibility

What it measures: how many vehicles and situations the product covers, and what extra parts you need to get there. Universal vs. model-specific, power source, size limits, parts sold separately.

  • 1 — Narrow fit, and the listing doesn't tell you until the box is open.
  • 3 — Works for most common vehicles; limits are stated clearly.
  • 5 — Broad fit, a precise size guide or vehicle lookup, and everything needed in the box.

4. Warranty and support

What it measures: warranty term in months, what it covers, whether the brand has a reachable US contact (phone, email, registered address), and whether firmware and spare parts are still offered.

  • 1 — No warranty stated, or a marketplace-only seller with no reachable support.
  • 3 — 12 months, written terms, email support.
  • 5 — 24 months or more, US-based support with a phone line, and a published return process.

5. Value

What it measures: what you get per dollar of MSRP compared with the two or three closest rivals in the category. MSRP comes from the manufacturer's site with the date we checked it. We don't use the Amazon price because it changes daily and we can't display it anyway.

  • 1 — Charges a premium for specs that cheaper rivals match or beat.
  • 3 — Priced in line with what it offers.
  • 5 — Undercuts rivals on price while matching or beating them on the core specs.

Reading the "why"

Each score has a one-line reason, and every reason names a number. "4.5 — 4K front + 1080p rear; most rivals under $200 are 1440p/1080p" tells you exactly what we compared. If a "why" ever reads like an adjective instead of a comparison, email us and we'll rewrite it.

What we can't measure

We don't own these products. Some things sit outside the method, and we'd rather say so than fake it:

  • Durability over time. A 24-month warranty hints at it. It doesn't prove it.
  • Noise. Vacuums, inflators, and LED cooling fans all make noise. Decibel figures are rarely published and rarely measured the same way.
  • Real-world suction, inflation time, cold-start performance. Published numbers are lab numbers. Your trunk carpet, your truck tire, and your January morning are not the lab.
  • Night image quality, lens flare, app stability. All real. None on the spec sheet.
  • Whether the listing you land on matches the spec we scored. Amazon listings change. Check the model number on the listing against the product page before you buy.

For those things, read a hands-on review from a site that owns the product, or read owner reviews on the listing. We're the first filter, not the last word.

How often we re-check

  • Every product page shows an Updated date. That's the last time a human re-checked the spec sheet and MSRP against the manufacturer's page.
  • MSRP and warranty terms: every 90 days. Full specs: every 180 days, or sooner when a manufacturer releases a revision.
  • A reader report of an error triggers a check within 3 business days (About BaniCar).
  • When a manufacturer changes a spec, we update the score and leave a note on the page. We don't silently overwrite.

Things that never move a score

Sponsorships (we don't take them). Affiliate commission rates (identical across products in a category). Brand reputation. Amazon ratings. How good the listing photos look.

Questions about the method: hello@banicar.com, or see Frequently Asked Questions.

Sources

  • 49 CFR § 571.108, Standard No. 108 (FMVSS 108), eCFR — https://www.ecfr.gov/current/title-49/subtitle-B/chapter-V/part-571/subpart-B/section-571.108 — accessed 2026-10-16
  • UL Solutions, "Portable Power Pack Testing" (UL 2743) — https://www.ul.com/services/portable-power-pack-testing — accessed 2026-10-16