How We Pick
The whole method, including the parts that make us look worse than the competition. If you only read one page here before trusting a recommendation, read this one.
We have not tested any product on this site. What we do instead: read every manufacturer's published specification, compile them into one comparable table, work out what those numbers mean on a cold morning, name every source, and check prices live. Every roundup names at least one product to skip.
What we do not do
This comes first because it is the part that matters. Eight of the nine competing pages we read use language that implies hands-on testing without ever quite claiming it - "chosen for durability, ease of use and cold-weather performance", "we thoroughly researched various products", an unattributed first-person "I focused on performance" sitting over an anonymous byline. We are not going to do that, so here is the list in plain terms:
- We do not test products. No lab, no cold chamber, no fleet, no borrowed units. Nobody here has held most of these items.
- We do not accept free products. Nothing arrives from a manufacturer and nothing is loaned for review.
- We do not sell placement. No brand has ever paid to appear here and none ever will. There are no sponsored posts.
- We do not invent credentials. Scooter M. is an enthusiast, not a mechanic, not an engineer and not certified in anything relevant. There is no expert panel and no editorial board.
- We do not print numbers we cannot source. If a manufacturer does not publish a figure, our table says so rather than filling the cell with an estimate.
The four things we actually do
- Compile the published specifications into one table. Blade width and extended length for scrapers. Fastening method, wiper coverage and dimensions for covers. Rated melt temperature for washer fluid. Peak amps and engine ratings for jump starters. Cold cranking amps and group size for batteries. Length and load rating for traction mats. Manufacturers publish all of this and nobody puts it side by side, which is the single most useful thing an outsider can do in this category.
- Do the arithmetic nobody bothers with. A traction mat is only useful if it gives more than one wheel revolution of grip, so we work out what that means in inches. Cold cranking amps is defined at 0F, so we say what that definition actually implies rather than quoting the number. Lithium jump packs lose deliverable current as they get cold, so we say where to store one. None of that needs a laboratory. All of it is checkable.
- Name every source. Every roundup, comparison and objection guide carries a sources list at the bottom naming exactly what we read. Of the nine competitors we surveyed, one published a sources list.
- Check the prices live. Prices come from a data layer that refreshes against the retailer, and every price on the site carries the date it was checked. If that check is more than 48 hours old the number disappears on its own and the button falls back to "check price". It is structurally impossible for this site to show you a stale figure.
The criteria, by category
Different products are decided by different specifications. These are the axes we rank on, and they are the same ones every time so a pick in one roundup means the same thing as a pick in another.
How products are excluded
Every roundup on this site contains at least one entry telling you not to buy something, and that is a rule rather than an accident. A list where all eight products are excellent is a list that has told you nothing.
Products get excluded or demoted for these reasons, in roughly this order:
- The category is wrong for the job. A claw scraper is genuinely good at bonded ice and genuinely likely to mark trim, so it is ranked last with the reason stated.
- The maker publishes nothing checkable. Where two similar products exist and one prints its specifications, the one that prints them wins.
- The design does not survive contact with cold. A telescoping lock that fills with water and freezes is a real failure mode, and it is a mark against a tool that has one.
- It solves a problem you probably do not have. More product is not better product, and an oversized recovery board in a sedan trunk is a worse purchase than a smaller one.
Every individual pick also carries a "skip this if" line naming the buyer it is wrong for. That is not hedging - it is the most useful sentence on most product cards.
Where we deliberately do not pin a product
Some things are fitment parts: wiper blades come in sizes that differ between the driver and passenger side of the same car, and batteries come in group sizes that physically will not swap. For those, our links go to a search rather than to one specific listing, because sending you to a 20 in blade when your car takes a 22 in one is worse than no link at all.
For the same reason we do not publish a tire roundup. Tires are fitment-heavy, expensive to return, and we would need vehicle data we do not have to make a recommendation honestly.
How prices work
Prices are never written into a page. They are injected at render time from a live data layer that checks the retailer, and they carry a visible "as of" date. There is a hard freshness gate at 48 hours: past that the number is suppressed automatically and the button reverts to "check price on Amazon".
The 48-hour window is deliberately tighter than it needs to be. Amazon's terms want displayed prices refreshed roughly daily, and 48 hours allows exactly one missed refresh before the site stops showing numbers rather than showing old ones.
How often this gets redone
Every page displays a "last reviewed" date, and that date only moves when somebody has actually re-read the page. Bumping a date without re-checking the content is the thing that makes review dates meaningless everywhere else, so we do not do it.
The single test every sentence has to pass
Could a competitor, a journalist or a Google quality rater check this claim and find it false? If yes, it does not ship. That rules out every testing claim, every credential we do not hold, every remembered price and every invented rating - and what it leaves is the part that was worth reading anyway.
If you spot something on this site that fails that test, tell us: info@scrapeandshield.com. Corrections are made visibly and the review date moves when they are. The editorial policy covers how that works.