How we pick and rate

Every pick on this site carries a Build a Boyfriend score out of five. It is an editorial opinion — ours — and not a customer rating.

What we are honest about first

We have not lab-tested these products. Nobody here has worn all 700 of them. This is a curated shopping site, and the score reflects how strongly we would recommend a thing to a friend with that budget and that problem — not a controlled comparison. Anywhere you see a number on this site, read it as an editor's call.

We also do not republish retailer star ratings as our own. Amazon's review average belongs to Amazon and to the people who wrote those reviews. It informs whether a product gets onto the site at all, but it is not the score.

Getting onto the site at all

Before anything gets a page, it has to clear a sourcing bar. There are two, because the risks are different at different prices.

The mass-market bar

For ordinary Amazon picks — the jumpers, the tools, the kitchen equipment — the danger is counterfeits and listings that vanish. So, at the time we sourced it:

A real brand. No rotating three-letter labels, no listings that will be a different company next quarter.
At least 900 ratings, at 4.2 stars or better. Not because volume equals quality, but because a deep, established listing is far less likely to be a counterfeit or to vanish mid-season.
Not a variant of something already here. One product, one page. We do not pad the site with the same jumper in four colours.
Men's, and actually the product. Not the carrying case for the product, not the replacement filter.

Plenty of things fail this. A Confederate-uniform costume, a shelf of no-name shirts, three near-identical wallets and a nursery glider marketed as a recliner have all been thrown out rather than listed.

The luxury bar

The review test does not work at the top of the market. A four thousand dollar Swiss watch sells in tens, not tens of thousands, so demanding 900 ratings would exclude every genuinely expensive thing on this site. Requiring it would not make us careful; it would just make us cheap.

So for luxury picks we use a different test: the piece must come from a brand-operated storefront — Longines, Vilebrequin, Giuseppe Zanotti, Stefano Ricci, Moon Boot, MATEO New York, SIMKHAI and the like selling directly, rather than a third party reselling them. That is what answers the counterfeit question here, and it answers it better than a review count would.

These carry no rating volume, and we are not going to imply otherwise. Where a pick came in on the luxury track, the review underneath tells you what we think of it and what the trade-off is — including, in several cases, that we think it is too expensive.

What the score weighs

Once something is on the site, the score out of five reflects four things:

Does it do its one job? The single biggest factor. A $20 griddle press that stops bacon curling scores well against its job. A gadget that half-solves three problems does not.
Value at its price. Judged inside its own band, not across the site. A $25 cast iron skillet and a $6,000 watch can both score highly; they are answering different questions.
Brand track record. Whether the maker has a history of the thing still working in five years, and of honouring a warranty when it does not.
Listing stability. How established and deeply reviewed the listing is — our proxy for “this will still be here, and still be the same product, when you click”.

Reading the number

The scale runs from 3.5 to 5, and that is deliberate. We do not score things we would not recommend — anything that would land below 3.5 is simply left off the site, which is the more useful editorial act than publishing a bad review of it. So the number is not “how good is this out of everything”; it is “how strongly do we stand behind this, among things worth buying”.

4.8 to 5.0 — the default answer in its category. If someone asked us for one recommendation and nothing else, this is it.
4.4 to 4.7 — excellent, and what most of the site looks like. A safe buy with no serious reservation.
4.0 to 4.3 — a good buy for the right person. Usually specialised, or the value case depends on the price holding.
3.5 to 3.9 — it earns its place for a specific reason, and there is a real trade-off in build, longevity or how narrow its use is.

Most picks carry the score on its own. Where we have written a longer note underneath it, that is us flagging something specific worth knowing before you buy — usually the catch.

Commission does not move the number

Commission rates vary by category and we ignore them when scoring. They are also not the reason anything is on the site. Our affiliate disclosure sets out how we get paid, and shipping & returns covers who actually fulfils an order.

Corrections

If a pick has got worse — a brand sold, quality dropped, a listing hijacked — we would rather hear it and pull the page than defend a score. Nothing here is too old to change.