Skip to content

The method

How we research.

Every recommendation on this site is an opinion with a paper trail. This page is the paper trail.

What we will not do

  • No brand can buy a verdict.

    Not a placement, not a score, not a position on the board. There is no rate card because there is nothing on sale.

  • The verdict is written before the link.

    Research finishes, the verdict is written, and only then does anyone attach a retailer link. The order matters, and it is not negotiable.

  • We say when something is not worth it.

    A recommendation nobody can fail is not a recommendation. If the answer is 'buy the cheaper one' or 'wait', that is what it says.

  • Your rating and our score never merge.

    The SortedChoice Score is ours and is measured. The community rating is yours and is reported. Averaging them into one number would hide both.

How a verdict gets made

Nine stages, in this order, every time. None of it is automatic: software and an AI reader help us cover more ground, but the score at the end and the sentence next to it are written by a person who can be argued with.

  1. Set the rubric before anything is opened

    What counts is decided per category, in advance. Headphones are judged on noise cancellation and call quality; a laptop on thermals and battery. Deciding what matters after seeing the contenders is how a favourite quietly gets a rubric built around it.

  2. Do the reading by hand

    Someone sits with the primary material — spec sheets, manuals, teardowns, independent measurements, the fine print on the retailer listing. It is slow and it is not delegated, and it is where most of the errors that would survive every later stage get caught.

  3. Watch the reviews. Plural.

    Long-form video and written reviews from several independent sources, picked to disagree with each other rather than to agree with us. One reviewer's verdict is one pair of ears, one desk, one review unit. A claim is not evidence until it survives more than one of them.

  4. Read the owners, not the launch week

    Owner reports across retailers, forums and communities, read for pattern rather than sentiment: the hinge that fails at month eight, the app that got worse after an update. Launch coverage cannot see any of that, and one furious review is not a fault.

  5. Use it ourselves where we can

    Some of what we cover we buy, borrow, or already own — and then it gets lived with rather than unboxed. Hands-on changes a verdict more than anything else on this list, so we say plainly when it happened and never imply it when it did not.

  6. Argue the case with an AI, on purpose

    The assembled evidence is put to an AI model and pushed back on: what have we missed, which claim is thinly sourced, what is the strongest case against the conclusion we are heading towards. It is a tireless second reader, and it is checked exactly like any other source. It scores nothing and it writes nothing. Our policy on automated tools.

  7. Score each criterion on its own

    Every criterion is scored from 0 to 10 against the category rubric, not against the other products in the round-up. A weak field does not promote a mediocre product.

  8. Write the verdict a person can be held to

    One line on who this is for, one on who should skip it, and the reason attached to both. If neither line can be written yet, the research is not finished and nothing is published.

  9. Revisit when the facts move

    Prices move, firmware changes behaviour, successors arrive. A score carries the date it was last reviewed so you can judge how much of it is still true.

A real one, taken apart

Not a diagram. This is a verdict currently published on the site, and every criterion behind its score.

Our score
  • Camera9.6 / 10
  • Performance7.5 / 10
  • Battery8.0 / 10
  • Display9.5 / 10
  • Software & updates9.8 / 10
  • Build9.0 / 10
  • Value7.5 / 10

These criteria belong to Smartphones. A different category is judged on a different list — that is the point of the rubric, and it is why scores are comparable within a category and not across the whole site.

Two numbers, kept apart

The SortedChoice Score

Ours. One number out of ten, built from the category rubric above, carrying the date it was last reviewed. It answers how good is this, and we are accountable for it.

The community rating

Yours. Stars from people who own the thing, shown with how many. It answers what is it like to live with — a question our rubric cannot reach.

They are never averaged together. A single blended figure would let a strong measured result bury a chorus of owners saying the hinge breaks, and it would let a review brigade overwrite a measurement. You get both, side by side, and you decide which one matters more for you.

Where our verdict stops

Everything above produces one opinion, held by us, written for a reader we had to invent. It is the start of your decision, not the end of it.

The rubric in §02 is a set of choices — which criteria count, and how much. We make those choices for a general reader in India buying at today’s price, because a score has to be written for somebody. It is not written for you specifically, and the distance between the two is real: what you are buying it for, what you already own, what you would pay, where you live, and plain preference all move the answer, and none of them are visible from here.

So the pages are built to be argued with rather than obeyed. Every score publishes its criterion breakdown, so you can ignore the total and read the one line you care about. Every verdict carries Best for and Not ideal for, so you can find yourself in the second column and buy it anyway on grounds we could not see. A reader who does that has used this site correctly.

What we will not do is hedge our way out of a conclusion. “It depends” is not a verdict, and a recommendation that cannot be wrong is not worth reading. We say what we think, we show the working, and we leave the disagreeing to you — the full version of that boundary is in our disclaimer.

How we pay for this

When you follow a link to a retailer we may earn a commission, at no extra cost to you. That is the whole business model, and it is the reason the order of operations in §01 is written down: the verdict is finished before anyone knows which retailer will carry the link, and the commission does not vary with what we conclude. We also do not sell products, run a basket, or hold stock — when you decide, you leave.

If you ever find a page that reads like it was written for a retailer rather than for you, tell us. That is a bug, and we treat it like one.