Methodology

What the figures beside each score mean

Market

Where the brand was measured. Brand strength is specific to a market: a brand measured in Canada is being read on its Canadian brand strength, which is not the same as its strength anywhere else, and the page says which market applies rather than leaving it to be assumed. A small number of brands read Not location-based, because the reviews behind them describe customers of an app or a product rather than visitors to a place, so no country attaches to them.

Score computed

The date the arithmetic ran. Not the date the evidence was collected - the reviews behind a score span many dates, and a single date over them would claim a freshness the underlying data cannot support. It says when this number was last worked out from the evidence on file, which is the honest version of the same question.

Category rank

Where a brand sits among the brands it is genuinely comparable with in its category. Brands shown without a rank - because they are measured differently from their peers, or are kept on record rather than scored - are left out of both the position and the total, so they do not push other brands down a place.

Sub-category rank

Where a brand sits among the brands it competes with most directly. A category like Retail holds shops that do quite different jobs, so each brand also sits in a smaller group - office and electronics, beauty and pharmacy, home improvement - and this is its position inside that group. A brand can rank differently in its sub-category than in its category, and both are true at once. A brand can lead a small group of close competitors while sitting mid-table against its whole category, or the reverse. Both are shown because they answer different questions: the category rank says how the brand does against everyone it is comparable with, the sub-category rank says how it does against the brands a customer would actually weigh it against.

It also explains the comparative measures. Three of the measures behind a score are comparisons - search demand, audience resemblance and price - and each is taken within this smaller group rather than across the whole Index, which is why a brand can read 100 on one of them: it leads its group, not the Index. Where the group is too small for a comparison to mean anything, no rank is shown - the card says the brand is the only one in its group, or that there are too few brands to rank, and the comparative measures are not published for that brand either. A position among one or two brands would claim a comparison we have deliberately withheld everywhere else on the same page.

Category Audience Fit

One of those comparisons needs saying plainly, because its name is shorter than what it measures. Category Audience Fit is how much a brand’s audience resembles the audiences of the brands it is compared with - the publications, voices and interests those audiences follow, and how much of that they hold in common. A high score means a brand reaches much the same people as the brands beside it; a low one means its audience sits somewhere else. Neither is good or bad on its own.Brands in a category usually share an audience, so resembling your competitors is ordinary rather than weak; and a distinct audience can mean a brand has specialised deliberately as easily as it can mean the brand is isolated from the conversation its category is having. The number says which people, never whether that is working. It is one of the measures being reviewed after launch, and the intended replacement asks a different question: not whether a brand’s audience looks like its competitors’, but whether it connects to the occasions the category is actually bought for.

Category index

A brand's score against the average for its category, where 100 is that average. A brand at 110 scores ten per cent above the typical brand it competes with. It is only ever comparable within its own category - the averages differ from one category to another, so a 110 in one is not a 110 in another. One further caution: Retail's brands sit closer together than any other category's, so a small difference in score moves the index more there than it would elsewhere.

Evidence level

How much a score rests on, in one word. Supported means all six factors scored, at least 750 reviews carrying written text, both of the measures that compare a brand against its peers, and a comparison group of at least five brands. Directional is the common case: a real score with a real basis, and a gap somewhere. Indicative means the reading is thin or uneven and should be treated as a first look. Unscored means we have not published a score at all. There is deliberately no level above Supported: nothing in this edition has been tested against survey research or business results, and a stronger word would claim exactly that.

How much we read from any one source

There is a limit on how many reviews we take from a single place, product or app. Without one, a brand's busiest location or its bestselling product would supply most of what we read, and the score would describe that one thing rather than the brand. The limit trades depth for balance deliberately - it is why a brand with one very busy store does not outweigh one with many quiet ones.

The limits differ because the unit differs. A location and a product are each one of many a brand has, so a modest limit on each still adds up to a broad reading. An app listing is the only one a brand has, so the same limit would leave a much smaller corpus - and that would measure how much we collected rather than how much evidence exists.

Which products the reviews are of

Where a brand is read from product reviews, we take the five products with the most reviews. That is a narrow reading of a brand that may sell hundreds of things, and it has a consequence worth stating: the most-reviewed products are often not the ones the brand is known for. Accessories and consumables - covers, filters, refills, replacement parts - sell in far greater numbers than the machine or the appliance itself, and they collect far more reviews.

Where a brand's main product is not sold on the marketplace we read, or is sold in numbers too small to review, the score describes what customers say about that brand's consumables rather than about its products. That is a real reading of a real thing, and it is not the same reading as one taken from the product itself.

The share differs between brands in the same group, which is the part that affects a comparison. One brand may be read almost entirely from its machines and another from its attachments, and their scores then describe different things while sitting side by side. We measured this across the Hard Goods brands in August 2026 and are correcting it where the gap is widest; a full correction needs a collection pass that this edition did not have room for. Where you are comparing two brands closely, it is worth knowing.

What a brand is compared against

Every comparative measure is taken within a brand's comparison group - the brands in this Index it actually competes with - rather than against its whole market. A brand leading its group scores 100 on Relative Search Salience; that says it is the most searched-for brand among its peers here, not that it holds all of the search in its category.

It follows that these figures move as the Index grows. Adding a larger competitor to a group changes what the others are measured against, which is why we show the monthly search volume beside the score: the volume is the brand's own, and it does not move when the group does.

When we cannot measure something

A brand only carries a score once there is enough evidence behind it. Below that threshold the page says so rather than publishing a thin number. Where a measure cannot apply to a kind of brand - a gym has no product price history, a card network has no premises to review - the page marks it as such rather than counting it as a failure. And every number opens back to what produced it: the measures behind a factor, the sources behind those, and the date each was collected. If a score looks wrong, it can be checked rather than argued about.