How we rate

Every week we read the Connect IQ Store: the first 100 entries of 31 store categories, 1,065 watch faces and 1,780 apps, widgets, data fields and music apps in all. Each entry gets a score from 0 to 10, built from seven public signals:

• Description (0–3): Does the store description explain what the face or app really does? Length, structure and real features count; hashtag stuffing and empty phrases do not.
• Rating (0–2): The average star rating, pulled towards 4.0 when there are only a few reviews, so three five-star reviews do not beat three thousand 4.8 ratings.
• Popularity (0–1.5): How many reviews it has collected in total (on a logarithmic scale).
• Momentum (0–2): How many new reviews arrived in the last 90 days: is it popular right now?
• Care (0–1): Updated within the last 6 or 12 months, and translated into several languages.
• Reach (0–0.5): How many watch models it runs on.
• Originality (0–1.5): Is it an idea of its own or one more variant of something common? Rated by hand for the top entries and the hidden-gem candidates; everything else gets the middle value.

The seven parts add up to at most 11.5 points; the sum is scaled to 0–10. A developer who publishes five or more near-identical variants of one design loses up to 2 points, so the lists show different ideas instead of the same face ten times.

Hidden gems

A hidden gem has fewer than 30 reviews, a clear description, is kept up to date, has no variant penalty and an originality of at least 1.0. We show 50 watch faces and 50 apps.