The best of Ibiza 2027

Each experience scores up to 100 points. Most come from its Google rating; the rest from our editorial 1-to-5 scores (crowds, weather dependence, fit for children) or from the price, depending on the list. The weights for each list are in the table.

The adjusted rating, in plain words

A 5.0 from three reviews says little. That is why we use a Bayesian average: it is as if every place started with 100 imaginary reviews at the average score (4.67, the mean across the whole directory). The more real reviews it has, the more its own score counts. We then map the result onto the useful range: an adjusted 3.5 or lower earns 0 points and a 5.0 earns them all, because almost everything on Google falls between those two numbers.

adjusted = (reviews × rating + 100 × 4.67) ÷ (reviews + 100)

Example: a 5.0 from 25 reviews ends up at 4.73; a 4.8 from 3,000 at 4.80. The second one wins.

The rules

• To be ranked, an experience needs a current Google rating (retrieved less than 30 days ago through the official API) and at least 20 reviews. Everything else is listed separately as “not enough data”.
• One experience per operator: the Google rating belongs to the business or place, not to each individual trip. Its other experiences are shown under its entry.
• If an editorial score or the price is missing, it counts as the list average (neither rewarded nor penalised) and we flag it in the breakdown.
• Ties go to the one with more reviews. We show up to 10 places and only publish a list when at least 5 experiences from different operators can be scored.
• Nobody pays to appear or move up. Affiliate links, advertising or an operator contacting us play no part in the formula.
• We do not score value for money or duration: hardly any listing has reliable data for the former, and the latter depends on what you want. We show them so you can decide.
• The lists are recalculated every time we update the data; Google ratings once a month.

Atlas methodology