Methodology
How My Cruise Intel works
No AI guesswork. Every recommendation traces back to a fixed, rule-based method applied the same way to everyone.
Methodology version 0.2 · ship catalog snapshot last reviewed 2026-08-20. Facts are shown with source and confidence context where available.
It's a deterministic score, not an AI opinion
My Cruise Intel doesn't ask a language model what it thinks you'd like. Your answers are run through a fixed scoring method: each of the 14 preference dimensions you rate (family time, luxury, dining, nightlife, relaxation, and so on) is compared against the same, consistently-rated dimensions on all 231 ships, and combined into a single match score. Give the same answers twice and you get the same result twice — nothing is generated, reworded, or reinterpreted by a model in between.
A few things you tell us — like traveling with kids, cruise lines to avoid, or a target price range — aren't scored at all. They're hard filters: a ship that doesn't meet them is removed from consideration entirely, not down-ranked.
Confidence is a separate signal from match quality
A high match percentage and high confidence aren't the same claim. Match percent answers "how well does this ship fit what you told us." Confidence answers "how much should you trust this specific result," and comes from three things looked at together:
How much you told us
A fully answered questionnaire supports more confidence than a partial one.
Whether your priorities agree
If your highest-weighted preferences point in different directions, no ship can satisfy all of them equally well, and confidence reflects that tension honestly rather than hiding it.
How far ahead the top pick is
If the #2 ship scores almost identically to #1, that's a genuinely close call, and confidence is lower even though the match percent might look strong.
We'd rather tell you a result is a reasonable starting point than overstate it as a sure thing.
Every "why" traces back to something you answered
The reasons shown with a recommendation aren't generic marketing copy — they're generated from the specific dimensions you rated highest, the specific ship attributes that scored well against them, and (where it's genuinely true) the specific reasons a close alternative ranked just below it. When a ship is a compromise on something you said mattered, we say so, in the same place as the strengths, not buried in fine print.
Data honesty: verified, derived, or left blank — never invented
Ship and cruise-line facts come from a mix of editorial research and, where noted, official cruise-line sources with a visible confidence level and, when available, a source link. Some attributes (like a specific ship's entertainment mix) are reasonably derived from a cruise line's general pattern rather than individually confirmed — those are labeled as derived, lower-confidence information, not stated as fact. If we don't have reliable information about something, we leave it out of a ship's page rather than guess. See it labeled on a real ship page →
How My Cruise Intel knows what it tells you
Not every fact on this site is equally certain, and My Cruise Intel is built to show that rather than hide it. A growing set of ships have a full evidence file — each important detail (dining, dress code, kids' programming, Wi-Fi, gratuities, common extra charges) carries its own source, a "nature" label, a confidence rating, and a date it was last checked.
Published policies and counts, sourced to the cruise line's own material wherever possible, reputable trade press or review sites otherwise. No cruise-line source caps it at "medium" confidence.
Inferred from a cruise line's fleet-wide pattern, not individually confirmed for that ship, and labeled as such.
How a ship "feels", how crowded it gets, nightlife intensity, the main trade-off — what guests commonly report. Never presented as hard fact, never "high" confidence.
Computed from how long ago a detail was verified against how fast that kind of thing changes. Anything overdue for review is flagged on the page.
When two sources disagree, the most recent primary (cruise-line) source wins, and a note records the disagreement. If something can't be found reliably, the attribute is left off — a missing detail is honest; a guessed one is not. See the Coverage page for exactly how many ships and attributes are sourced today.
Recommendations aren't for sale
No cruise line or travel partner can pay to rank higher, rank lower, or be excluded, and commission never enters the scoring.
Ranking is based only on how well a ship matches your answers. My Cruise Intel may show clearly labeled advertising, and some outbound links to cruise lines or booking partners may earn My Cruise Intel a commission — but that revenue never touches the scoring above. See our affiliate disclosure. If you opt in to being contacted by a travel advisor after seeing your results, that's a separate, explicit choice you make — it never changes which ship was recommended or why.
See it for yourself
The fastest way to understand My Cruise Intel is to run it once and read the reasons next to your own result.
Start my recommendation →