Cruise Price Index
Methodology
The index measures how the price of the same cruise changes between two Monday loads, then chains those changes into one series. This page says exactly how, and where it can mislead.
What goes in
Each Monday we load every future sailing from our supplier feed, with its fare in four cabin categories: inside, outside, balcony and suite. Fares are whole US dollars. We keep every load, so each sailing has a price history of its own. The first load in the series is 6 September 2026; a second followed on 7 September, and weekly loads have run since.
A cabin with no fare is not offered or sold out. The feed does not tell us which, and we do not guess.
The pair statistic
Take two consecutive loads. Keep the sailings that have a fare in both. For each, compute the natural log of the new fare over the old fare. Clip that value to the range minus 0.5 to plus 0.5, which stops a handful of corrupted or one-off fares from steering the result. The week's statistic is the mean of those values.
Using logs makes a rise from $1,000 to $1,100 and a fall from $1,100 to $1,000 cancel, which a plain percentage change does not. Averaging over matched sailings means a new, expensive sailing entering the feed cannot lift the index; only a price moving on a sailing that was already there can.
The chain
The index is 100 at the first load. Each later value is the previous value multiplied by the exponential of that week's mean log change. A reading of 100.49 means that, after chaining every week since the start, the same sailings cost about 0.49% more than they did on 6 September.
Which fare counts
The headline series is labelled any. For each sailing it takes the lowest fare among the four cabins, and it uses the same cabin in both loads, chosen by the older load. That avoids a sailing "falling" because a cheaper cabin type appeared that was never compared against its past. Per-cabin series (inside, outside, balcony, suite) compare that cabin alone.
Slices
Beyond the headline, the index is published for each cruise line, and for segments:
- Booking window, measured from the older load to departure: 0 to 90 days, 91 to 180, 181 to 365, and more than 365.
- Sailing quarter, for the next eight quarters.
A slice is only published when at least 30 sailings match in both loads. The all-lines series is always published.
The other numbers
- Cut and raised
- The share of matched sailings whose fare fell, or rose, between the two loads.
- No fare
- Of sailings that had a fare in the older load and are still listed, the share that no longer show one. This is the nearest thing we have to a sell-out rate, and it is an upper bound on one, since a fare can also be withdrawn.
- Matched
- The number of sailings behind the figure. Read small ones with care.
- Dollars a night
- The mean of the new fare divided by the sailing's length in nights, over the matched set.
Caveats
- Snapshots, not live fares. We see one price a week. A fare that dropped on Tuesday and recovered by Sunday never appears.
- Lead-in fares. The feed gives one price per cabin category, which reads as a "from" price. It moves with promotions and with whichever cabin is cheapest, and it is not what a particular traveller pays. We do not break out taxes, port fees or gratuities.
- Short history. The series is weeks old. Seasonal patterns, which dominate cruise pricing, cannot show up until a full year of loads exists, and until then weekly moves should not be read as a trend.
- Departed sailings leave the set. A sailing drops out of the matched set once it has sailed. The index follows the sailings still on sale.
- Clipping. Winsorising at 0.5 in log terms is far wider than a normal weekly move, so it matters only for outliers, but it is a choice and we state it.
Reproducing it
On the Business plan, the API returns the full series for any line, cabin and segment, and every sailing's fare history, so you can recompute the index yourself. Questions or corrections: [email protected].