What an Aurora Trip Costs

This page used to say that we would not publish price figures until we could calculate them properly, from a full catalogue, with a stated sample and a date. That condition is now met, and the four rules we set ourselves are the ones the tables below follow: median rather than average, a declared minimum sample, a date on every number, and like compared with like.

Everything here comes from the 307 departures we list, read on 10 August 2026. Prices are the lowest published adult rate for each departure, in euros, as the provider feed reports them.

Two figures for each place. The first covers every departure we list, including private vehicles and multi-night packages. The second covers shared seats only, which is what most people are actually buying, and is the number to compare across destinations.

Five destinations are missing from this table on purpose: Bodø, Saariselkä, Anchorage, Whitehorse and Abisko each have fewer than five shared departures, and a median of three is not a median. They get no figure rather than a weak one.

The 400 EUR line separating a shared seat from a whole vehicle is our threshold, not an industry standard, and we state it here for the same reason we state the cloud thresholds: someone rerunning the calculation should be able to get our answer, or explain why they got a different one.

Three things fall out of that table. Yellowknife is less than half the price of anywhere else, at 79 EUR against a European cluster around 140 to 190. Small markets are expensive: Alta at 242 and Kiruna at 184 have five and six departures respectively, and no competitive pressure at all. And Tromsø costs a third more than Rovaniemi, which is worth holding onto for the next table.

A tour price on its own is half a number. What a traveller is buying is not an evening in a minibus, it is a chance of seeing something, and those chances are wildly different between destinations. Divide the median price by the share of season nights on which the sky is usable and you get a rough index of what one successful night costs.

Nobody publishes this, as far as we can find, and both of its inputs are on this site with their methods declared.

Yellowknife comes out roughly four times cheaper than Tromsø, and it is not because the tours are worse. It is a cheap market sitting under a good sky, and the busiest aurora town on earth is an expensive market sitting under the worst sky in Norway.

Now the caveats, because this index is easy to over-read.

It assumes one tour per night at the median price, which is a simplification. It uses a fixed-point cloud statistic, which measures the sky above the town and not what a vehicle can reach: that penalises exactly the destinations whose operators drive furthest, and Tromsø, with working radii of 200 to 350 kilometres, is penalised hardest of all. And it ignores flights and accommodation, which for most people are larger lines than the tours.

So read it as a comparison between places, not as a budget. What it does show reliably is the shape: the market prices tours on demand and competition, and prices them almost independently of the odds.

Search any of the main cities and the same night's tours will span a range of several times over. The listings look alike. The products are not, and a few things account for almost all of the spread.

The one worth dwelling on is the guarantee. It adds far less to the price than group size or duration does, which tells you what operators think it is worth. If you are choosing between a guaranteed short tour and an unguaranteed long one, the long one is usually the better purchase. Our page on what guarantees cover explains why, and includes a census showing that only fourteen of these 307 departures carry a money-back clause at all.

People optimise the tour price and then spend several times that amount on everything around it.

The consequence is worth stating directly: optimising the tour price is optimising the fourth largest line. Moving your dates out of the Christmas peak affects the first two, which are several times bigger, and it usually improves your cloud odds at the same time. In Rovaniemi the difference between December and March is 22.3% and 50.0% of usable nights, in the cheaper direction.

• Travel in shoulder season, with one exception. September, October, February and March are cheaper than December and, in Europe, statistically better. In Fairbanks the rule reverses: January is the best month there at 58.1%, so the cheap months and the good months do not coincide.
• Book two cheaper nights instead of one expensive one. Two chances at a moderate tour beat one chance at a premium one, because the binding constraint is weather rather than service quality.
• Skip the guarantee premium. Put the difference into a second night.
• Book the tour locally, late. Once you can see a two-day forecast you are buying with information the advance booker did not have. The risk is that good small operators sell out.
• Share a private tour. Four people in a private car often costs about the same per head as four separate small-group seats, with far more control.
• Consider the cheap market. Yellowknife at a 79 EUR median under a 47.8% sky is the best value in this catalogue by a wide margin, and almost nobody in Europe considers it.

• Prices: the lowest published adult rate for all 307 departures in our catalogue across Norway, Iceland, Finland, Sweden, Alaska and Canada, read on 10 August 2026. Reykjavík's sample is 60 of 61 departures.
• Cloud statistics: ERA5 (ECMWF, Copernicus Climate Change Service) via Open-Meteo, eleven seasons; full method on our when the sky opens page.
• Our thresholds, stated so the calculation can be repeated: 400 EUR separates a shared seat from a private vehicle or multi-night package; a destination needs five shared departures to receive a median.
• Consumer rights on cancellation and refunds: Your Europe, package travel.