We Read Eight Aurora Guarantees in Full. Not One Promises You the Sky

This is a snapshot with a date on it, and the date is the point. Our permanent page on what an aurora guarantee actually guarantees is a census: it counts how many of the 307 departures in our catalogue carry a money-back clause, and it deliberately names nobody, because operator wording changes between seasons and a named quotation ages into a misquote. This article does the opposite job. It names the companies, quotes the clauses, and prints the day they were read, so you can tell how stale it has become.

The eight: Aurora Experts, Ultimate Iceland, BusTravel Iceland, Arctic Adventures and Iceland Adventure Tours in Iceland; Rødne Fjord Cruise, NorthernShots Tours and Arctic GM in Tromsø. Every quotation below comes from a page we opened ourselves.

Two things we went looking for and did not get. The clause that circulates most widely in this industry is the one making a photograph the test of success, and the Tromsø operator most associated with it publishes that policy through resellers; its own site did not open for us, so it is not quoted here. We also found no operator setting a minimum duration, the "five minutes of visible aurora" test that gets repeated in forums. If it exists, it was not on any of the eight pages we read.

Start with the disclaimers, because they are the most honest sentences in the sample and they all say the same thing.

Rødne, which sells the remedy hardest of the Tromsø three, uses one line: "But there is no guarantee with such a natural phenomenon." NorthernShots writes "The Northern Lights are a natural phenomenon and cannot be guaranteed." Aurora Experts sets out the three conditions it cannot control, "For a sighting, three things must come together: solar wind, darkness and a clear sky," and then "Neither displays nor their magnitude can ever be guaranteed." Arctic Adventures: "sightings can never be guaranteed."

So the industry knows, writes it down, and sells the word anyway. That is not quite hypocrisy. The promise is real; it is a promise about something else. Rødne is the clearest about what the something else is, and the sentence repays reading slowly, because it is the entire business model in one line.

Everything downstream depends on this, and most of the eight do not define it at all.

Three say something explicit. Aurora Experts and Ultimate Iceland both use that identical five-word sentence, "Low-level lights count as sightings." A low-level aurora is the faint grey-green smear that people who have travelled a long way often fail to recognise as the thing they came for. Under those terms it is a successful tour, and the remedy does not trigger.

Then there is the one that runs the other way, which is the finding we did not expect. The terms on the Iceland Adventure Tours private aurora page set the test at the eye: "A sighting means that the Northern Lights can be seen with the naked eye," with the page adding that even a faint green line for a short time counts. And then, explicitly: "If the aurora is visible only through a camera, the supplier considers this a non-sighting."

That is the reverse of the clause this industry is rumoured to run on. A camera at a fifteen-second exposure records a green band where the eye sees nothing, so whoever controls the definition controls the refund. One operator in our sample wrote the definition down and wrote it in the customer's favour. Five left it blank, which in practice means the guide on the night decides.

The blank rows are not an oversight. A test that is never written down cannot be argued with afterwards, and on a marginal night the person deciding whether you saw the aurora is the person whose company pays if you did not.

Five of the eight rule out refunds for a failed night in language so close that it suggests a template circulating rather than eight independent decisions.

Aurora Experts: "Under no circumstances will a refund, full or partial, be given for a non-sighting tour." Ultimate Iceland: "Under no circumstances will there be a refund, in full or partial, for a Non-sightings tour." BusTravel Iceland: "Refunds are not issued for unsuccessful tours." Arctic Adventures: "Refunds aren't issued if the tour takes place, but the Northern Lights don't appear." Iceland Adventure Tours: "Refunds are not provided when the tour operates but a Northern Lights sighting is not possible," and separately, "Bookings are non-refundable. All sales are final."

The other two rule out cash on the remedy itself. Rødne's free ticket "is non-refundable and cannot be exchanged for cash." NorthernShots, whose remedy is a discount rather than a free night, writes that "Refunds are not available for this offer if you cannot take a tour on another day."

Arctic GM is the outlier, and the most interesting case here. Its tour page carries the flattest promise any of the eight makes: "We guarantee that you will see the Northern Lights during our tours." Its terms page, which is where a promise like that would have to live to mean anything, contains no definition of a sighting and no money-back clause for a failed night. What it contains is discretion. Refund requests are approved or denied "at its sole discretion," and where the terms do offer a remedy, customers "will be offered rescheduling or a refund, at Arctic GM's discretion." The strongest wording in our sample sits on the page that sells rather than the page that binds.

Here is where the eight separate properly, and it has nothing to do with how generous the word guaranteed sounds.

Aurora Experts gives a free Re-Try Tour "valid for up to four years from the date of your original tour," and adds that "The number of free Re-Try Tours is not limited." Ultimate Iceland: two years, also unlimited. BusTravel Iceland: "The offer for rebooking is valid for three years from the date of your tour departure," but it "can be used once during the validity period." Arctic Adventures: up to three years, unlimited, with a catch worth noticing, because "retries are operated on a minibus, regardless of the original tour type." The replacement can be a cheaper product than the one you paid for.

Then Rødne, at the far end. The code you are handed is "subject to availability, valid only for this season," and "This offer does not apply to the final cruise of the season on March 31."

Four years against one season. Same remedy, same phenomenon, and a difference in real value of roughly an order of magnitude, because almost everybody books an aurora tour inside a trip of a few nights. A re-try valid for four years is an asset: you can spend it on a different trip, in a better month, at a better point in the lunar cycle. A re-try valid only this season, handed to you on a Tuesday in March when you fly home on Wednesday, is worth nothing at all, and it is the version most buyers are holding.

A failed tour and a night that misses our cloud threshold are different events. Keeping them separate matters when interpreting a free second attempt.

In the previously published fixed-location model, Reykjavík has a usable-sky frequency of 35.7%, and a conditional estimate of 30.2% for the night immediately after one that did not meet the cloud criterion. For Tromsø the corresponding figures are 25.4% and 19.1%. These describe the historical cloud series at each location, not the outcomes of the operators’ tours. The destination context is on our Reykjavík page.

The difference illustrates weather persistence in that model. It does not measure a voucher’s cash value, and it cannot be applied automatically after a failed sighting: auroral activity, timing and the route also affect what a tour sees. A retry taken weeks or years later is a different situation again.

Nor can the complement of the usable-sky percentage tell us how often a guarantee is triggered. A short opening can allow a sighting without satisfying our threshold, and a tour may travel beyond the fixed location. We have not measured the operators’ retry or success rates.

Our screening rule counts a night as usable when at least two of the nine hourly samples from 18:00 through 02:00 local time have cloud cover at or below 50%. Those samples need not be consecutive. This is a historical planning indicator, neither an aurora-sighting probability nor a measured upper bound on sightings. Clouds, darkness, geomagnetic activity and a moving tour’s route are not combined into an observed tour-success rate here. The assumptions and source limitations are explained on our weather-method page.

A retry can be useful when you have another available evening and its conditions allow you to use it. Its value depends on that practical opportunity.

The previously published Reykjavík model estimates the chance of at least one night meeting its fixed-location cloud criterion as 35.7% over one night, 55.1% over two, 68.6% over three and 84.7% over five. These are model outputs with weather persistence assumptions, not observed aurora or tour-success rates. A five-night stay is the model’s time window; a voucher does not itself create those weather conditions or guarantee that five departures are available.

For Tromsø the corresponding published one- and five-night estimates are 25.4% and 68.0%. They should not be compared directly with an operator’s advertised sighting percentage, which may use a moving route and a different definition of success. The calculation and its limits are discussed in our article on trip length in Tromsø.

Read the remedy on its own terms: a discount requires another purchase; a retry requires time and availability; a refund returns money under the conditions stated. In our dated catalogue census, fourteen of the 307 listings carried a money-back clause, all in Finland. That count is a description of the listings reviewed, not a prediction of a claim’s outcome.

Our guide already lists the six questions to put to an operator before booking, and that list has not changed. What this reading adds is which answers to insist on having in writing, because these are the points that moved most between the eight companies.

1. The expiry date, in months or years. Not "valid for a future tour".
2. Whether the re-try is unlimited or single use.
3. What vehicle the re-try runs on, if you booked anything other than a coach.
4. Who decides on the night, and whether a camera counts.

Leave time for another attempt if your itinerary allows it, and compare the cost, route and conditions of each option. We have not established that two cheaper tours always beat one tour with a guarantee. Weather cancellations are a separate matter, covered in our guide to refunds, cancellations and cloudy nights.

Every clause quoted above comes from one of these pages, each opened on 14 October 2026. Operator wording changes between seasons, and the version that binds you is the one on the page you book from.

• Aurora Experts, terms: auroraexperts.is/terms
• Ultimate Iceland, terms of service: ultimateiceland.is/terms-of-service
• BusTravel Iceland, northern lights tour: bustravel.is
• Arctic Adventures, northern lights FAQ: adventures.is
• Iceland Adventure Tours, private northern lights tour: icelandadventuretours.is
• Rødne Fjord Cruise, northern lights chase: rodne.no
• NorthernShots Tours, aurora chase: northernshotstours.com
• Arctic GM, northern lights tours and terms: arcticgm.com

The cloud and persistence values are the previously published outputs of an Open-Meteo archive analysis spanning ten September–March seasons within September 2015–April 2025. The saved request did not specify a weather model, so exclusive ERA5 provenance is unverified. This correction of 20 October 2026 does not recalculate those values. See our weather-method page for assumptions and limits. The operator terms remain a snapshot read on 14 October 2026; this article is not legal advice.