How to Plan a Trip in the Order That Keeps It From Falling Apart

Trip plans rarely collapse at the airport. They collapse in the weeks before, when one decision gets taken ahead of the information it depended on.

Most advice on how to plan a trip hands you a numbered list of steps. The harder question is the order, and the order is not about effort — it is about which decisions are still recoverable when the next one comes due. A museum may not sell you a ticket until a quarterly release opens, a railway may not list a train until a seasonal sale begins, and a passport takes weeks you cannot compress by looking harder. Every rule below was checked on 20 September 2026 and belongs to the country or company that sets it, any of which can change it — the official page for your own destination and nationality is the one that decides. Locking the wrong thing first is what makes everything after it unrecoverable, and every deadline here comes from one of three constraint types. Telling them apart is where the sequence starts.

Why trip plans fall apart before the trip does

Three constraint types sit behind every booking, and they fail in different directions. Something is not yet possible: a sale has not opened, a booking window has not begun, a release date has not arrived. Something is no longer free: a deposit started a clock, a free-cancellation window closed, a fare was always non-refundable. Or something simply takes longer than expected: a document is issued on the government’s schedule, not yours.

None of the three announces itself as a deadline. You meet them as a search that returns nothing, a refund that does not arrive, or a departure date that arrives before a passport does.

Ordering decisions by price is what makes the failure expensive, because the biggest number is usually the least flexible thing in the plan. A traveller who booked a Milan–Rome–Athens routing described what that costs. Writing on Travel Stack Exchange in October 2018, a traveller called Aditya explained that he had booked everything well in advance, as separate tickets on Air Italy and Ryanair. When Air Italy cancelled two legs and offered other days, the accepted answer was blunt: “because you booked your Air Italy and Ryanair flights separately, your connections are not the airline’s concern.” A commenter put it more plainly — the Italian carrier had no responsibility for flights booked with other carriers. Nothing about the fare was unfair. The ticket structure had decided, months earlier, who would absorb the loss, and that structure was chosen at the moment of the very first purchase.

That is the general shape of the problem: correction cost rises with time, so the decisions worth slowing down are the ones that cost the most to take back. A hotel booking is nothing like that. On a booking-error thread in January and February 2025, a traveller called Steve T described a booking made six months ahead through a credit-card travel portal: a browser tab had silently swapped his dates, he confirmed a non-refundable reservation by mistake, and he was on the phone within minutes. The hotel would not move the dates, on any terms he offered. A $100 credit was the whole recovery on a booking he had paid about $500 for. One reply on that thread, from a poster called Kate, matters more than the outcome: “non-refundable nearly always means exactly that: non-refundable for any reason including an unintentional error.”

What made Steve T’s mistake survivable for someone else is what had not been bought yet. Another poster, Wanderlust58, described the same error with a cheaper ending: the bookings “were made far enough in the future and I had not yet purchased train/bus tickets, so I adjusted my itinerary around my mistake.” The hotel was the same kind of commitment. The downstream commitments were not, and that is the difference between a detour and a loss.

That is why these decisions need ranking rather than listing. Lock in the unrecoverable, keep the recoverable open, and sort the awkward middle by who actually holds the change. Three questions do that work, and the next two sections turn them into a test you can run on your own trip.

The rule that decides what you book first

The rule is: book the scarcity first, in the order that a published deadline forces — and cheap-to-change purchases fill in around it, never lead.

The popular version of this rule says to book flights first, and it comes with a reason. On a thread asking what to book first, opened on 24 September 2021, the original poster was worried about “getting locked into airline tickets and then finding out that there are no rooms available where I am going.” The most cited reply inverted his fear: “There are thousands of hotel rooms available compare to a couple hundred airline seats. The probability of being locked into hotel rooms with no appropriate flights available is greater than the reverse.”

That argument is about scarcity, and scarcity is only one of the two things the rule has to weigh. On the same thread, a poster called Dean described the opposite habit — booking several hotels in different neighbourhoods, then cancelling down to one once the plan settled. Another, Jean, called her early bookings “placeholder” lodging and made sure every one of them could be cancelled. One line from a poster called Chani states the second principle better than any list: “I usually wait to book trains since the tickets are usually no-refund, no-exchange. I book rooms as early as possible to get the best selection and best prices, but I make sure everything can be canceled without penalties.” She was not ordering by price or by scarcity. She was ordering by what she could take back.

Airlines are not universally scarce either. On a companion thread, one contributor from the same community reported booking Venice rooms in early August for Carnival and not buying the flight until the following month. The scarce item there is a bed in a small city during a festival, not a seat on a plane. Rebecca, on that thread, described the check that produces the right answer: “We check events first, such as special exhibitions at museums. We check for Bank Holidays, school holidays in England… Then we check hotel prices… Then we check airfare prices.” The order of research is fixed. The order of purchase falls out of what those checks show.

Run the test on each purchase before you pay for anything.

• When does it stop being free? Find the cancellation deadline, the deposit trigger or the fare rules, in the seller’s own terms and in local time. If the deadline is close, or the rate cannot be changed at all, the item moves up.
• Who can undo it? You can, before a stated date. The operator can, subject to its own discretion. Or nobody can. Purchases in the third group get made last and made late, because nothing downstream can rescue them.
• How thin is the supply? Small properties, limited-capacity venues, event weeks and single daily departures are scarce in a way that chain hotels are not. Where supply is thin, scarcity outranks reversibility for that one item.

Two checks answer most of it in practice. A refundable reservation does not shut a door: book the bed you like, keep the cancellation date visible, and cancel once something better or more certain appears. And a booking page will happily show you availability without your money — pick your dates, look at what exists and at what price, then close the tab before entering payment details. The availability question gets answered either way.

When two purchases are both constrained and neither can be moved, the tie goes to the one that carries the other. Book the flight in the middle of the trip before the short hop that depends on it, and never buy an onward leg before you have confirmed the connection can physically exist. Rail is where that bites, because operators open their sales at different times in the year. Booking the first carrier before the second one’s timetable is published commits you to a connection you cannot yet verify. The rolling calendars in the fifth section are what to check first.

Some decisions will fail that test every time, and they come before all of this. Dates and documents are two of them.

Fix the dates against a calendar, not a wish

Choose your dates, then attack them: before they are allowed to become fixed, make them survive four external calendars that have nothing to do with your preferences.

The first two are other people’s holidays. A practitioner on a thread about top travel mistakes from 17 March 2023 gave the instruction as a precondition to any reservation: “before you make any flight reservations or hotel reservations: check when school vacations will be, holidays taking place in the cities you will be visiting as well as trade fairs. All of those can massively affect prices.” She added the part that catches people out — that school holidays are staggered rather than national, so Berlin can be on break while Munich is not, and that a national holiday can close a country down. The cleanest example is France, which splits its school holidays into three zones; the Ministry of Education publishes the school holiday calendar by academic year, and official public holidays sit on a separate government page verified on 1 January 2026. Two calendars to check, and neither of them is a travel site.

The third calendar is maintenance, and this is where the date rule gets its teeth. Venues close, and a closure is a date constraint rather than an itinerary problem. The Louvre is closed on Tuesdays, and its own page states that it stays open on other public holidays “unless they fall on a Tuesday, the museum’s day of closure.” A four-day Paris trip built across a Tuesday cannot contain the Louvre, whatever the rest of the plan says. The Cenacolo Vinciano in Milan closes Mondays, Christmas, New Year’s Eve and 1 May, per Milan’s visitor site. This is a date-selection input, not a scheduling detail, because finding it after the hotels are booked means one of the two things you most wanted is simply not available on the days you bought.

The fourth calendar belongs to transport and to release schedules, which is what most of the fifth section is about. The reason it belongs in date-setting is narrower: some date changes are free and some are not, and the ones that are not should be discovered before you commit. Under SNCF Connect’s own published terms, train tickets can be changed and refunded free of charge up to seven days before departure, after which a fee applies, and once you are within 30 minutes of departure only one change is allowed and no refund is possible after it.

One more force fixes dates legitimately, and it is not laziness. A conference, a wedding, a school term or a ferry that does not sail in February can pin the calendar before any of these checks are run. That is not a mistake; it just moves the constraint list earlier, because the dates are now an input to the plan rather than an output of it. What does create a problem is treating a preference as a constraint. A Saturday departure is not fixed until you have seen what leaving on Wednesday costs.

Check the documents before you defend the dates

Documents come next because they are the only gate that can cancel a trip outright rather than reschedule a purchase. A sold-out ticket changes what you do that afternoon. A passport that will not arrive changes whether you go.

The arithmetic is a subtraction with more terms than people expect. Those terms differ by issuing country, so the numbers below are the issuing authority’s own and apply to that country’s passports only. A British passport renewed online costs £102 and the renewal guide states plainly not to book travel until you have a valid passport. Its urgent routes are the useful part: 1 day premium is limited to renewing an adult passport and costs £239.50, with the earliest appointment two days after you apply and collection four hours after the appointment; the 1 week fast track costs £192 for an adult passport, with the earliest appointment the next day and courier delivery a week later. Those options exist, and they are priced for people whose lead time has already gone.

The second term is the validity condition, and it belongs to the destination rather than to you. Foreign governments set how much remaining validity a passport must have, and no amount of checking your own government’s page settles it, because this is a border rule belonging to the country you are entering. The consistent guidance from passport-issuing authorities is to check the destination’s own entry requirements rather than assume a single global figure, because the requirement is set country by country and some countries exempt some nationalities from their own general rule. Check yours on the destination’s own page.

The third term belongs to whichever country you are entering, and Schengen short-stay visas are a clean example because the European Commission publishes both ends of the window. It writes that an application must be lodged “at least 15 days before your intended journey and no earlier than 6 months beforehand,” and adds that “you may have to book an appointment before lodging the application” — which is the detail that has already sunk more trips than the window itself. Its visa-application page states that normal processing takes 15 days and can extend to 45 days if the application needs closer examination or extra documents, that the passport must remain valid at least three months after departure from the area, that medical insurance covering emergency care, hospitalisation and repatriation is a required document, and that the stay limit is 90 days in any 180-day period.

Read those two figures together and the schedule changes. Six months is the earliest you may apply. Fifteen days is the latest you may be allowed to, and it is not a promise — it is the floor under a process that can run to 45 days and that sits behind an appointment queue you do not control. Anyone on a nationality that needs a Schengen visa should treat the appointment as the real deadline and work backwards from it, not from the 15-day line.

One consequence is easier to miss than the rest, and it is about the order itself. Pressing an authority by buying a non-refundable trip is a bad trade: if entry permission is still open, keep the arrangements around it changeable until the decision lands. Refundable costs something. It costs less than a trip that cannot be used.

Documents also move, and this is why they come before the money. A passport number changes on renewal, which affects everything booked against it. A name on a ticket has to match the name on the passport exactly. And a document problem does not just delay a purchase — it can invalidate a destination you have already paid to reach, which is a category of error no refund policy covers.

Because those requirements depend on your nationality, your transit points, your destination and the date, confirm your own case with the issuing authority or the destination’s consulate before you pay for anything, and re-check it close to departure.

What remains once the passport is valid and the dates are pinned is a set of windows that open and close on other people’s schedules. Those have published dates, and the next section collects them.

The booking calendar nobody publishes for you

Every popular guide tells you to book early. Hardly any tells you when early starts, so a search that returns nothing looks like a sell-out when it usually means the sale has not opened.

That confusion is expensive in both directions. Travelling in the wrong season costs money, and buying too late costs the trip. The table below is a calendar of published openings, deadlines and closure days, drawn from the issuing operator’s or authority’s own page, with the date each was checked.

Three patterns fall out of that table, and they are worth more than the individual rows.

The first is that rail in this part of Europe sells in seasonal batches rather than on a rolling clock. Tickets went on sale on 11 March 2026 for travel between 4 July and 12 December 2026, according to SNCF Connect’s sales-opening page, which also lists a six-month horizon for services to Germany and a 60-day window for OUIGO Classic Train routes. The consequence for planning is direct: searching for a July train in January does not tell you the train is full. It tells you the sale has not opened, and the fix is to note the opening date rather than keep searching.

The second is that timed-entry venues release in batches too, on schedules that differ between venues in the same city. The Cenacolo Vinciano opens bookings on a three-month basis and then releases additional tickets every Wednesday at noon for the following week, which means a Wednesday lunchtime is a real appointment if that visit matters. The Cenacolo also enforces two rules that change what the day looks like: the ticket must bear the visitor’s full name, and visitors must be at the ticket counter 30 minutes before the slot or the reservation is lost. Separate terms apply elsewhere — the museum confirms that the Last Supper is not included in the city tourist pass, because of its entry quotas. Whatever pass you buy, the one visit it will not cover may be the one you came for.

The third is that airfare behaves unlike everything else in the plan, which is why it produces so much anxiety. There is no published date on which a route’s fares open or change — the only way to see a fare is to look for that route on those dates. Search results answering “how far in advance should you book a flight” disagree with each other by months, and they are mostly restating one or two underlying fare datasets rather than quoting the airline. So no window is published here, and none should be trusted from anyone claiming one. What replaces the window is a number you set yourself: decide the fare you are willing to pay for an itinerary you are willing to fly, watch the route, and buy when both are true.

One venue can also run several different windows at once, which is where most of the confusion about “when do tickets go on sale” actually comes from. The Eiffel Tower’s operator states on its own ticket page that tickets to the top by elevator “go on sale up to 90 days in advance,” at €36.70 for an adult. Its buying-tips page then sets out the rest of the ladder: the summit by stairs-then-elevator opens 60 days ahead and only in low season, the second floor by elevator 90 days, and the second floor by stairs 60 days. Four windows, one monument. Reading the 90-day figure and assuming it covers every ticket is how people arrive at a sold-out screen and conclude the day is gone.

The same page is blunt about who may sell you that ticket: “Only the Eiffel Tower sells single tickets for the ascent,” and any other site offering one “may be fraudulent,” with authorised resellers permitted to sell only bundles that add a service such as a guided tour. It also warns that fraudulent sites posing as official ones are being promoted through search engines. That is the issuer-versus-reseller distinction with money attached, and it generalises: the page that tells you a rule is rarely the page that issues it, and the trip planning websites that rank above a venue are not the venue. Worth knowing before you buy the fallback option, too — the operator states that second-floor tickets are non-changeable and non-refundable.

Not every deadline that matters is a release date. Some start the moment you pay, and one of them is a reasonable candidate for the most commonly missed deadline in trip planning.

Route it before you book a bed

Book the shape of the trip before you book a bed in it, because the route decides which bed is correct and the wrong bed is where a trip’s costs begin to compound.

The failure mode is specific. A non-refundable room booked before the route was settled cannot follow you to a different city, and a room in the wrong neighbourhood costs you the trip’s spare hours in taxi rides. Sellers set the terms, and the terms are not uniform. According to Booking.com’s guest FAQ, “every room or property has an individual policy determined by the property,” and the same page states without qualification that “changing your dates for non-refundable bookings isn’t possible.” There is no platform-level cancellation window to appeal to.

That is exactly what turned a mis-click into a loss in the forum case described earlier: the traveller booked the right city on the wrong dates and there was no mechanism to correct it. But the sequence that produced that $500 loss was not really about the hotel’s willingness. The booking was non-refundable and the dates were not yet stable, which is the combination to avoid. If the dates are still moving, the rate should be one that can move too.

Route first, then bed, then the fine-tuning — with one exception already noted in the second section, where an event makes a particular property the scarce item and the bed moves ahead of everything else.

Working in that order changes what a “route” even looks like. In practice it is three lists.

Start with the shape: the cities and the order, including whether you fly home from somewhere other than where you landed. An open-jaw or multi-city ticket that drops you in one city and collects you from another is usually cheaper than returning to the starting point and adding a train back to it, and it removes a whole backtracking day with it. The number of nights per stop decides more of the trip than the list of sights does — a late arrival and an early departure leave no usable day in between.

Then the bed, chosen against a map rather than a rating. Pin the arrival station or airport, the one thing you came to see, and the place you will want to eat, and the right neighbourhood tends to appear between them. This is also where the price is most easily misread: what matters in a hotel comparison is the total payable, with local taxes, breakfast and the transfer folded in, and the cancellation deadline sitting beside the quoted rate.

Then the return trip and the internal transport. The forum tactic of holding several refundable properties in different neighbourhoods and cancelling down to one is a legitimate way to keep the bed decision flexible while the route settles — provided the bookings are cancellable, which is the entire point of the tactic. Cancelling down is cheap. Cancelling out is not.

The route decision also fixes what the rest of the trip will cost to change. Two hotel bookings with free cancellation and a long deadline are almost free to correct. One non-refundable booking plus a separate-ticket flight between two cities is not, and neither is a connection built on separate tickets with no protection if the first leg slips. Choose how much of the trip you want to be able to un-take.

That ranking is where the planning actually gets prioritised — and it is the subject of the next section.

What you should leave unbooked (and what each booking costs to change)

Not every useful booking is worth making early. Some are better made late, and a plan with the option to change is worth more than a plan with everything fixed.

The argument for fixing everything in advance is that it saves money. The counter-argument is that it removes your ability to react. Practitioners on the forum put it as a warning about a different kind of cost: one poster described the failure as having “every hour of every day scheduled and prebooked – no room for serendipity, delays or when ‘stuff’ gets in the way,” and another listed the events that break a tight plan without warning — a train strike, a flight delay, an unexpected museum closure, a day of really foul weather. None of those are planning errors, and all of them are survivable if the plan has slack in it.

The workable balance is to sort every purchase by a single question: if this turns out to be wrong, what does fixing it cost? That puts the decisions roughly in these groups.

• Expensive and hard to reverse. Long-haul flights, non-refundable lodging, separate-ticket connections, and the deposits that start a clock. Everything here is worth getting right before you buy, and there is no hurry to buy it early.
• Cheap to reverse, worth securing. Refundable accommodation, changes that are free up to a stated date, tickets that can be re-timed. Booking these early is sensible precisely because the exit is free.
• Cheap and flexible, best left late. Local transport, restaurant reservations, hop-on transport, day tours, museum admissions on days that have no quota. Booking these early buys little and costs you the ability to change your mind about the day.
• A short-lived, valuable window. Anything with a genuine free-cancellation period. A hold is a cheap option, and an option you can exercise late is worth more than a small discount taken now.

Two clocks in that list deserve their own attention, because both expire while you are still making other choices.

The first is the insurance clock. Pre-existing-condition waivers on travel insurance are typically conditioned on time rather than on the trip, and the condition is usually triggered by your first non-refundable payment — sometimes described as a deposit, sometimes as an “initial trip payment.” The specific number belongs to the policy, and the number on a plan you are considering is the only one that matters to you; read it on that plan’s own terms, because waiver conditions and look-back periods differ between insurers and between plans. The planning consequence is simple and easy to act on: the moment you make the first payment that cannot be undone, whatever purchase that was, check the policy’s timing condition that day rather than a week later.

The second is the group trip. On a trip with several travellers, the earliest deposit made by anyone usually starts the clock for the policy, and one traveller’s document problem can invalidate a route that everyone else has already paid for. On a group trip, coordinate payments rather than making them independently, and settle the documents question before anybody’s non-refundable money moves.

What is left, once those are sorted, is the day of the trip itself — and this is the part where the widespread advice to pre-book everything does the most damage.

What should you leave unbooked?

A plan with no unbooked space in it has no capacity to absorb a bad day, and bad days do not announce themselves. The practitioners’ rule of thumb is to leave real slack in the schedule and to concentrate it where it can help: keep one day in four or five genuinely unscheduled, and put the unbooked time where the plan is most exposed, such as the day after a long transfer rather than a day in the middle of a settled city. Two heuristics have survived years of forum use. Book the few things that genuinely require it — a timed entry, a train that sells out, a table at the one restaurant that matters — and leave the rest. And if the trip is built around one specific anchor, protect the day after it, because the anchor is fixed and everything around it should be able to move.

The practical version of that is what to keep open. Transport within a city, meals that are not an occasion, museum admissions on days with no quota, and day trips with an hourly service all belong in the late group.

What genuinely needs booking early?

The bookings that deserve an early date are the ones with a published opening and a real quota: the timed entry released on a schedule, the train whose sale opens on a calendar date, the small property in a city during an event week, and any permit or ticket whose loss would change why the trip exists. Each of those has an issuer, and the issuer’s page is what to check before assuming an item is unavailable.

Everything else can be bought in the last week or in the last day, and in most destinations that is when a flexible plan is worth more than a fixed one. The bookings you actually need are the ones with a date and a limit attached. The rest is a preference, and preferences can be changed the morning they come up.

That logic runs backwards cleanly for anyone who has already booked out of order.

When the order has already gone wrong

Triage first: sort the damage by what is still reversible, who holds the change, and which rule applies — and only then start making calls.

The order matters because the recovery routes have very different costs, and two details decide which ones are open to you.

Start with the items that are still free to correct. Anything inside its cancellation window, anything refundable and anything changeable without a penalty can be fixed today by you alone, and it is the only part of the situation that is truly under your control. Clear that first, because it may change what the rest of the problems even are: moving one hotel can make a bad connection irrelevant, and shortening a leg can remove the need for an expensive ticket.

Then identify who holds the change on everything else, and ask them, not the platform. In the Verona case, the booking platform could only act with the hotel’s consent and the hotel declined, which is what the platform-level terms on cancellation already anticipated. The escalation order that generally works is the one the travellers in that thread described: request the change through the channel that produced the confirmation so a written record exists, then go directly to the property if the intermediary cannot obtain consent, and ask specifically about the parts of the charge that may be separable from the non-refundable room itself. City taxes and cleaning fees are often distinct from the room rate, and asking for them item by item is a different request from asking for a refund.

Then work out which rule actually governs, because a claim made under the wrong rule wastes the goodwill you will need for the one that fits. The operator’s own terms govern what that operator must do. A statutory refund right may or may not reach your booking, depending on the ticket and the route. The distinction is easiest to see in a single US rule, which applies to flights to, from and within the United States. According to the U.S. Department of Transportation, a passenger is entitled to a refund when the airline cancels a flight or significantly delays or changes it and the passenger chooses not to accept the alternative. The rule is specific about what “significant” means — a departure three hours or more early for a domestic itinerary or six hours or more early for an international one, an arrival delayed by the same margins, a change of origin or destination airport, more connections than originally ticketed, or an involuntary downgrade in class. Where the conditions are met and the passenger rejects the offer, the refund must be issued automatically, within seven business days for a credit-card purchase or 20 calendar days for other payment methods. And a passenger cannot be pushed onto a voucher instead: any airline credit must be valid for at least five years, with its restrictions disclosed, and a refund remains available.

Now read the limits, because they are the reason this is a section rather than a paragraph. The same page states that the rule does not cover a passenger who chooses to travel on the changed flight, and — the line that matters most here — that a separately ticketed leg operated by another carrier is not covered by the first carrier’s obligation. That is the Italian traveller’s case again, and it is why his remedy was limited to what the cancelling carrier chose to offer. A right that exists is not the same as a right that reaches your booking. And one more line worth knowing before you telephone: the 24-hour rule, which requires an airline either to allow a penalty-free cancellation within 24 hours of booking or to hold the reservation without payment, applies to tickets bought from the airline itself — it does not apply to tickets booked through an online travel agency or another third-party agent. If you booked through an intermediary, the place to start is the intermediary.

Two things are worth doing after a recovery succeeds. Put a note of what happened next to the booking itself, and keep the correspondence where you will find it — the travellers in that thread noted that a modified reservation can lose the earlier correspondence trail. And expect the outcome to depend on the operator’s discretion where the terms permit it, not on a rule that obliges them: the same thread contains cases where a property moved a non-refundable booking at no extra cost, and cases where it point-blank refused. Goodwill is real, and it is not a right, which is why the earlier sections put the price of changing a booking above the hope of being forgiven one.

The master planning checklist and deadline timeline

Everything above reduces to one page you own. The point of the table is the last two columns, because a checklist that cannot tell you when a decision expires cannot help you order the decisions.

A long-running thread on the Rick Steves travel forum from 3 November 2023 shows what an experienced planner actually writes down. One contributor’s itinerary records lodging with a “cancel by” date beside it, “so I don’t have to go back and fish thru emails if my plans change.” The same person puts the verification habit next to it: “I do not trust ANY guide book on museum, cemetery or church hours especially since Covid… I always check the official website for any place that is important for me to visit.” A second contributor keeps six sheets — identification, lodging, daily itinerary, transport, costs, packing — with transport carrying whether an item needed booking ahead. It is a better trip planning checklist than most published ones, for a single reason. The thread is full of ways to store this. Almost none of them store the deadline.

That is the gap this table closes. Print it, or copy the columns into whatever travel plan template you already keep.

Three notes on using it. Where a cell names an issuer rather than a number, the number belongs to that issuer and changes without notice, so it belongs in the “where to re-check” column rather than in a fixed figure. Where your own passport, nationality or destination differs from the examples above, the whole row changes, because every deadline in it belongs to a specific country. And the columns are a method, not a template you fill once: the same three questions run the calendar, the route, the bed and the tickets.

A planning method built in public can look interchangeable, so here is what this one actually is: it is a practical structure built for this guide from the three constraint types described above, the reversibility ranking, and the published dates of the operators and authorities cited in it — not an industry standard, and not a replacement for the issuers’ own pages. The deadlines are theirs. The ordering is what this guide adds.

One column in that table is doing more work than the others, and it is the one to keep open until the last week before departure.

O Erro Que Mais Custa Caro

The mistake that costs the most is rarely a wrong choice; it is a right choice made in the wrong order, and its price is set by what it forecloses rather than what it costs. The document problem that ends a trip is the clearest version of it, so the most useful habit to carry away from all of this is to check your own case on the issuing authority’s page before you pay for anything — the destination’s consulate or immigration service for entry requirements, and IATA’s Travel Centre if you want a single passport, visa and health check built from your own nationality and itinerary. Run that check while your bookings can still be changed, not after they cannot, because the point of this ordering is that everything behind a document is refundable and nothing behind it is recoverable.

Frequently asked questions

What should I book first, flights or accommodation?

Neither, until the test decides. Book whichever item has the nearest published deadline and the thinnest supply — usually the flight on a common route, but the bed first during a festival or in a small city. Hold refundable accommodation early and settle it later.

Should I cancel a flight or take the travel credit when my plans change?

Take the refund if the airline cancelled or significantly changed the flight and you do not want the alternative. Under U.S. DOT rules that refund is automatic, within seven business days for a card purchase, and a voucher needs your consent.