How to Find Cheap Flights (A Step by Step Method That Actually Works)

Shopping for a flight on a Wednesday saves about 1.3%. Flying on one saves roughly 13% against a weekend departure, and up to 20% on domestic routes, according to Thrifty Traveler's analysis of Google Flights' own price-tracking data. That gap between the two numbers is the whole method. Almost everything travellers are told to do at the moment of searching is worth about a percentage point. Almost everything that moves the price is a decision about the trip itself: which day it starts, which airport it leaves from, whether it tolerates a connection, and how many weeks before departure the card gets charged.

This page is for someone who already has dates, a route in mind and a ceiling number, and wants the order in which to do things. The conclusion, stated up front: pick the booking window from the destination and season, read a whole month of prices rather than one date, test two airports and one connection, and then price the total rather than the fare. The single most expensive mistake is the last one. On a budget carrier, the advertised fare is a component, not a price. Over half of Ryanair's 2024 revenue came from ancillary fees rather than base tickets, per Kiwi.com, which is a polite way of saying the headline number was never meant to be what anybody pays.

Airlines do not price a seat. They price a bucket of seats, sell it, and move to the next one. Thrifty Traveler puts fare-class depletion behind 90% or more of the price changes a shopper observes, which explains the two things that confuse people most: why a fare jumps overnight, and why it jumps for everybody at once rather than for the person who searched twice.

This matters because it rules out the most repeated trick in the genre. Incognito mode deletes browsing history, cookies and temporary files on one device. It does not touch the airline's inventory. Kiwi.com's phrasing is exact: it "does not change pricing or availability as it does not alter airline inventory, prices, or how sites present offers," and it does not stop server-side identification by IP address or device fingerprint either. Thrifty Traveler is blunter: "Searching for flights incognito does nothing. Absolutely nothing. Nor does clearing your cookies."

The myth has an unusually good alibi, which deserves stating plainly. On 18 April 2026 a JetBlue customer reported a fare rising $230 in a single day while booking a flight for a funeral, and JetBlue's customer service replied, in a message later deleted, advising the customer to clear cache and cookies. Tech Times reported that the episode helped trigger letters from Rep. Frank Pallone Jr. to eight US carriers on 11 and 12 August 2026, with a 25 August deadline, asking how AI systems and personal data shape ticket prices and whether browsing history, income or location feed into them. One study cited in that inquiry found an airline lifting revenue by up to 6% through AI pricing built on consumer data. Thrifty Traveler's position is that individual-level personalised pricing is not happening yet in 2026. Both things can be true: the mechanism is under investigation, no findings are public, and clearing cookies still does nothing about it.

What the traveller controls is narrower and duller. Timing inside a booking window, day of departure, airport, connection, and the add-ons. The sources disagree about the first of those, and the disagreement is worth reading closely rather than averaging away.

The columns are not measuring the same thing, which is why they look contradictory. Expedia is reporting where the lowest observed prices sat. Going.com and NerdWallet are recommending when to buy. Those diverge sharply at the short end: Expedia's cheapest international window of 8 to 15 days out describes fares that existed, not a strategy anyone can execute, since the same shopper arriving at day 12 to find nothing under the ceiling has no move left. The safer reading is to use the advice windows as the outer boundary and the Google Flights averages as the target: for Europe, that means starting to watch three to four months out and buying near the 94-day mark rather than hoping.

Before opening a search engine, five things need to be settled. Searching without them produces hours of browsing and no decision.

1. Set the booking window from the destination and the season, not from a general rule. Domestic US: 1 to 3 months ahead, targeting the 21 to 52 day zone. Europe or another long-haul: Going.com's range of 2 to 8 months, with the Google Flights average sitting at 94 days. Peak season pushes both out: 3 to 5 months domestic, 4 to 10 months international. Takes five minutes and produces a two-week span on the calendar during which to buy.

2. Search the whole month before searching the date. A month grid or flexible-date view shows the spread across the period, which is the only number that tells the traveller whether flexibility is worth anything on this route. If the cheapest and the most expensive day in the month differ by 15%, shifting dates is worth the effort. If they differ by 3%, it is not, and the search can move on to airports.

3. Test the day of the week on the specific route, not the national ranking. The published rankings conflict. Expedia's 2026 Air Hacks report puts Friday first and attributes the change to fewer business travellers flying into the weekend; Going.com lists Tuesday, Wednesday and Saturday; Thrifty Traveler's data favours Monday to Wednesday. They agree only that Sunday and Monday are the worst domestic days to fly. The grid from step 2 settles it for one route in about thirty seconds, which is faster than reading any of the rankings.

4. Add the second and third airport. One search per airport, same date range. The cost of a coach or train to the alternative airport goes on the total before comparing, or the comparison is fictional.

5. Price one connecting itinerary against the nonstop. Nonstop flights run about 25% more than connecting ones on average, per the Google Flights data. On a $600 long-haul that is roughly $150 for the convenience, which is a real thing to buy and a real thing to decline; on a two-hour hop it is rarely worth the extra half day.

6. Set a price alert on the winning combination and keep the ceiling in mind. Kiwi.com recommends alerts alongside date and airport flexibility, and the mechanism from section one explains how to treat them. If roughly 90% of price movements are a cheap bucket emptying, then an alert reporting a drop is reporting that a bucket has opened, and the bucket will close when it sells. Waiting a week to think about it is a decision to lose it.

7. Compare across several search engines, then check the airline's own site. Kiwi.com's guidance, from an online travel agency with no obvious incentive to say so, is to compare multiple engines before buying direct. The airline site is where fare bundles and bag prices are visible in full, which matters for the next step.

8. Price the total. Fare, plus checked bag if needed, plus seat if seats matter, plus any card fee. On Ryanair this is done from the published Table of Fees, and it changes which airline wins more often than most people expect.

9. Buy, and in the US use the 24-hour rule as free optionality. A US flight booked at least seven days before departure qualifies for a 24-hour free cancellation window under DOT rules, per Going.com. That converts a purchase into a one-day hold: buy at the ceiling, keep looking for a day, cancel if something better appears. Points and award tickets are the exception and should be booked the moment availability appears, since the inventory is fixed and small.

10. Buy the extras at booking, not later. This is where the money is, and it is covered in full below.

11. Check in the moment check-in opens. On Ryanair, paid seat selection opens 60 days before the flight and, per MoneySavingExpert, a seat price locked in at that point does not rise as departure nears. Passengers who pay for nothing are still guaranteed a free seat, assigned automatically between 24 and 2 hours before departure. Skipping online check-in entirely is the expensive version: Ryanair's airport check-in charge was listed at £55 in an August 2026 fee roundup.

Waiting for the last-minute low that the averages describe. Expedia's finding that international economy fares were cheapest 8 to 15 days out is a description of where the bottom sat, not an instruction. On the other side of the same coin, Kiwi.com reports budget carrier fares rising 200% to 300% between early booking and near departure. The recognisable symptom is a shopper at day 20 watching the number climb with no fallback date. The fix is structural: pick the buy-by date at step one and treat it as a deadline rather than a suggestion.

Treating browser hygiene as a lever. The tell is a traveller who has spent twenty minutes on private windows and VPN settings and has not yet looked at the airport 60 miles away. Thrifty Traveler's argument is the one that settles it: professional deal hunters search constantly and still find cheap fares, which they could not do if repeated searching raised prices.

Adding the bag afterwards. A 20kg checked bag on Ryanair was listed at about £20 to £25 when bought with the ticket and over £40 when added later or at the airport, per the same August 2026 roundup and MoneySavingExpert. Priority boarding with a second cabin bag ran £2 in advance against £4 at the airport on MoneySavingExpert's figures, and the fuller roundup put the later-addition range at £20 to £60, with the option unavailable at the gate. Excess weight is charged at £13 per kilogram at the airport against a 32kg maximum, which turns a slightly heavy bag into a fare of its own.

Booking under a name that does not match the passport. Ryanair's name change or correction fee was listed at £115 online and £160 through the contact centre or airport in August 2026. It is the most avoidable large number in this article and it is caused entirely by typing quickly.

Buying nonstop by reflex. A 25% average premium is defensible on a long-haul with a tight connection and indefensible as a default.

Comparing an unbundled fare with a bundled one on the base price. Ryanair's Regular fare includes a standard seat, priority boarding and two cabin bags; the Plus fare adds a 20kg checked bag. Where a traveller needs those things anyway, the bundle can undercut the Basic fare plus the same items bought separately. It can also not. The only way to know is to price both, and bundle pricing is dynamic, so the answer changes by route and date.

Peak summer in Europe and other high-season long-haul. Going.com's peak international window is 4 to 10 months out, against 2 to 8 in normal season. The same source puts European shoulder season fares, April to May and September to October, at 20% to 40% below peak summer rates, which is a larger saving than any booking-timing tactic in this article. Moving the trip by three weeks beats optimising the purchase of a July ticket.

Thanksgiving. NerdWallet reports the lowest prices at about 35 days before departure with a sweet spot of 24 to 59 days, and advises starting the search in early October. The Google Flights averages put the low at about 45 days with a 26 to 59 day range and an optimal search window of 1 October to 2 November. The overlap, roughly late October, is where both agree.

Christmas. Lowest prices around 51 days out, sweet spot 32 to 73 days, per NerdWallet; about 58 days with a 36 to 72 day range from the Google Flights data, and a search window of 14 October to 19 November. Flying on the holiday itself, Thanksgiving Day or Christmas Day, is substantially cheaper than the days around it.

Mexico and the Caribbean. Lowest average around 44 days out, range 26 to 68 days.

Summer domestic travel in the US. An outlier worth knowing about: the Google Flights average low for summer domestic trips falls at just 21 days out, with a range of 13 to 44 days, far shorter than the general domestic pattern.

Point of sale. Distinct from the incognito myth and frequently confused with it. Thrifty Traveler reports a real case of a flight priced at $46 when bought as though from Colombia against $60 from the United States, achieved by changing the apparent point of sale. The saving is genuine; it is also a currency and residency question with implications for payment and refunds that the sources here do not cover.

Travelling with children. Following a UK Competition and Markets Authority review, Ryanair's random free-seat allocation was changed as of June 2026 to guarantee children under 12 a seat next to an accompanying adult at no charge. Families on that carrier no longer need to buy seats to sit together, which removes a line from the arithmetic below.

Cherry blossom season in Japan, festivals, and any other fixed-date surge. The published data here does not cover them. The defensible move is to treat them as peak international and work from the 4 to 10 month window, then verify against the month grid, which shows the surge on the specific route whether or not anyone has published a study about it.

The clearest way to see what timing is worth is to price the same trip twice. Take one passenger on a Ryanair Basic fare who needs a 20kg checked bag, wants a reserved standard seat and wants priority boarding with a second cabin bag, and price the extras at booking against the same extras added late or at the airport, using the low end of each published range.

Bought with the ticket: about £20 for the 20kg bag, £5 for the standard seat and £2 for priority, or roughly £27. The same three items left until later: over £40 for the bag, £7.50 for a seat bought at the airport and from £20 for priority added after booking, or roughly £67.50. Identical flight, identical passenger, identical luggage, about £40 apart, before the fare itself is counted. Add Kiwi.com's observation that budget fares can run 200% to 300% higher near departure than early on, and the two versions of the same trip are not in the same price bracket. The seat figures come from MoneySavingExpert as of June 2026 and an August 2026 fee roundup; Ryanair itself publishes only the categories and confirms in its seat policy that prices move with route, season and time to departure, so these are floors rather than quotes.

What is worth buying, on these figures: the checked bag at booking if it is needed at all, since the gap is roughly double. What is worth skipping: the seat, in most cases, because a free one is assigned at check-in between 24 and 2 hours before departure and, since June 2026, under-12s are seated with an adult at no charge. Extra-legroom rows were listed from about £8 to £14 and up, front rows from about £7.50, standard seats from about £3 to £5, all rising with demand. Fast Track security ran roughly £5 to £12 depending on the airport. None of those change whether the plane arrives.

The one add-on that is never worth its price is the one bought by accident: £55 for airport check-in, £13 per kilogram for excess weight, £45 online or £60 by phone for a date change. All three are charges for a decision not made in advance. The cheapest baggage strategy remains the oldest one, which is to fit the trip inside the 40 x 20 x 30 cm bag the Basic fare already includes.

The strongest objection to everything above is that it assumes fares are set by inventory rather than by the shopper, and that assumption is currently under federal examination. Eight US carriers, including American, Delta, United, Alaska, JetBlue, Southwest, Frontier and Hawaiian, received information requests in August 2026 asking how often AI systems change fares and whether income, browsing history or location affect what a customer sees. The FTC has had a surveillance pricing investigation running since July 2024, with preliminary findings released in January 2025. Four states had enacted surveillance-pricing bans by mid-2026, none of them explicitly covering airlines. Separately, Rep. Shontel Brown and Sen. Ron Wyden asked the GAO in August 2026 to investigate why the Department of Transportation, holder of exclusive authority over airline passenger data privacy for more than 40 years, has never brought a single enforcement action. Their letter notes that a broker co-owned by major airlines sold access to about 722 million passenger travel records to federal agencies without warrants.

So the industry demonstrably holds the data. What it has not been shown to do, in any published finding as of the material here, is price individual travellers differently on the basis of it, and no amount of cookie-clearing would counter it if it did, since incognito mode leaves IP address and device fingerprint intact. The practical position is unchanged: the levers that have measurable effects are dates, airports, connections, timing and add-ons, and a reader who works through those in order has captured everything the evidence currently supports capturing.

A shopper who does all of this well on a July departure to Europe will still pay more than one who does none of it in late September, because a 20% to 40% seasonal spread swamps an 8% weekday spread and a 25% connection premium. Which is the real conclusion hiding in the figures: the cheapest flight is usually found before the search begins, in the choice of when to go.