When to book flights: 95 fares show prices jump inside three weeks

Book about three to six weeks before you fly, and don’t leave it until the last three weeks. In Google Flights’ daily price histories for 95 round trips on 16 routes, the median fare sat about 7% to 10% below its two-month price from 42 to 21 days before departure. Then it turned: back to even at 19 days, 14.1% higher at 14 days and 28.8% higher at 12 days. Booking four to five months ahead wasn’t cheaper either.

That pattern was remarkably consistent. Of the 47 trips we followed into the final two weeks, 46 cost more 12 to 14 days before departure than they had at their cheapest point three to five weeks out. None of the 47 got cheaper over the course of its history.

How we tracked 95 fares

Google Flights keeps a daily price history for a trip: the lowest price it found for those dates on each day. On November 3, 2026, we collected 61 days of that history for 95 round trips. Every trip was a seven-night stay, in economy, for one adult, priced in US dollars, on one of 16 routes ranging from Atlanta to Orlando to New York to Paris.

The trips fell into two groups, because a 61-day history can only show one stretch of a trip’s life:

• November trips (47 trips) departing November 12, 14 and 17, 2026, which was 9 to 14 days after we collected. Their histories run from about 74 days before departure to 9 to 14 days before. This group shows what happens in the final weeks.
• January trips (48 trips) departing January 13, 20 and 27, 2027, which was 71 to 85 days away. Their histories run from about 145 days before departure to about 71 days before. This group shows whether booking four to five months out pays off.

Trips on different routes cost very different amounts, so we measured each trip against itself: its price on a given day compared with its own price 60 days before departure. Then we took the median across trips, the middle value, so one wild swing on one route can’t drag the result around.

Fares dipped from six weeks out, then jumped inside three weeks

For the first few weeks, very little happened. At 56 and 49 days before departure, the median trip was within half a percent of its 60-day price.

Around six weeks out, prices eased. The median was 6.9% below the 60-day price at 42 days, 7.5% below at 35 days, 7.6% below at 28 days and 9.8% below at both 24 and 21 days. Looked at day by day, the median stayed below the 60-day price on every single day from 39 to 20 days before departure, usually by 5% to 11% (the smallest dip was 1.8%, at 27 days). That is the cheapest stretch in our data.

Then the floor gave way. At 19 days out the median was back to 0.5% below the 60-day price, at 17 days it was 8.4% above, at 14 days 14.1% above and at 12 days 28.8% above. From 21 days to 12 days, the median moved from 9.8% below to 28.8% above, a swing of about 39 percentage points (9.8 + 28.8 = 38.6) in nine days. At 10 days it was 23.6% above, but only the 16 trips departing November 12 had history that close to departure, so treat that point, and the 12-day point (32 trips), as less certain than the rest.

We also found the single cheapest day in each trip’s history. For the November trips, the median cheapest day was 38 days before departure, with a range from 16 to 73 days.

Not every trip moved in lockstep. At 21 days out, the middle half of trips sat anywhere between 22.8% below and 2.1% above their 60-day price, which means at least a quarter of trips were already pricier than they had been two months earlier. By 12 days out, the middle half sat between 11.2% and 54.9% above the 60-day price, so at least three in four trips cost at least 11.2% more than they had two months earlier.

Across each November trip’s whole history, from about 70 days out to the last day we have (9 to 14 days out), the median trip rose 32.3% and the average trip 44.5%. Forty-six trips went up, one stayed flat and none went down.

What the jump looks like in dollars

In dollars, the median round trip in the November group cost $259 at 60 days out, $240 at 35 days, $250 at 28 days, $213 at 21 days, $260 at 14 days and $339 at 12 days (32 trips). These are medians of the dollar prices, a separate calculation from the percentage changes above, which is why, for example, the dollar median ticks up at 28 days while the percentage median barely moves.

To see what the percentages mean for a single booking, take a round trip that costs $259 two months out and apply the median changes:

On that trip, waiting one more week, from three weeks out to two, would cost about $62 per traveler ($295.52 − $233.62). It’s an illustration built from medians, not a forecast for any particular flight, but it shows why the last three weeks are the expensive ones.

The jump wasn’t a Thanksgiving effect

All of the November trips left in mid-November, so it’s fair to ask whether Thanksgiving, on November 26, 2026, was behind the rise. The November 17 trips come home on November 24, two days before the holiday, so some holiday demand could be in their prices.

But every one of the three departure dates showed the same jump, including the November 12 trips, which return on November 19, a full week before Thanksgiving. Here is what the median November 12 trip did:

These trips kept rising right up to the last day we could see.

So why do fares climb in the last three weeks?

Our data shows when prices moved, not why. A likely part of the answer is the advance-purchase deadlines that many fares carry: some lower fares can only be bought a certain number of days before departure, and as those deadlines pass, the cheaper options drop out of the search results. The timing in our data fits that idea, with prices turning at about 19 days and climbing from there, but we didn’t look at fare rules, so we can’t confirm it.

One more signal from the same searches: on the day we collected, Google Flights rated the current price of these 95 trips “typical” for 57 of them, “high” for 34 and “low” for just 4. A “low” label was rare, so don’t count on seeing one before you buy.

Is it cheaper to book flights months in advance?

Not on these trips. For the 48 January trips, we compared the first day of each price history (about 131 to 145 days before departure) with the last (about 71 to 85 days before). The median trip got 9.2% cheaper over that stretch. The average trip got 4.2% cheaper, a smaller fall because some trips rose. In all, 26 of the 48 trips got cheaper and 22 got more expensive.

So for these late-January trips, booking four to five months ahead was not cheaper than booking about two and a half months ahead, and the typical trip was a little cheaper later. At the level of a single trip, though, it was close to a coin flip, and the cheapest day could fall almost anywhere: the median cheapest day for the January trips was 104.5 days before departure, with a range from 71 to 144 days.

Put the two groups side by side and the shape is clear, even though they are different trips and can’t be joined into one curve. Waiting from five months to two and a half months out saved the typical trip 9.2%. Waiting from ten weeks to the final two weeks cost the typical trip about a third more (a median rise of 32.3%).

How this compares with Google’s own booking advice

Our numbers line up closely with Google’s published analysis. In October 2026, Google said that, based on five years of aggregated Google Flights data, prices for U.S. domestic flights hit their lowest 39 days before departure, with a low-price range of 22 to 57 days. Our November trips had their median cheapest day at 38 days, and their cheap stretch ran from about 42 to 20 days out.

Google’s figures for other kinds of trips are a reminder that one window doesn’t fit every flight. It put the lowest international prices at 89 days before departure, with a low range of 50 to 133 days (though it called the average price drop across that range insignificant), and the lowest Christmas fares at 56 days before departure (33 to 67 days). For trips to Mexico or the Caribbean, it gave 43 days (24 to 69 days). Eleven of our 16 routes are domestic, so our results speak most strongly to domestic trips. For holiday dates, see how to afford Christmas and New Year’s flights.

The 16 routes in this study

Every route was priced as a seven-night round trip in economy for one adult.

We aren’t publishing results route by route. The November group has 47 trips across 16 routes, about three per route (47 ÷ 16 ≈ 2.9), which is too few for a single route’s median to mean much. The pattern above is the pattern across all of them.

How far in advance should you book flights?

Here is the whole study as a booking timeline, using the medians above:

Turned into a playbook:

1. Start watching about two months out. Price the trip around 60 days before departure so you know what normal looks like. Booking much earlier than two and a half months didn’t save money on our January trips.
2. Aim to buy between six and three weeks out. That is where the median trip was cheapest, and where the typical single cheapest day fell (38 days out).
3. Treat three weeks as a deadline, not a target. The discount was gone by 19 days and the median was 14.1% higher at two weeks. Booking at 21 days means booking at the very end of the cheap stretch, with no room for a delay.
4. Don’t hold out for a perfect dip. At 21 days, at least a quarter of trips were already pricier than two months earlier. If a fare in the window fits your budget, waiting for it to drop further risks running into the climb that started at about 19 days.
5. Be flexible on the day you fly. Timing your purchase is only half of it. In our cheapest day to fly study, Tuesday and Wednesday departures averaged 7.5% below their week’s average fare, while Sunday averaged 13.4% above.
6. Check what the fare includes. The cheapest fare is often the most restricted one, and the extras can cost more than timing saves. Our basic economy study found that moving up from basic economy cost a median $45.
7. Go earlier for holidays and long-haul trips. Our data covers off-peak economy trips, mostly domestic. Google’s own analysis puts the low point for international and Christmas trips further out.

What this data can’t tell you

This study is a careful look at a narrow slice of the market, and its limits matter:

• Two snapshots. We followed trips departing in mid-November 2026 and late January 2027. Summer, spring break and peak holiday dates can behave differently.
• Economy, one adult, seven-night round trips, 16 routes. Premium cabins, one-way trips, other trip lengths and other routes weren’t tested.
• The final days are missing. Our November histories end 9 to 14 days before departure, so we can’t say what happened in the last week or two.
• Two different sets of trips. The November and January groups are different trips on different dates, so they can’t be stitched into one curve from five months out to departure day.
• The lowest price Google showed, not every seller. The history tracks the lowest price for the trip each day, which may not be the same flight from one day to the next. Other sellers may have had other prices.
• Timing, not cause. We saw when fares rose, not why.

Booking in the cheap window with a payment plan

If the fare you find in the cheap window, roughly 42 to 21 days before departure, is more than you want to pay in one go, a payment plan is one way to book it. With Defera, you book now and pay over time in weekly or bi-weekly payments, with no credit check and no interest. Once your booking is approved, your agreed-upon trip price is secured: if fares rise after that, as they did in the final three weeks for 46 of our 47 November trips, your price doesn’t, as long as you stay in good standing on your plan. How your trip price is protected explains the details.

Two things to know. First, every Defera trip departs at least 22 days after booking, so the very end of the cheap window can’t be booked on Defera at all (how far ahead you need to book). Second, the deposit depends on how far away your trip is. Deposits start as low as 5% on trips more than 90 days away, 20% at 60 to 89 days, 40% at 30 to 59 days and 60% at 22 to 29 days. So the cheapest fares in this data and the smallest deposits don’t fall in the same weeks. If a lower upfront payment matters more to you than the dip, which was 6.9% to 9.8% on the median trip, booking 60 or more days out, where deposits start at 20%, can be the better fit for your budget. Choosing your deposit walks through the options, and you can search flights to see the plan for your dates.