Strawberries run on a price calendar. It has not missed once in thirty-five years.

SECTION 01The claim

A pound of strawberries is not one price with some noise on it. It is a steep, repeating annual cycle that has run in the same direction, in the same months, for as long as anyone has been writing it down.

Here is the whole finding, computed from the United States producer price index for strawberries, monthly, 1990 through 2025.[2]

• Compare the winter quarter to the summer quarter, year by year. Winter came in dearer in thirty-five years out of thirty-five. There is no exception in the record. The median gap is 1.73 times, the narrowest year 1.08, the widest 2.35.
• Take the single dearest month against the single cheapest instead, in the twenty-three years with a complete twelve-month record, and the gap runs a median of 3.28 times. It was never once smaller than double. The narrowest of those years still came in at 2.08.
• The peak landed in November, December or January in twenty-one of those twenty-three years. The trough landed in May, June or July in sixteen.
• Comparing the two anchor months directly, December ran dearer than July in thirty-one of thirty-two years.

A dessert shop buying the same fruit in December and in July is running two different businesses and only has numbers for one of them.

None of this is secret and none of it is expensive. It sat in two files that cost nothing, and the work took an afternoon.

SECTION 02Two files, no money

The first is the National Oceanic and Atmospheric Administration's daily weather record for Plant City, Florida, station USC00087205.[1] Plant City sits in the middle of the Florida strawberry belt. The station has daily minimum temperature going back to 1892. We pulled 46,448 daily records.

The second is the Bureau of Labor Statistics producer price index for strawberries, series WPU01110222.[2] Monthly, back to January 1947, 975 observations, published every month, free.

We mention both because this is the part that gets sold to small businesses as impossible. It is not impossible. It is two downloads and an afternoon of care.

SECTION 03The shape of the year

Here is the median index for each month across 1990 to 2025, with July set at 1.00 so the ratios are readable.

The year splits cleanly in half. October through March sit above the all-month median. April through September sit at or below it. There is no ambiguous middle.

The reason for the shape is agronomic and it is published. Florida plants from late September into early November and fruits from November into April or May.[4] Winter strawberries in the United States are largely a Florida and Mexico story, at a time of year when California is not picking. Less fruit, harder to grow, in the months when people want it most.

SECTION 04Can you see the bad years coming?

If the cycle is that reliable, the obvious next question is whether the exceptions are predictable too. A hard freeze in Hillsborough County is a news event. Does it move the national price?

We built a cold measure for every season from 1990 to 2026: the coldest night between December 1 and the end of February, the count of nights at or below 32 Fahrenheit, and the count at or below 28. Then we stripped the seasonal cycle out of the price index and compared.

Thirty-six seasons. Read the third row against the first, because that is where the interesting thing is.

How many times it froze matters. How cold the worst night got does not.

That inverts the way almost everyone talks about it. The story is always the record low, the one terrible night. But Florida fruits in cycles, and each frost event costs a picking cycle whether the thermometer reads 30 or 22. Four ordinary frosts take more fruit out of the market than one spectacular one. The same finding is in the disease work: the university's account of the outbreaks that destroyed whole fields in 2018-19 and 2019-20 points at consecutive and prolonged rain, not at one storm.[5]

A correlation of +0.35 across thirty-six seasons is a real relationship and a weak one. It is worth about as much as knowing a coin is slightly bent. Which brings us to the year it failed completely.

SECTION 05Where it broke

The winter of 2026 was the worst documented freeze in this record. Plant City had five nights at or below freezing and two at or below 28, bottoming at 25 Fahrenheit on February 1.[1] Florida's Department of Agriculture put strawberry losses at $306,965,897, part of an estimated $3.1 billion across the state's agriculture.[3] The federal disaster declaration that followed was reported as roughly 80 percent of the remaining harvest.[8]

If the simple story were true, the price should have gone vertical. Here is what it actually did, against each month's own 1990 to 2025 median.

February, the month the freeze happened, came in 19.4 percent above its median. Then March, with most of the crop gone, came in 23 percent below normal.

A third of a billion dollars of fruit was destroyed and the national price in the following month went down.

Fruit destroyed in a field never becomes supply that has to clear. The freeze took the sellers and the buyers out together, and a market with nothing to sell and nobody expecting to buy does not bid. The thing everyone was sure would happen did not happen, and it did not happen for a reason you can only see by looking.

SECTION 06What it is worth to a shop

Take an invented dessert shop. It is not a real business and these are not a real business's invoices. It uses 40 pounds of strawberries a week, 2,080 pounds a year, and it pays $1.80 a pound in July, its cheapest month.

Apply the shape of the year and nothing else. Buying the same amount every month, it spends about $5,754 a year on strawberries. Now move the volume: feature the strawberry items hard from April through September, rotate to something else from October through March, so that 70 percent of the pounds land in the cheap half.

Eleven percent of a line item, for changing which weeks a dish is featured. The fruit is the same fruit. The recipe is the same recipe.

The version that lands harder is the single week. On this invented shop's numbers, a 40 pound week costs about $72 in July and about $205 in December: $133 more for the identical box. Most owners feel that as a vague sense that winter is expensive. Very few have ever seen it as a number with a date attached.

SECTION 07What this does not say

The index is the price at the farm. It is not the price on an invoice. Packing, cooling, freight and a distributor's margin sit in between, and none of them swing 2.85 times. Government figures put the farm share of a restaurant food dollar at 7.1 cents, against 18.5 cents for food bought to cook at home.[6]

So the honest version of the claim is this. The shape of the year transmits to what a shop pays. The full size of the swing does not. A buyer who wants their own number has a better source than an index: the Agriculture Department publishes shipping-point and terminal-market strawberry prices by origin, daily, in dollars per flat of eight one-pound containers.[7] That is much closer to a real invoice, and it is also free.

SECTION 08The point

None of the work here needed a single number from inside a business. Two public files, one of them older than the business anyone reading this owns, and the annual cost shape of a menu item falls out of them.

Every business has a version of this. A price, a season, a weather pattern, a school calendar, a fuel index, a tourism cycle. Something outside the four walls that moves the numbers inside them, on a schedule, reliably, and that nobody has ever plotted because plotting it was somebody's whole week and nobody had the week.

That is most of what a data team does. Not clever forecasting. Joining the thing you already have to the thing that was already public, and then being honest about the years it does not work.