Halloween demand on Amazon is already doubling — count your inbound weeks backward

Halloween is October 31. Halloween demand on Amazon is an August phenomenon, and this year’s curve is already steep.

We pulled the US search data on August 23. In the three weeks to August 15, weekly exact search volume for “halloween decorations” went from 293,526 to 570,445 — a 94% jump. “Halloween costumes” grew 74% (29,543 → 51,262). “Halloween candy” grew fastest of all: 15,728 → 54,842, up 249%.

If your Halloween inventory plan assumed September was for shipping and October was for selling, the buyers disagree. Here is what the data says, and the calendar math that follows from it.

The curve is early, and it is not uniform

Three head terms, three different shapes.

Decorations demand is the largest and compounding weekly — the classic early-decorator pattern, where the front porch is done weeks before the candy bowl. Costume demand is climbing steadily, and the apparel side skews even earlier: “halloween costumes for women” already shows August 2026 as its best sales month in the current data. Candy demand is smaller in absolute terms but accelerating hardest, growing from a low base as consumables buyers start stocking.

The operational point: “seasonal” is not one calendar. Each SKU family has its own curve, and the curve — not the holiday date — is what your inbound plan should be built against.

The head is rich, review-locked, and already discounting

The “halloween decorations” niche is a real market: 198 tracked products doing roughly $9.8M in combined monthly revenue — about $49,700 per product — at an average price of $46.57.

It is also defended. The average product in that set carries 2,004 reviews. A listing launched this month does not cross that moat by October; there is no compliant review path that fast. And the head is already fighting on price: one of the top listings tracked under the keyword shows a −34% price trend over the last 90 days, while an adjacent head term shows price moving the other way — dispersion that says promo season has effectively started in the head while some segments still hold price.

If you are already established in the head, this is your demand signal to re-forecast sell-through upward and check stock cover. If you are not established, entering the head six weeks before peak is a donation to the incumbents’ ad auctions.

The long tail is where a late entry can still work

Two examples from the same August 14 impression window, found in under a minute of filtering:

• “halloween shower curtain” — 60,562 exact impressions in 30 days against roughly 6,000 search results. The top tracked competitor has 129 reviews, and the segment’s 12-month sales are up 166%. A functional home product with a seasonal print, defensible with double-digit reviews.
• “halloween squishies” — 51,274 impressions against roughly 3,000 results, and the tracked competitor set averages a 3.4-star rating on thin review counts. Weak incumbents, real demand.

That is the general shape of a late seasonal entry that works: demand in the tens of thousands of impressions, competition in the low thousands of results, review bars in the low hundreds, incumbents with quality problems. Those setups exist every season; they are just invisible from the head keyword.

The backward calendar

Work from the shelf, not the factory.

To be selling when decorations demand peaks (October has been the best month for the head keywords in each recent year of this data), inventory needs to be received and available around October 1. Peak-season receiving is not instant — budget one to two weeks of buffer between carrier delivery and sellable status, more if your shipment splits across placements. US inbound transit adds several days to a week. Ocean freight from China to a US port is roughly two to four weeks depending on lane and port congestion.

Run that subtraction and the conclusion is uncomfortable but simple: ocean-bound Halloween inventory needs to be moving now — this week, not after Labor Day. Cargo that ships in mid-September arrives into the teeth of peak receiving with the season half over, and whatever misses the window converts into January storage fees and removal orders.

The remaining levers, in descending order of margin: a smaller ocean buy now, air freight for the proven sellers only, or cutting the SKU this year and planning a January-booked program for next season. What is not a lever is the calendar.

The Operator Read

Seasonal planning fails quietly, in August, when the demand curve starts without you — not loudly in October.

The data says Halloween shopping on Amazon US has already begun: decorations demand nearly doubled in three weeks, candy demand is accelerating fastest, and parts of the head are already discounting. The head niche is review-locked at ~2,000 reviews per incumbent; late entries belong in long-tail segments with thin, weak competition. And the freight math works backward from an October 1 shelf date to a ship date of approximately now.

Pull the curves for your own catalog this week. The sellers who do this check in August are the ones who spend October reading sales reports instead of storage-fee estimates.

Data in this note: Helium 10, US marketplace — weekly exact search volume through August 15, 2026; keyword and niche estimates pulled August 23, 2026. Revenue and sales figures are model estimates, not Amazon-reported actuals.