WAYAMZ Demand Index, September 2026: Halloween is seven weeks early and Christmas is eleven days late
Our first monthly index of 24 tracked US Amazon keywords. Halloween decorations already sit at 99.2% of last October's peak; Christmas decorations are running 11 days behind their own 2025 curve; and the back-to-school window closed 36.5% smaller than last year. Plus the measurement trap that makes 23 of 24 unrelated keywords spike in the same week.
By WAYAMZ Team
This is the first edition of the WAYAMZ Demand Index, a monthly measurement of 24 US Amazon keywords — twelve evergreen, twelve seasonal — published on the first of every month. The one-sentence version of September: seasonal demand did not move forward as a block, it fractured, and Halloween is seven weeks ahead of Christmas decor in a race they normally run about ten days apart.
Signal one: Halloween is at last October’s peak, in August
Halloween decorations recorded 794,104 monthly US searches in the reading ending August 22, 2026. The highest reading that keyword produced in the entire 2025 season was 800,776, in the week ending October 11.
August is at 99.2% of last October.
Year over year against the equivalent week, the term is up 120.1%. But the more useful way to state it is as a phase shift. Take the 2026 level of 794,104 and find where the 2025 curve first crossed it: interpolating between the readings of October 4 (702,733) and October 11 (800,776), that is roughly October 10, 2025. The equivalent-week comparison point is August 23, 2025. Halloween demand in 2026 is therefore running about 48 days ahead of its own 2025 curve.
Seven weeks. For an operator that means the decision window for Halloween inventory, creative, and bids did not open in September this year — it opened and largely closed in July.
Signal two: Christmas decor is behind, and Christmas gifting is ahead
The obvious inference from signal one is that shoppers moved earlier across the board. The data says otherwise, in the most closely related category available.
Christmas decorations recorded 136,201 searches — up 89.3% month over month, which reads like a strong ramp until you compare it to last year’s same week, which was 250,302. That is down 45.6% year over year. Run the same phase calculation and the 2025 curve crossed 136,201 around August 12. Christmas decorations in 2026 are running roughly 11 days behind their 2025 pace.
Then look at advent calendar, which is a Christmas keyword but a gifting one rather than a decor one: 35,970 searches, up 85.5% year over year, and about 10 days ahead of its 2025 curve.
Same holiday. Two terms. Twenty-one days of separation in opposite directions.
The read is that the pull-forward is real but it belongs to gift buying, not to decorating. Gift-shaped seasonal SKUs need their listings, inventory, and bid ramps live earlier than the calendar suggests. Decor-shaped seasonal SKUs got a reprieve of about a week and a half — which is a reprieve worth having, but only if you know it applies to you and not to the gift SKU sitting next to it in the same account.
Cold-weather demand, for what it is worth, is on schedule: space heater is running about 1 day ahead of its 2025 curve and electric blanket about 3 days ahead. Both are inside the noise. Not everything moved.
Signal three: back-to-school closed early and closed smaller
Back to school supplies peaked at 244,832 in the reading ending August 1, 2026. The 2025 peak for the same term was 385,495, in the reading ending August 2.
The peak arrived on essentially the same calendar day and came in 36.5% smaller. By August 22 the term was at 91,819 — it has already surrendered 62.5% of this year’s peak.
Two things follow. First, the same-week year-over-year figure for this keyword is +13.3%, which is positive and completely misleading: the term is positive on that comparison only because the 2025 curve had already collapsed further by that date. A single YoY number taken at the wrong point on a steep curve will tell you a growing category is shrinking, or the reverse. Second, the adjacent evergreen terms tell the more actionable story — desk lamp is up 39.2% month over month and kitchen organizer up 14.7%, both still climbing on August 22 while the explicit back-to-school term collapses. The dorm and study-space purchase runs about three weeks longer than the supplies purchase, and it runs on evergreen keywords where the competition has already stopped bidding.
The full index
Current reading is the week ending August 22, 2026. MoM compares it to the reading ending July 25, 2026; YoY compares it to the reading ending August 23, 2025.
Evergreen basket
| Keyword | Searches | MoM | YoY | Read |
|---|---|---|---|---|
| desk lamp | 424,850 | +39.2% | +19.4% | Still climbing on Aug 22; in 2025 it had already peaked and fallen by Aug 30. The window is running longer this year. |
| water bottle | 1,143,873 | +24.6% | +7.7% | Back-to-school carry. The weakest YoY in the top half — volume is up but the category is not growing. |
| kitchen organizer | 44,517 | +14.7% | +61.0% | Smallest term in the basket and the second-strongest YoY. Dorm setup plus the autumn home reset. |
| yoga mat | 567,433 | +11.9% | +57.3% | September resolution demand is already loading in August. |
| air fryer | 1,102,468 | +5.9% | +48.5% | Largest evergreen term we track and still adding 48% YoY. Cooking-at-home demand has not normalised. |
| shower head | 611,438 | +5.4% | +17.3% | Steady. Useful as a low-seasonality reference point inside the basket. |
| car phone mount | 88,453 | +3.5% | +40.0% | Flat month, strong year. Accessory demand tracking device replacement cycles. |
| wireless earbuds | 233,911 | +3.4% | +22.6% | Pre-Q4 accumulation, not yet the gifting ramp — that arrives in the November readings. |
| laundry basket | 567,950 | +2.0% | +54.5% | Dorm-driven, and the +54.5% YoY says the category itself expanded. |
| resistance bands | 406,510 | −2.4% | +111.0% | Only keyword in the index that more than doubled year over year. Flat month, transformed year. |
| dog toys | 291,817 | −4.4% | +34.2% | Slight cooling ahead of a category that historically multiplies 5x into December. |
| phone case | 164,364 | −7.4% | −24.4% | The only evergreen term down on both timeframes. Worth watching as a possible term-fragmentation effect rather than lost demand. |
Seasonal basket
| Keyword | Searches | MoM | YoY | Read |
|---|---|---|---|---|
| halloween decorations | 794,104 | +170.5% | +120.1% | At 99.2% of last October’s peak, 48 days ahead of its 2025 curve. The signal of the month. |
| christmas decorations | 136,201 | +89.3% | −45.6% | Strong month, weak year. Running about 11 days behind 2025. |
| space heater | 47,492 | +60.8% | +11.4% | On schedule — roughly 1 day ahead of last year. The cold-weather ramp has not been pulled forward. |
| electric blanket | 34,473 | +29.4% | +24.7% | Also on schedule, about 3 days ahead. Confirms the space heater read. |
| advent calendar | 35,970 | +21.9% | +85.5% | About 10 days ahead of 2025 — Christmas gifting moved earlier even though Christmas decor did not. |
| easter basket | 1,983 | +10.0% | +25.0% | Deep off-season floor. Its 2026 peak was 2,523,887, so the floor is 0.08% of peak. |
| humidifier | 215,761 | +3.8% | +58.1% | Early in its own ramp; last season it roughly tripled from here into December. |
| patio furniture | 69,653 | −8.1% | −14.0% | Season closing, and closing weaker than last year. Clearance decisions are already late. |
| valentines gifts | 125 | −35.2% | +557.9% | Statistically meaningless at this volume. Included to show what a dead-zone reading looks like: the YoY is +557.9% on a base of 19. |
| sunscreen | 319,166 | −43.3% | +46.8% | Steep seasonal decay but 46.8% above last year at the same point — the category is genuinely bigger. |
| beach towel | 117,615 | −57.7% | +25.3% | Same shape as sunscreen: sharp decay, higher floor than 2025. |
| back to school supplies | 91,819 | −59.4% | +13.3% | Peaked 36.5% below 2025 and has already given back 62.5% of that peak. |
Basket aggregates: the twelve evergreen terms are up 9.5% month over month and 30.8% year over year in combined volume; the twelve seasonal terms are up 3.4% and 42.1%. Sixteen of the 24 terms rose month over month.
The measurement trap: why 23 of 24 keywords spiked in the same week
This is the part of the index we would most like other people to copy, because it invalidates a lot of casually produced seasonality analysis — including analysis we would have produced ourselves a year ago.
In the reading ending June 27, 2026 — a major Amazon shopping event week — we compared each keyword against the average of its two neighbouring readings. 23 of the 24 keywords came in at least 1.15x above that baseline. The median was 1.85x. Air fryer hit 3.19x. The only term that did not move was valentines gifts, which sits at a volume where nothing is measurable.
That list includes christmas decorations in late June. It includes easter basket, eleven weeks after Easter. There is no commercial mechanism by which an Amazon shopping event creates genuine demand for Easter baskets in June.
The pattern repeats. In the reading ending October 11, 2025 — another event week — 24 of 24 keywords came in above their neighbours, median 1.37x. In both cases the spike reading is followed immediately by a below-trend reading, then a return to the prior level. Demand does not behave that way. Measurement does.
The practical rule: never compute a month-over-month change using a reading from an event week or the week immediately after it. Every figure in this index uses August 22 against July 25 — four weeks apart, both outside the event week and its rebound. If you are running this analysis on your own keywords, check whether either endpoint sits on an event week before you act on the delta. If it does, the number is an artifact of the measurement, and you will make an inventory decision against a spike that never existed.
Operator takeaways
Split your seasonal catalog by gift versus non-gift, not by holiday. The 21-day gap between advent calendar and christmas decorations is inside the same holiday. Holiday is not the variable that predicts timing this year; purchase intent is.
If you sell Halloween, the decisions are already behind you — audit rather than plan. At 99.2% of last October’s peak in an August reading, the useful question is no longer “when do we ramp” but “is our inventory position and our bid ceiling correct for demand that is already here.”
Buy the adjacent evergreen term when the explicit seasonal term collapses. Back to school supplies fell 59.4% while desk lamp rose 39.2% and kitchen organizer rose 14.7%. Sellers who bid the obvious seasonal term follow it down; the demand moves to durable keywords where competition thins out at exactly the moment volume holds.
Treat a single YoY reading on a steep curve as unreliable. Back to school supplies shows +13.3% YoY while its peak came in 36.5% smaller. Both numbers are correct. Only one is true. Compare peaks to peaks, or compare phase, before you conclude a category is growing.
Check your endpoints for event weeks before trusting any delta. 23 of 24 unrelated keywords moved together in one week. If your month-over-month math touches that week, throw the number out.
Methodology
The index tracks 24 US-marketplace keywords: twelve evergreen terms selected for category spread and low seasonality, and twelve seasonal terms weighted toward the holidays adjacent to the publication month. All search volume comes from Helium 10 (get_keywords_search_volume_history, US marketplace, exact monthly search volume, weekly resolution). Nothing here comes from a third-party estimate or a modelled figure.
The current reading is the week ending August 22, 2026, the most recent available on the pull date of September 1, 2026. Month over month compares it to the reading ending July 25, 2026 — four weeks apart, and chosen so that neither endpoint falls on an event week or its rebound week. Year over year compares it to the reading ending August 23, 2025, the nearest equivalent week 52 weeks prior. Phase-shift figures (“48 days ahead”) are computed by locating the current level on the prior-year curve using linear interpolation between the two bracketing weekly readings, then differencing that date against the equivalent-week comparison point.
Two limitations we would rather state than have someone find. First, aggregate year-over-year figures across a whole basket can absorb changes in the data provider’s own estimation model as well as real demand, so treat basket-level YoY as directional and treat within-month rankings and phase shifts — which are computed against the same source in the same pull — as the firmer readings. Second, terms below roughly 5,000 monthly searches produce percentage changes that are arithmetically valid and practically meaningless; valentines gifts is retained in the table as an illustration of that failure mode rather than as a finding.
All 24 tracked terms returned data this month, so no substitutions were required. This dataset is free to cite with attribution to WAYAMZ. The next edition publishes October 1, 2026.
The Operator Read
Seasonal demand in 2026 did not shift forward — it fractured. Halloween decorations are running 48 days ahead of their own 2025 curve and already sit at 99.2% of last October’s peak, while christmas decorations run 11 days behind and advent calendar runs 10 days ahead, meaning the split is gift-versus-decor rather than holiday-versus-holiday. Back-to-school peaked 36.5% smaller and has already given back 62.5% of that peak, pushing the live demand onto adjacent evergreen terms like desk lamp (+39.2% MoM) where bidding thins out. And before you act on any of your own month-over-month numbers, check the endpoints: in an Amazon event week, 23 of our 24 unrelated keywords rose a median 1.85x, which means the most confident-looking spike in your data is the one most likely to be an artifact.
The Operator Brief
One email a week, operator's cut.
The week's sharpest Amazon signals from the journal — policy changes, fee math, and the data reads worth acting on. No spam, unsubscribe anytime.