A trend report tells you what is rising this month. It never tells you what happened to last month's risers, which is the only thing that decides whether building a listing on one was worth it. So we kept the receipts. Since 23 August 2026 we have recorded, every day, whether Google demand for each of 1,481 Etsy keywords is rising, flat or falling, and where it sits against its own yearly peak — 37,711 daily readings over 76 days. This post follows the keywords that were rising in the first week and asks how long that lasted, whether flat keywords ever turn, and what the daily curves of four seasonal phrases look like from the first flicker to the fall. The answers are shorter and sharper than ‘trends last a season’.
The short version
- Half of all ‘rising’ signals are over within 8 days; 35% last a week or less. Only 1 in 7 keeps rising for more than two months.
- Of the risers still readable two months on, 64% were still flagged rising — but only 40% were actually higher than when they started.
- Flat keywords almost never turn: 1.4% of 328 flat keywords became rising within 30 days.
- Seasonal demand switches on in steps and falls off a cliff: ‘beach wedding guest’ lost 60% of its demand in a single week in early October.
- Generic terms spike last: ‘halloween decor’ peaked from 16 September; plain ‘halloween’ tripled only on 1 November — too late to list.
Most ‘rising’ keywords are not trends; they are blips. Of 179 keywords flagged rising in the first week of our daily record, the median run of consecutive rising days was 8; 35% lasted a week or less, 27% one to two weeks, 18% two weeks to a month, and 14.5% kept rising for more than 60 days — those are the ones worth a listing. Two months later, among the risers we could still read, 64% still carried the rising flag but only 40% sat above their starting level, because ‘rising’ measures change against the previous period, not height. Flat keywords stay flat: 1.4% turned rising within a month. Seasonal phrases do not drift; they step up weeks before the date and drop off a cliff after it, while the generic term spikes in the final days. Build on a riser only after two weeks of persistence, and on a seasonal phrase by its five-year shape, not this week's arrow — Keyword Research carries both, and flags the shape before the date.
→ track your phrases daily, with the five-year shape and Next-Month Radar
01How this was measured
Each day since 23 August 2026 we read, for every keyword in the store, Google Trends' demand index (0–100, relative to the keyword's own peak in the last twelve months) and its direction — rising, flat or falling against the previous period. By 6 November that is 37,711 readings across 1,481 keywords and 76 dates; the median keyword has 19 readings, 546 have thirty or more.
For the survival questions we took the 629 keywords whose first reading fell in the first seven days and split them by that first direction: 179 rising, 328 flat. We then looked at each one 14, 30, 45, 60 and 75 days after its first reading, using the latest reading within the preceding week. One caveat stated up front: keywords are re-read when sellers research them, so the later checkpoints hold fewer keywords (117 at two weeks, 56 at 75 days) and lean toward phrases people kept looking at. The run-length figures, which only need each keyword's own consecutive readings, do not have that problem. Nothing here is an Etsy search volume; Google demand is the signal, as it is throughout the search-volume truth post.
02How long a ‘rising’ signal lasts
Take the 179 first-week risers and count the consecutive days each stayed flagged rising from its first reading. Median: 8 days. 34.6% were over within seven days, 26.8% within eight to fourteen, 17.9% within fifteen to thirty, 6.1% within thirty-one to sixty — and 14.5% were still rising after sixty days.
That distribution is the whole answer to ‘is this trend worth a listing?’. Six risers in ten are a one- or two-week event: a news item, a holiday spike, a viral product that was already peaking when the flag appeared. Three in ten run two weeks to two months, which is a real season or a real climb. One in seven is the thing trend reports promise — a phrase still climbing two months later. You cannot tell which one you are looking at on day one. You can on day fourteen, which is why the rule in reading the direction signal is persistence before action.
Median run: 8 days. Six in ten rising flags are over within two weeks; one in seven is still rising after two months.
03Where the risers were two months later
Follow the same cohort forward. Of the risers still readable at each checkpoint: at 14 days 87.2% were still rising; at 30 days 57.7%; at 45 days 58.7%; at 60 days 63.8%; at 75 days 66.1%. The survivors look persistent — but remember the caveat: the keywords still being read at day 60 are the ones sellers kept researching, which selects for the phrases that kept moving.
The level tells a different story from the flag. At 60 days, 39.7% of readable risers sat above their starting demand index, 8.6% at the same level, and 46.6% below it. At 75 days: 39.3% higher, 48.2% lower. A keyword can be ‘rising’ against last period while sitting well under where it was when you first saw it, because the direction compares adjacent windows. Trend lists publish the arrow; the arrow is not the height. When you read a riser, read the index next to it, and read it against the five-year shape in the Keyword History Explorer — a phrase at index 20 ‘rising’ in October may simply be returning to its usual November level.
The flag persists more than the level does: at 60 days only 39.7% of these keywords sat above their starting demand index; 46.6% were lower.
A rising arrow is an eight-day fact. A season is a shape you can plan around. Confusing the two is how sellers build listings for a week of demand.
04Flat keywords stay flat
The other half of the question is whether a phrase that is not moving will start. Of 328 keywords flagged flat in the first week, 143 were readable 30 days later; 1.4% of them had turned rising. At 60 days the readable set was down to 28 keywords and 14.3% had turned — four keywords, too few to build on. The practical reading: a flat phrase is a stable phrase, which is a fine thing to list for if its five-year shape says it sells all year, and a poor thing to wait on in the hope that it turns. Evergreen vs seasonal Etsy products covers how to tell the two apart from the shape.
05Four seasonal curves, day by day
Averages hide the part sellers need, which is the timing. Four phrases from the daily record, with the full curves charted below.
- ‘halloween decor’ sat at 47 all of late August, stepped to 73 on 16 September, held that plateau for a month, and eased to 60 on 16 October and 56 on 23 October — before Halloween. The buying window for decor opened six weeks ahead of the date and was already closing ten days before it.
- ‘halloween’ on its own read 11 through September, 16 through October, and 52 on 1 November. The generic word spikes in the final days, when buyers want something now and sellers who list for it are two months late.
- ‘ceramic advent calendar’ was 0–3 until 15 September, 8 from 22 September, 24 from 30 September, dipped to 16 in late October and climbed again to 29 in November. Demand switched on in two steps across one week — the window for an Advent product opened in the last week of September.
- ‘beach wedding guest’ held 57–60 from August to 27 September, dropped to 47, then to 23 on 5 October — 60% of its demand gone in a week — and stayed there. Off-season is a cliff.
The same shapes are in the store for 1,481 phrases, and the point of a daily record is that the step and the cliff are visible the day they happen, not in next month's report.
Stepped up on 16 September, held a month, and was already easing ten days before Halloween.
From 4 to 79 in the last ten days. A listing made when this term moved would have gone live after the buying ended.
Two steps inside one week, 22–30 September. The Advent window opened then — and the five-year shape showed it in advance.
From 57 to 23 between 27 September and 5 October — 60% of demand gone in a week, then flat.
06What to do with it
Three rules that fall out of the numbers.
- A riser earns a listing after two weeks, not on day one. Six in ten rising flags are gone within a fortnight; the ones still rising on day 14 are the 40% worth your $0.20 and your afternoon. Check the index next to the arrow — rising from 4 to 6 is not a market.
- A seasonal phrase is listed by its shape, not its arrow. ‘halloween decor’ was already at its plateau when the generic term was still asleep; the five-year curve showed the September step in advance, and when to list on Etsy turns that into a list-by date six to eight weeks ahead of the peak.
- Plan for the cliff. Beach wedding demand did not taper; it fell by 60% in a week. Seasonal listings need their tags and photos swapped before the drop, not after the sales stop — the seasonal keywords guide covers the swap.
If you would rather not read 1,481 daily curves, Keyword Research tracks the phrases you care about and its Next-Month Radar flags the ones whose shape says they switch on next month.