Scryup reads an Amazon product's sales-rank history the moment it starts trending and calls it by day 3: keeper or fast fade. Stock the keepers, skip the duds.
A tool for Amazon sellers that reads a product's sales-rank history the moment it starts trending and predicts, by day 3, whether the trend will keep selling for months (a keeper) or fade fast (a dud) — so you stock the keepers and skip the duds.
On its top-confidence calls, Scryup is right about 7 in 10 times at its three-week grade — versus a base rate where only about 3 in 10 trends last. The calls we publish stay up either way, so the record can be checked.
The method comes from quantitative trading: telling a signal that persists from noise that snaps back. We analyzed years of daily sales-rank history on thousands of Amazon products. A model then learned which early-days shapes preceded demand that held and which preceded a fade — trained on outcomes it could not see, and graded only on surges it was never trained on.
When a product blows up, you have to decide now: stock up, or stay out. Get it right and you ride the wave. Get it wrong and you're sitting on dead stock, marking it down, and paying to store it — monthly storage from the day it lands, aged-inventory surcharges from day 181, and in Q4 the storage rates sting from week one.
And "everyone can see it's hot" includes the twenty other sellers about to jump on the listing and split the buy box. In our data, most duds are back near their old rank within about two weeks, and only about 3 in 10 trends clear our three-week keeper bar. You can't just wait for proof: by the time a trend is obviously real, the listing is crowded and your sourcing window has closed. The margin goes to whoever knew at day 3.
No vague hype. Measured across about 94,000 US Amazon products we track daily in 21 major categories.
A sudden jump. At day 3, you can't yet tell if it'll last.
A product selling far faster than its own recent normal — sales rank improves by at least half — and holding that jump for at least a few days. Not a one-day blip.
Decision time. The sourcing window is open for days, not weeks — this is exactly when the keeper-or-dud call has to land.
Jumps — and stays up for weeks.
A trend with staying power. Three weeks in, it's still selling strong — and keepers tend to keep going for months.
Worth stocking — a keeper usually has months of demand ahead, not weeks. Only about 3 in 10 trends ever clear that bar.
Jumps the same way — then falls back to normal within ~2 weeks.
A fast fade. The trend collapses back toward the product's old, ordinary sales level within the first week or two.
Don't stock up — you'd be racing the storage clock to sell off at a loss. About 7 in 10 trends end this way.
Both surge the same way. The split shows up later — right when most sellers have already committed inventory. Scryup makes the call on day 3, while the two are still indistinguishable by eye.
Illustrative curves, shaped to the medians we measure across about 94,000 US Amazon products. Lower = selling better.
Paste a product link or ASIN — the 10-character Amazon ID that starts with B0. No signup, no email.
We pull its full daily sales-rank, price, and seller-count history and score the surge against thousands of past trends with the same shape — including whether a price cut started it.
Keeper, Dud, Borderline, or Not Trending — with a confidence rating, the two reasons behind it, and the product's own trend track record. Pro adds the market snapshot: estimated units/day and seller count with a crowding flag.
Other tools either show you a chart and make you guess, or promise "winning products" and never say how often they're wrong. We do the opposite.
We make a real forward call: sustain or fade, over a defined window.
We tested it on data the model had never seen. On hundreds of top-confidence calls held out from training, we're right about 7 in 10 times at our three-week grade, against a base rate where only 3 in 10 trends lasts.
And our keepers keep going. On a top-confidence keeper call, measured from the day we make it and counting every miss, about 2 in 3 are still selling strong two months later — and still about 2 in 3 at three months, at least double their old pace. Long enough to source, land, and sell through.
We tell you when we're not sure. We can't get a read on every trend. When that happens, we say so honestly.
Only 3 in 10 trends last, so buying on instinct loses twice for every win. Scryup moves that ratio. It doesn't remove the risk.
Scryup scores every surge it finds across the products it tracks — 13,757 so far in 21 categories, 10,928 of them already graded against what happened next. Far too many to post, so a couple go on this board each day as the audit trail: posted the day we make them, then graded three weeks later against what the sales rank actually did. Hits and misses both stay on the board — follow along on X @scryup.
| Called | Product | Call | Confidence | Rank jump | Result at +21 days | Result at +42 days |
|---|---|---|---|---|---|---|
| Eucerin Radiant Tone Dark Spot Corrector Body Cream, Thiamid Beauty & Personal Care | DUD | High | 3× | Pending — grades ~2026-10-14 | Pending — grades ~2026-11-05 | |
| Magic Sleepsuit Baby Merlin's Cotton Baby Swaddle Transition Home & Kitchen | DUD | High | 14× | Pending — grades ~2026-10-13 | Pending — grades ~2026-11-04 | |
| Cetaphil Ultra Gentle Refreshing Body Wash, Refreshing Scent Beauty & Personal Care | DUD | High | 5× | Pending — grades ~2026-10-13 | Pending — grades ~2026-11-03 | |
| ijbnhd Cross Diamond Art Kits for Adults, Jesus Pray Cross D Arts, Crafts & Sewing | DUD | High | 3× | Pending — grades ~2026-10-11 | Pending — grades ~2026-11-02 | |
| 3W Floor Mats&Cargo Liner Fit for Jeep Wrangler JL 2018-2025 Automotive | DUD | High | 9× | Pending — grades ~2026-10-11 | Pending — grades ~2026-11-01 | |
| Swihauk (variant family, 2 listings) Industrial & Scientific | KEEPER | Top | 16× | Pending — grades ~2026-10-10 | Pending — grades ~2026-10-31 | |
| Bag Balm Vermonts Original Moisturizing And Softening Ointme Pet Supplies | DUD | High | 4× | Pending — grades ~2026-10-09 | Pending — grades ~2026-10-31 | |
| Pokémon TCG: First Partner Illustration Collection —Series 2 Toys & Games | KEEPER | Top | 522× | Pending — grades ~2026-10-10 | Pending — grades ~2026-10-30 | |
| Lemict 6FT Yellow Table Cloth - 2 Pack Fitted Stretch Rectan Home & Kitchen | DUD | High | 4× | Pending — grades ~2026-10-08 | Pending — grades ~2026-10-30 | |
| Microphone Transmitter Audio Cable Coiled for Sennheiser SK5 Musical Instruments | DUD | High | 5× | Pending — grades ~2026-10-07 | Pending — grades ~2026-10-29 | |
| Ramalto 114 Pcs Ring Sizers for Loose Rings, Invisible Ring Arts, Crafts & Sewing | DUD | High | 2× | Pending — grades ~2026-10-05 | Pending — grades ~2026-10-27 | |
| Mega Man Star Force Legacy Collection - Nintendo Switch Video Games | KEEPER | High | 9× | Pending — grades ~2026-10-04 | Pending — grades ~2026-10-27 | |
| Mario Vs. Donkey Kong™ - US Version Video Games | KEEPER | High | 3× | Pending — grades ~2026-10-03 | Pending — grades ~2026-10-26 | |
| Pets First unisex adult NFL Pet Bandana, Washington Commande Pet Supplies | DUD | High | 4× | Pending — grades ~2026-10-04 | Pending — grades ~2026-10-26 | |
| Tiki Cat Wet Cat Food Mega Pack, Seafood Selects, 2.8 oz Can Pet Supplies | DUD | High | 6× | Pending — grades ~2026-10-04 | Pending — grades ~2026-10-25 | |
| Absorbine Veterinary Liniment Topical Analgesic Sore Muscle Pet Supplies | KEEPER | High | 8× | Pending — grades ~2026-10-03 | Pending — grades ~2026-10-25 | |
| BESAQUO Shower Head, 9 Settings High Pressure shower head wi Tools & Home Improvement | DUD | High | 11× | Pending — grades ~2026-10-02 | Pending — grades ~2026-10-24 | |
| LITTLE TREES Car Air Freshener | Hanging Tree Provides Long Automotive | KEEPER | Ultra | 39× | Pending — grades ~2026-10-04 | Pending — grades ~2026-10-24 | |
| Wooden Busy Board with LED Light Montessori Baby Learning To Toys & Games | DUD | High | 4× | Right — Faded — back near its old rank by day 21. | Pending — grades ~2026-10-23 | |
| TONIFUL 16 Inch 4th of July Balloons Banner Fourth of July L Toys & Games | DUD | High | 3× | Right — Faded — back near its old rank by day 21. | Pending — grades ~2026-10-22 | |
| MONGERY Carrying Case for Nintendo Switch 2 (2025), Hard She Video Games | DUD | High | 8× | Right — Faded — back near its old rank by day 21. | Pending — grades ~2026-10-21 | |
| Space: 1999 - The Complete Series Health & Household | KEEPER | Ultra | 85× | Miss — Faded — a miss, it stays on the board. | Pending — grades ~2026-10-20 | |
| Glad ForceFlex Tall Kitchen Drawstring Trash Bags, 13 Gal, F Health & Household | DUD | High | 16× | Right — Faded — back near its old rank by day 21. | Pending — grades ~2026-10-20 | |
| Avanti Ice Maker Mighty Ice Countertop Design Makes 26 Pound Appliances | KEEPER | Top | 11× | Right — Held — 8.5× its old pace on day 21. | Pending — grades ~2026-10-17 | |
| Icon Sports Officially Licensed US Soccer Federation Soccer Sports & Outdoors | KEEPER | Ultra | 679× | Miss — Faded — a miss, it stays on the board. | Pending — grades ~2026-10-16 | |
| Copper Pearl Baby Bows: Stylish Headbands for Baby Girls, Pr Baby Products | DUD | High | 3× | Right — Faded — back near its old rank by day 21. | Pending — grades ~2026-10-16 | |
| H7 LED Light Bulbs, Pack of 2 Automotive | DUD | High | 3× | Right — Faded — back near its old rank by day 21. | Pending — grades ~2026-10-15 | |
| Redragon K521KS Wireless Gaming Keyboard, Tri-Mode BT/2.4G/U Video Games | KEEPER | Top | 12× | Right — Held — 8.3× its old pace on day 21. | Pending — grades ~2026-10-15 | |
| Hydro Flask Micro Bottle (variant family, 4 listings) Sports & Outdoors | KEEPER | High | 12× | Right — Held — 2.7× its old pace on day 21. | Pending — grades ~2026-10-13 | |
| Hobonichi Techo Accessories Hobotoridori: Pencil Board for W Office Products | DUD | High | 16× | Right — Faded — back near its old rank by day 21. | Pending — grades ~2026-10-13 |
How to read this board. Rank jump is how many times the product’s Amazon sales rank improved over its own pre-surge baseline at the moment of the call. High and ULTRA are our keeper-confidence tiers: on held-out data about 7 in 10 ULTRA calls are still selling at three weeks, and about 3 in 5 High calls — against a ~3-in-10 base rate.
LIVE rows were posted in real time. SEMI-LIVE rows were posted a few days after their call date, but while still pending, before any verdict existed. BACKTEST rows apply the same rules day by day to held-out history the model was never trained on (trained through Feb 2026; every backtest call starts after its last training label).
Every call — all three kinds — is graded against one bar, fixed before the call is made: three weeks after the surge began, is the product still selling at at least 2× its pre-trend pace AND ranking among its category’s proven sellers? Clears both = a keeper call was right; fails either = a dud call was right. The six-week counter asks that same question again over days 31–42 instead of 10–21 — nothing is redefined, just asked later, because sourcing takes weeks.
No judgment calls, no moved goalposts: every miss is shown, and nothing is ever removed or edited. Fewer calls carry a six-week verdict than a three-week one — some are simply too young, and a few can never get one, because the product stopped selling so completely that its rank data goes flat and there is nothing left to measure. Those are left out rather than counted as wins.
The free check is yours forever. Scryup Pro removes the limits and watches your inventory for you.
Only want the Q4 pre-season list? It is $79 on its own, and the full $79 comes off an annual plan if you upgrade within 30 days.
Secure checkout by Stripe · cancel anytime · any applicable tax is shown at checkout. The free instant lookup stays free — Pro just removes the limits and adds the weekly report and watchlist.
Not sure it's really trending? Send it anyway — "Not Trending" is a real verdict, and knowing that before you buy is the whole point.
One email a month: the products that surged on Amazon, our keeper-or-dud verdict on each, and how earlier calls actually graded. Pro members get it weekly, the week each surge breaks. No spam — unsubscribe anytime.
The 10-character Amazon product ID (starts with B0), in every product URL.
They show you history and let you guess. Scryup makes the forward call — will this trend last — and tells you how confident we are.
On top-confidence calls, right about 7 in 10 at our three-week grade — when only ~3 in 10 trends last at all. It's a prediction made from data analysis, not a guarantee.
The right question — sourcing takes weeks. Measured from the day we make a top-confidence call and counting every miss: about 2 in 3 are still selling strong two months later, and still about 2 in 3 at three months. Long enough to source, land, and sell through.
We check. Pro flags any surge that began alongside a price cut of 15% or more, so you don't mistake a promo bump for organic demand.
Pro shows the seller count now vs. before the trend, with a crowding flag — so you can judge whether the buy box is about to get split before you commit units.
A top-confidence keeper call still misses about 3 in 10 times at three weeks. We publish that so you can size your buy accordingly — being right about 7 in 10 is still more than double the rate at which trends last on their own, not a guarantee.
Amazon US (amazon.com) for now, working from full daily sales-rank, price, and seller-count history. Expansion into other Amazon marketplaces is on the roadmap — members get a say in which comes first.
Selling elsewhere? Tell us your marketplace and we'll notify you when we land there →
We'll tell you exactly that — plainly, with the data. "Not Trending" is a real verdict, and knowing it before you buy costs you nothing.
The method comes from quantitative trading: telling a signal that persists from noise that snaps back. We analyzed years of daily sales-rank history on thousands of Amazon products. A model then learned which early-days shapes preceded demand that held and which preceded a fade — trained on outcomes it could not see at call time, and graded only on surges it was never trained on. That is why the calls we publish stay up with their results: the track record is the test.
No. It's a prediction made from data analysis, to inform your own decision. Not sourcing or investment advice.
Scryup reads amazon.com today. Tell us where you sell and we'll email you the moment we expand there — the most-requested marketplace gets built first.