Repository navigation
Product Selection Methodology
See the market clearly with platform data — then decide what to sell and how. Covers shelf e-commerce (Amazon, Walmart, Shopee, Temu, AliExpress, Mercado Libre, etc.) and interest-based e-commerce (TikTok Shop).
Decision boundary: every dimension, index, ranking and calculation in this document is reference and advice only, meant to help you see the market faster and more completely. Whether to enter, how much to invest, and when to act is always the seller's decision. Conclusions are based on historical and current data; markets change constantly, and no business outcome is guaranteed.
Data caveats: sales volume, monopoly coefficients, CPC and similar figures are largely estimates (e.g., sales inferred from BSR). Use them for elimination and structural judgment, not for precise rankings. Cross-check before entering (Amazon Brand Analytics, manual front-end sampling, your own store data). Cases marked with a pull date are live API measurements; all others are historical teaching cases and do not represent current market conditions.
Scope: Amazon Haul (the low-price mall) is an invitation-only program and is out of scope here.
One shared goal: with profit secured, earn platform weight (sustained free traffic / organic exposure) in the shortest time at the lowest risk, and grow organic orders quickly, safely, and steadily.
"Weight" takes different forms on different platforms. Identify the platform's logic first, then apply the matching playbook:
| Dimension | Shelf e-commerce (Amazon / Walmart / Shopee / Temu / AliExpress / Mercado) | Interest-based e-commerce (TikTok Shop) |
|---|---|---|
| Traffic nature | People find products: search-driven; rankings and keywords allocate traffic | Products find people: algorithmic recommendation; content pushes products to potential buyers |
| Buyer mindset | Rational brain: shopping mode, arrives with a need to solve | Emotional + visual brain: entertainment mode, 3-second hook, impulse buying |
| Core question | Is this product better? | Is this product more interesting? |
| Form of weight | Listing weight, organic keyword ranking, Best Seller lists | Content recommendation weight, organic product sell-through, creator matrix |
| Success factors | Product strength + listing weight + keyword ranking | Content strength + topicality + visual impact + creator fit |
| Selection priority | Battlefield first: pick the market, then refine the product (data-driven) | Content first: assess content potential, then validate the market (creative-first, data-validated) |
The platform logic chain (shelf):
Better customer experience → impressions → more clicks than peers → higher conversion
→ higher and more stable prices → more and better genuine reviews → weight (= free traffic)
The interest-based equivalent: great content → higher watch time and engagement → more recommended impressions → higher conversion → more genuine sales and reviews → product weight and more organic reach. The shared core never changes: better conversion and reputation earn free distribution.
Weight is an outcome metric, not an operating metric. The algorithm is a black box, and "higher, more stable prices → weight" fails in downward price-spiral categories (raise price → conversion drops → weight drops). Operate only on what you control: conversion rate, negative-review rate, return rate, and staying in stock. Watch weight as a result, never chase it as an action.
| Era | Market condition | Selection orientation |
|---|---|---|
| Platform boom (pre-2016) | Search demand > product supply; listings get impressions for free, no traffic cost | Category demand and sales volume; SKU count drives traffic |
| Stock competition (post-2016) | Oversupply, homogenization, price wars + paid traffic; conversion rate and traffic cost become the decisive variables for margin and survival | Oriented by difficulty of earning weight + profit; listing weight drives free traffic and profit |
| Platform rule shifts (2023+) | "Reviews are king" weakens: front-end tilts toward new sellers/products; Review counts become Rating counts; AI review summaries | Review moats are actively weakened by the platform — the window for new products widens |
Both logics share one closed loop; only the signal sources differ at each step:
⓪ Hard resource screen (budget / certification capability / supply-chain radius / never-touch categories)
→ ① Full-set ranking inside the pool, discover opportunities → ② Final review of Top N, validate competition
→ ③ ROI calculation → ④ Differentiated build → ⑤ Post-launch monitoring & iteration (incl. exit)
↑ Throughout: risk flags, honest data, seller decides ↓
Resource constraints first, ranking second, deep-dive last. Do not binge-analyze all 17,000 categories. Rank only inside the qualified pool; how much you review each day is bounded by your budget and team capacity — this prevents "opportunity addiction" (everything looks like an opportunity, nothing gets a deep look).
| Step | Shelf e-commerce (Amazon baseline) | Interest-based e-commerce (TikTok) |
|---|---|---|
| ① Discover | All-category dashboard (17,000+ subcategories, 90+ dimensions, full-set ranking); Hidden Profit Index for 6-in-1 products; long-tail markets beyond the Top 100 (e.g., ranks 101–400) | Content potential first: trending videos / shoppable products / creator dashboards — find "scalable content formulas × selling products" |
| ② Validate | Three monopoly coefficients (Listing/brand/seller + Amazon retail), new-product share, review moat, price bands | Seller/brand/creator concentration, content lifecycle (one-hit wonder vs. continuous new-product sell-through), AOV and conversion |
| ③ Calculate | COGS + first-mile + FBA + commission + returns + CPC × conversion → ACOS ceiling | COGS + fulfillment + commission + returns + creator commission / samples / content production / paid boosting → contribution per order and break-even |
| ④ Build | Differentiated product + listing (title / bullets / main images / A+) | Camera-ready selling points + content formulas (script / first 3 seconds) + creator mix |
| ⑤ Monitor | Rankings / keywords / price / rating / inventory + exit triggers | Video performance / product sell-through / creator output / creative fatigue and refresh + exit triggers |
Full-set ranking, no hard thresholds (both logics, inside ①): rank everything by composite dimensions and keep it all visible. Hard filters ("monthly sales ≥ 1000 AND reviews ≤ 500") make borderline candidates disappear; ranking merely places them lower — the "gray zone" excluded by everybody else's filters deserves a second look. Rankings are reference, not an entry list.
Snapshots and cadence: indices, rankings, price bands and content heat are all dynamic snapshots — re-run regularly (every ~30 days); never let one analysis govern a whole year. Validate in small steps first (small test order, small budget), scale only after validation, and return to the pool without hesitation when it fails.
Analogy: in food-delivery, hotel and flight apps, consumers use "full list + sort + filter" to find what fits them fastest. Product selection works the same way — put 17,000+ subcategories on one dashboard and compare them side by side, instead of reviewing categories one by one.
Dashboard fields are organized into 12 dimension groups: market demand / market trend / seller profile / monopoly & competition / market volatility / new-product opportunity / capital cost / logistics cost / operating cost / after-sales cost / margin headroom. Key columns:
- Three monopoly coefficients: Listing monopoly, brand monopoly, seller monopoly (plus Amazon retail share and Chinese vs. overseas seller ratio)
- New-product opportunity: 1/3/6-month new-product count share + sales share, average days since launch
- Entry moats: average review count, new-product review average, A+ video share, brand-flagship-store ratio
- Cost risks: average CPC, return rate, weight & volume (→ first-mile cost and capital efficiency), listing-suspension ("dead dog") rate
- Averages: price, unit profit, margin, average hijackers, average brand/seller market size and monthly sales
- Forward trend: expected month-over-month change
Contrast examples (historical teaching cases):
- Coffee pods: seller monopoly 93.15%, Amazon retail share 90.37%, average reviews 22,476, new-product share 0%, average listing age 2,308 days → a battlefield owned by Amazon retail
- Diapers: brand monopoly 88.45% + seller monopoly 98.64%, retail 97.62% → dual brand + retail monopoly
- iPhone cases: low monopoly (Listing 10.60% / brand 36.78% / seller 35.37%) but 7.81% return rate and expensive CPC → low monopoly ≠ profitable; must pass the calculation gate
- Chin Strips: market size 15,652, 70 brands, 74% new-product share → a structure new products can still enter
Survival-line check (mandatory after comparison): the flip side of low barriers is a low ceiling and easy copying. Even when a small category is open, compute: category volume × reachable share × unit net profit ≥ monthly team cost. Example: a ~15k units/month category at 2% share ≈ 300 units/month — after ads and returns, that may not feed a team. "Easy to enter" is not "worth entering."
| Group | Dimensions |
|---|---|
| Market | Total sales (demand size), sales growth trend, multi-year seasonality for forecasting, total sellers and products, seller type (3P vs. Amazon retail), Chinese vs. overseas seller ratio |
| Competition | Sales concentration (single-ASIN dominance, sales distribution, sales of the #100 product), brand concentration, store concentration, unbranded share, brand-flagship ratio, hijacker count, rank-list churn, max CPC |
| Logistics | FBA vs. FBM ratio, weight & volume (→ shipping method, capital requirement, capital efficiency) |
| Listings | New-product share, star ratings (top-product rating vs. sales share; after-sales risk), average selling price, launch-time distribution of new vs. old products, A+ video share, variation ratio, review counts (top-product reviews vs. sales share; sales of the lowest-review products) |
18 reference criteria of a "good category" (six groups):
- No monopoly: no single product dominates sales and traffic entry; no brand owns consumer mindshare; no store (incl. Amazon retail) dominates the main traffic entries
- Low entry barriers: low review requirements; high new-product count and sales share with real momentum; Top 100 slots not filled; competitors are weak overall
- Healthy demand: stable or fast-growing demand, with other dimensions trending positively too; solid category-wide conversion
- Margin headroom: high and stable price ceiling with low product cost; rich price architecture with stable high-end demand and no monopoly; rich product elements with premium and packaging angles; broad attribute space
- Affordable promotion: low white-hat traffic cost that gross margin can fully cover
- Friendly fulfillment & after-sales: low first-mile cost and high capital efficiency; meaningful FBM share (FBA entry angle); low after-sales risk or large room for improvement
Reading the contrast cases (historical):
| Case | Signal | Reference reading |
|---|---|---|
| Kids' travel pillows vs. sheet sets (Listing monopoly) | Former: #1 takes a cliff-like share; latter: flat sales distribution | Flat distribution = traffic entry not captured |
| Magnetic dress-up dolls vs. wall mirrors (brand monopoly) | Former: one brand owns mindshare; latter: multi-peak brand spread | Spread brands = weak brand moat |
| Screen protectors (review moat) | Average reviews 30,364, max 296,985, new-product review average 22,231 | Huge review moat; costly for newcomers to catch up |
| Chin strip ranked #12 (new-product activity) | A few months old, 13 reviews, 3.60 stars, 371 units/month, 2.32% share | Few reviews and mediocre rating still reach the top = low moat |
| Evaporative coolers (new-product activity) | New-product count/sales share, A+ and FBM structure | New products holding share = rank list not frozen |
- Start from sales: read year-over-year curves for seasonality and time your stocking and listing to the demand ramp, not the peak
- Look beyond sales trends: average price, average reviews, average rating, 1/3/6-month new-product share, retail share, FBM share — all have trend curves; signals are more reliable when multiple dimensions improve together; cross-validate with category comparison
- Historical teaching case: in May 2020, Pool Safety Covers demand was just lifting off-season (963 units/day) and was flagged as an opportunity; two years later — January 2021 YoY +78.64%, January 2022 YoY +49.02%
- Live API measurement (2026-11-02, Covers category incl. pool safety covers; sales estimates): strong seasonality — September peaks in both years (326,843 / 335,647 units), January trough (33,284 in Jan 2026), a ~10× peak-to-valley ratio; October 2026 YoY +28.3% (234,836 vs 182,980), yet May and June 2026 YoY were -11.8% and -2.9% — a few isolated months support opposite conclusions, so always read the whole year-over-year curve. Whether demand keeps rising needs further monthly validation; no extrapolation promised
Discover more reference-worthy products, faster and more completely. Six signals, all present:
| # | Signal | Direction |
|---|---|---|
| 1 | Ranking | Stable / rising |
| 2 | Sales | Stable / growing |
| 3 | Price | Stable / increasing |
| 4 | Rating count | Low / catchable |
| 5 | Ad spend | None / low |
| 6 | Time since launch | Short (within 6 months) |
In one sentence: a product launched within six months, with no ads and few reviews, that keeps climbing in rank and sales without cutting price — its playbook is worth reverse-engineering.
Tool form = the Hidden Profit Index (HPI): daily sales + subcategory rank + price as three historical curves, plus review count and ad-spend index; 20,000+ products per marketplace ranked as a full set (with 30-day sales change, review change, review-count coefficient and season flags). The index ranks leads; it is not a profit promise.
Replicability Checklist (mandatory when reverse-engineering, to beat survivorship bias): "sells organically with no ads" is an outcome, not a cause — it may come from off-platform communities, influencer deals, an established store's weight, exclusive supply, or loss-leading clearance. Check each:
- Traffic structure: on-platform ad share is only the tip — probe off-platform traffic, branded-search volume, deal events
- Store background: a new store with a new listing, or an old store with a new listing (store weight carries over)? A multi-store matrix?
- Price & promotion history: is the stable rise actually price-hiked clearance, big coupons, or off-platform flash sales?
- Supply replicability: can you source the same spec at the same cost? Exclusive molds / patents / certifications?
Public data holds no alpha: the 18 criteria and the 6-in-1 signals are visible to everyone — you can screen by them, and so can your competitors. Data eliminates mines; it does not create unique opportunities. Outperformance lives in supply chain, product strength and execution speed. A good category is a moving game — by the time you enter, it may already be deteriorating, so leave room to iterate.
Content first, data-validated: evaluate like a "director" whether the product can star in viral content, then use data to confirm the market is worth entering. Default robust path: if you can validate demand and profit on Amazon first and use TikTok as an amplifier, prefer that. For TikTok-native selection, follow the four steps below with stricter validation (data coverage, policy and fulfillment risks in §7).
Content genes (primary signals — the more the better; proceed only with at least two):
- Visual impact (Wow): unboxing / usage looks striking within 3 seconds
- Satisfying / therapeutic feel (ASMR): cleaning, organizing, ASMR-style emotional payoff — the process itself is watchable
- Dramatized problem (pain point): strong before/after contrast; the pain is amplified and solved on camera
- Aesthetics / design (beauty): colors, texture and scene fit that support premium-looking footage
Virality genes (secondary signals, bonuses): 5. Topic extensibility: supports ≥3 content angles (use cases, tutorials, reviews, skits) 6. Creator fit: matches ≥2 mainstream creator archetypes (lifestyle, beauty, tech, parenting, home…) 7. Music / rhythm fit: pairs naturally with trending sounds (beat-sync, transitions, reveals) 8. Replicable formulas: proven content formulas exist in the category ("immersive X", before/after, "365 days of Y") and are not yet over-consumed
Core question: can this product sustain a stream of 3-second-hook videos that also drive purchase? Evidence sources: shoppable-video dashboards, video tags, daily/weekly/monthly viral-video charts, trending content in the category.
Confirm the content-driven hotspot sits in a structurally healthy market that still admits new players:
- Demand: TikTok search volume and growth, product sell-through and video-driven sales
- Structure: beyond seller/brand concentration, add creator concentration — is volume monopolized by a few creators or viral videos? (If so, entrants need comparable creator resources)
- Sustainability: is the hotspot a one-hit wonder or does it keep producing new sellers? Do AOV, conversion and return rates support profit? Low AOV (roughly under $15) is usually unprofitable under commission + samples + boosting costs — be cautious
- Quantified proxies (instead of gut feel): product sell-through velocity (monthly sales, cumulative sales, days since launch), creator count, shoppable-video count, rating and review count, price. Creator and video data currently cover the US only — for other marketplaces, supplement with manual front-end sampling; do not force-fit
- Fit dimensions: creator persona and content tone, audience profile (age / gender / region / spending power), historical commerce performance (category, AOV, conversion), engagement rate and style
- Portfolio approach: never bet on one big creator; test "multiple creators of the same archetype × multiple content angles" in small steps and iterate on actual funnel results (views → clicks → orders); watch for sample abuse and fake engagement, and budget samples into your costs
- Core question: who should sell this, in what content format, to which audience?
- Translate selling points into camera-ready selling points: effects that can be shown beat specs that must be explained
- Produce assets from 2–3 content formulas; the first 3 seconds decide everything; test multiple assets on small budgets and scale the winners
- Keep organic reach and paid boosting on separate ledgers to see whether the product can sell without pure paid traffic
Directional signals, no hard lines:
- Sustained organic sell-through (not driven by pure paid traffic or a single viral post)
- Stable/growing sales contributed by multiple creators and videos (not one source)
- Stable pricing, no downward price spiral
- Few reviews but strong sell-through (catchable for entrants)
- New content formulas still emerging in the category (content dividend not exhausted)
- Product/shop recently launched and inside a growth window
| Product | Price | Monthly sales | Cumulative / launched | Rating | Creators / videos | Reading |
|---|---|---|---|---|---|---|
| Telescopic long-handle crevice brush | $12.99 | 11,047 | 17,963 / 2026-09-20 | 3.7 | 18 / 43 | ~6 weeks old with ~18k units sold; a 3.7 rating did not block sales — content persuasion replaced the review moat |
| Electric makeup-brush cleaner | $9.97 | 3,890 | 14,275 / 2026-03 | 4.1 | 298 / 417 | A textbook creator matrix: ~300 creators and 400+ videos contributing |
| Hair-brush cleaning tool (old listing) | $19.50 | 275 | 6,765 / 2023-09 | 4.9 | 107 / 168 | An old listing still selling on content's long tail — content assets compound |
| USB makeup-brush cleaner (control) | $4.65 | 128 | 156 / 2026-05 | 3.7 | 1 / 1 | Same keyword, near-zero content investment = near-zero sales; cheap doesn't save you |
Takeaway: in interest-based e-commerce, content and creator investment IS the traffic entry — cross-validate with "creator count × video count × sell-through velocity". Viral cases still need the Replicability Checklist (sample scale, paid-traffic share, single-post dependence), and lifecycles are short — price in the creative-fatigue risk from §7.
- Amazon → TikTok: within categories you know, reverse-lookup TikTok's fast-growing products and trending content, and reuse proven content formats for off-platform traffic and new-product testing, reducing dependence on on-platform ads
- TikTok → Amazon: re-validate TikTok-proven hits in Amazon data — search volume, competitive structure, review moat, profit headroom. Impulse buying ≠ search-driven repurchase: audiences, price bands and purchase motives differ, and hit migration usually fails. Validate strictly; this is not a mandatory pipeline
- The synergy loop (ideal form): content-based testing → data validation → brand building → multi-channel expansion. Each step stands on independent validation; never skip a step because the previous one succeeded
- Walmart: optimal price bands typically run ~10-15% below Amazon's; watch 1P retail share and WFS penetration; ad data is coarser than Amazon's, so leave margin for error; for cross-platform price gaps, weigh absolute profit and two-platform sales, not the gap percentage; mind pricing and hijacking rules
- Shopee: Southeast Asia is more price-sensitive with lower price bands overall; evaluate per site
- Temu / AliExpress / Mercado Libre: fulfillment and pricing mechanics differ greatly — confirm your operating model (managed / self-operated) before applying the validation dimensions
- Same product across platforms: identical appearance ≠ identical supply chain; verify cost and specs one by one; check brand authorization and platform rules first
Run the operating calculation at selection time, with traffic acquisition included: target sales + COGS + first-mile logistics + fulfillment & commission + return rate + traffic cost × expected conversion, yielding your affordable promotion ceiling (shelf: recommended ACOS control ceiling; interest-based: contribution per order and break-even after creator commission + samples + content + boosting). Where gross margin cannot cover promotion cost, more sales can simply mean more losses.
Three calculation rules (+ one measurement rule):
- Three-scenario sensitivity, never a single-point ROI: model conversion, CPC/commission and return rate under pessimistic / neutral / optimistic cases; ACOS above the ceiling during launch is the norm, not an accident — list ramp-up losses separately
- Cash flow before margin: with first MOQ + 45-60 days of first-mile lead time, compute months to positive cash flow and annual inventory turns — the same ROI at 2 turns vs. 6 turns is two different businesses
- Stop-loss before investment: maximum acceptable loss and the exact line that stops you, written down at selection time, with an exit plan (see §8)
- Measure fees, don't estimate them: recompute fulfillment/FBA fees from actual dimensions and weight — sourcing-page estimates run low; watch fee-tier cliffs: at $19.99 a higher fee tier can leave you less profit than $17.99, so price against the tiers
Live API measurement (2026-11-02, Amazon US, "iphone 17 case"): suggested CPC rose from $0.91 ($0.77-1.14) in 2025-03 to $2.04 ($1.49-2.80) in 2026-10 — ~2.2× in 19 months. Search demand grew over the same period (~9.49M yearly searches in 2025 → ~13.25M in 2026), but traffic cost grew faster — demand growth ≠ opportunity; the cost curve decides survival. First-page ad slots: 636 vs. 215 organic slots (~75% ads); estimated first-page ad spend ≈ $1.45M (API estimate).
Historical teaching cases:
| Case | Paper gross margin | Traffic-cost reality | Calculated result |
|---|---|---|---|
| iPhone cases (2022) | 51.84% | CPC $1.8, 100% ad-attributed sales | Monthly loss $11,790, ROI -62.38% |
| iPhone 13 cases (2024) | $9.99 price looks profitable | CPC $1.5, 80% ad share, ads = 62.6% of cost structure | -$6.98/unit, monthly loss $132,243, ROI -55.74% |
| Chin strips (three pricing scenarios) | 43.78% | $16.95 vs. $13.95 pricing; ad-attributed share 100% → 60% (as organic orders grow) | ROI 48.77% / 14.77% / 56.50% |
Every calculation is an estimate built on assumptions: sourcing price, conversion, CPC/creator commission and return rate can all drift — calibrate progressively with supplier quotes and small-batch tests.
Risks are flagged, never used to eliminate options on the seller's behalf; no high index score exempts a category's inherent risks. Four reference levels:
| Level | Typical categories | Risk substance |
|---|---|---|
| 🔴 Compliance & certification | Food / supplements / medical devices / baby food / pesticides / alcohol / functional cosmetics | Certification barriers; requires professional teams and compliance investment |
| 🟡 Capital & returns | Apparel / shoes & bags / jewelry / furniture / large appliances | Return rates up to 5-30%, size-variant complexity, inventory capital pressure |
| 🟠 Operational complexity | Electronics / liquids / flammables / licensed IP / auto parts / fragile glass | Certifications, hazmat logistics, infringement, breakage and after-sales |
| Phone cases / screen protectors / seasonal decorations / books | Extreme competition, thin margins, seasonal dead stock, review anomalies |
Store-level risks (check separately): category gating, brand registry and hijacking limits, patent TROs, account linkage, review-manipulation red lines. In some categories, white-hat-only ramp-up takes very long — price "time-to-rank without gray tactics" into the decision instead of assuming it works.
Two additions for interest-based e-commerce:
- Creative-fatigue risk: viral formulas have short lifecycles; creative and creator performance decays. Budget for continuous content investment; never extrapolate a full year from one viral hit
- Platform policy risk: TikTok's US policy outlook is uncertain, and creator/video data is US-only — assess platform dependence before going all-in and keep multi-channel exits open
Supporting actions: design-patent and IP clearance upfront; use review velocity to distinguish organic growth from anomalies; verify fulfillment fees and return rates before trusting demand.
① Discover high-potential references (shelf: 6-in-1 products; interest: content potential × selling products)
→ ② Validate competitive difficulty → ③ Quick operating calculation (sales + costs + fulfillment + returns + traffic, 3 scenarios)
→ ④ Contact suppliers, confirm specifics → ⑤ Detailed calculation (cash flow / turns / stop-loss)
→ ⑥ Collect and analyze top competitors' / viral contents' elements
→ ⑦ Build product + listing/content → ⑧ Post-launch monitoring and exit execution
From battlefield selection to weapon building, two paths:
- Differentiate the product itself to serve real consumer needs better. Examples: a commoditized phone case rebuilt as waterproof + stand + cooling + drop-proof; Bluetooth earbuds designed for "one pair, two listeners"; a body-fat scale built around sensor accuracy + app + multi-user packaging
- Zero/low-cost micro-innovation on commoditized products: sharpened selling points + presentation quality (shelf: main images / scene images / pain-point visual storytelling; interest-based: first-3-second visuals and content formulas) to outclass peer listings
The exit mechanism (equal weight with entry) — every "suggested entry" must ship with:
- Exit triggers: e.g., sales below a set fraction of projection at day 60/90, ACOS persistently above ceiling without improvement, rating below category tolerance, cash flow hitting the stop-loss
- Liquidation plan and cost: clearance channels (markdown / outlet / bulk liquidation / removal & disposal) and estimated clearance cost — dead-stock handling can exceed the first order's value, so price it in at selection time
- Feedback loop: exit reasons feed back into the Replicability Checklist and the next round's screening parameters
Three seller challenges: understanding the platform (values, rules, methods); judging the competitive landscape and rivals; uncovering real consumer needs. The operating goal: select and build differentiated products that serve real needs, aiming to keep earning platform and consumer recognition without sacrificing your target price — no price wars; defend price and weight through differentiation.
| Methodology step | Sorftime tools (MCP tool names / CLI endpoints) |
|---|---|
| All-category comparison |
category_search_from_top_node + category_report (monopoly coefficients / new-product share / reviews / CPC / return-rate field groups) |
| Single-category Top-100 panorama |
CategoryRequest + ProductSearch (sales distribution / review ranges / launch times / rating distribution); category_report history via query_date (up to 40 days) |
| Multi-year & dimension trends |
CategoryTrend with 40 indices (0=sales, 3=avg price, 6=new-product share, 9=retail share, 28=Top-3 Listing monopoly, 34=Top-10 brand monopoly; loop and stitch into time series) |
| 6-in-1 hidden-profit products |
potential_product (Hidden Profit Index; no hard filters before ranking) + product_trend (SalesVolume / Price / Rank, 6 months) + product_traffic_terms (organic vs. ad exposure) |
| Quick operating calculation | Profit calculator + keyword CPC (KeywordRequest, with CPC trends / slot breakdown / per-page estimates) + ali1688_similar_product (Chinese keywords; apply a ×1.3 correction to the lowest listed price to offset bait SKUs; reference only) |
| Competitor element collection |
product_content (main images / bullets / A+ / buyer focus) + product_reviews (sample by star bucket) |
| IP clearance |
patent_search_from_keyword / patent_search_from_image / patent_request
|
| Cadence & post-launch monitoring |
BestSellerListSubscription / KeywordBatchSubscription (rank and keyword monitoring with configurable schedules); 30/60/90-day reviews (re-evaluate on large sales deviation, pause above ACOS ceiling, investigate rating anomalies, reorder on low stock, execute exit on trigger) |
| Methodology step | Sorftime tools (CLI endpoints; same capabilities on the MCP TikTok endpoint) |
|---|---|
| Market & category validation |
CategoryTree / CategorySearchFromName / CategoryRequest / CategoryTrend (category structure, volume and trends) |
| Product discovery & validation |
ProductSearch / ProductSearchFromName / ProductRequest / ProductTrendRequest (products and sell-through trends; fields include monthly/cumulative sales, launch time, rating, creator count, video count) |
| Shop / seller structure |
ShopSearch / ShopSearchFromName / ShopRequest (seller concentration and shop performance) |
| Content potential & formulas |
VideoRequest / VideoSearchFromName / VideoTagSearch (shoppable videos, tags and viral content; US only) |
| Creator matching |
AuthorRequest / AuthorSearchFromName (creator commerce data and profiles; US only) |
| Brand structure | BrandSearchFromName |
- Walmart: category / product / keyword and sales-trend endpoints; pair with Amazon for cross-platform price-gap and validation work
- Shopee: category / product / shop / keyword endpoints; evaluate per site
- 1688 (sourcing, universal): image search for identical products, price and spec verification (apply the ×1.3 bait-price correction; reference only)
- Social demand listening (universal): Reddit / Quora post and comment mining for genuine off-platform demand and emerging trends
- 🆕 Closed-Loop Selection ← Full pipeline
- 📖 Case Study: Water Bottle ← Real data example
- Loop & Goal Templates ← 58 automation recipes