How Growtsy decides

Live market signals, per analysis

When you analyze a product, Growtsy runs its own market pipeline against real Etsy search results for that niche. What comes back:

• Search result composition — who ranks for the buyer phrases your product would target, and how crowded that page is.
• Tag & keyword frequency — which phrases top-ranking listings actually use, so tag suggestions come from evidence, not brainstorms.
• Price distributions — the real price band of the niche: median and favorite-weighted reference points, from listed prices.
• Review counts, shop age and favorites where exposed — the maturity signals Etsy makes public.
• Time-windowed review counts — how many new reviews the leading listings collected in the last 30 and 90 days, pulled per listing.
• Shop-level lifetime sales & shop age — public shop facts, for reading who dominates a niche and whether newcomers still break in.

Demand and competition are inferred from these signals — from how a results page is composed and how fast reviews accumulate. We name that inference instead of dressing it up as measurement.

If a live pull fails, Growtsy says so in the result and falls back to category baselines — clearly labeled as estimates. We never present a fallback as fresh data.

Reviewed purchases — our proven-sales floor

Every Etsy review requires a completed purchase. So when the leading listings in a niche collected, say, 132 reviews in the last 90 days, that is at least 132 real sales — counted, not modeled. Not every buyer leaves a review, so true totals run higher; we say exactly that and stop there. No multipliers, no conversion factors, no "estimated monthly sales." The floor is the claim.

Titles

Etsy allows 140 characters, but search results display roughly the first 50–70. So Growtsy front-loads the strongest buyer keywords instead of padding to the limit, and uses comma-separated phrases that mirror how Etsy parses queries. A shorter, front-loaded title beats a stuffed one.

Tags

Etsy gives 13 tag slots, each capped at 20 characters, and treats each tag as a separate chance to match a search — so all 13 are used. We prefer multi-word (long-tail) tags because they catch specific buyer intent with less competition, and we avoid repeating title words verbatim: Etsy's search has devalued repetition since its relevancy updates.

Descriptions

The first two lines appear before “read more” and are indexed by Google — so the intro sells in buyer language before specs and policies. We deliberately avoid keyword stuffing: Etsy reads descriptions for relevance, not density.

Prices

Charm pricing (ending in .99) is applied for conversion, but the value itself is positioned inside the niche's real price band from the live pull. When a suggestion sits far below the market reference, Growtsy flags the gap instead of silently overriding — a differentiated piece can legitimately price above the band. You decide; we show the evidence.

Photos & gallery

Etsy shows up to 10 images and the first one wins or loses the click. Growtsy's gallery slots each have a distinct job — hero, lifestyle, scale, detail — and listing-gallery images must represent the actual product faithfully. Scene and context can be generated; the product itself is preserved. That line is an Etsy policy line, and we build on the safe side of it.

Timing

Publish-time guidance follows buyer-market windows: ~74% of Etsy buyers are in the US, whose evening peak is 19–23 ET, with secondary UK/EU and AU/NZ windows — converted to your timezone in the Calendar. These are static market-structure facts, not per-listing predictions.

No per-listing sales counts. No revenue estimates. No magic scores.

Etsy's public data does not expose how many sales an individual listing makes. Any tool showing you “this listing earns $X/month” is estimating — sometimes wildly. Growtsy refuses to do that. The only sales-adjacent number we show is the review floor above: real, counted reviews framed as a minimum — never an estimate.

It's also why Growtsy gives you decisions instead of scores: “Publish now”, “Improve before launch”, “Fix visuals first”. A fake-precise 87/100 feels scientific but hides the uncertainty; a directional verdict grounded in named evidence respects it.

What you'll always get instead: the signals we read, the rule we applied, and the single highest-leverage next action.

Honest limits

• AI can misread a product photo — that's why verdicts show their reasoning and you can correct the product identification.
• Market signals are directional, not guarantees. A great listing in a validated niche can still take weeks to sell.
• Etsy changes its rules; when it does, we update the encoded rules — this page is the single source they trace to.