How our data works

Radical transparency. Here's exactly how we collect, process, and score every keyword.

Data collection

Every week, we query Etsy's search autocomplete — the same system that shows suggestions when buyers type in the Etsy search bar. We start with every possible single-character prefix (a-z, 0-9) and follow the tree deeper wherever Etsy returns enough suggestions.

For each prefix, Etsy returns up to 11 keyword suggestions ranked by popularity. Position #1 is the most popular suggestion for that prefix. We record every suggestion along with its position, the prefix used, and the depth in the search tree.

Example: How we map "phone case"

prefix "p" → "phone case" at position #1

prefix "ph" → "phone case" at position #1

prefix "phone c" → "phone case" at position #1

prefix "phone case " → "phone case cute" at position #2

prefix "phone case c" → "phone case cute" at position #1

...and so on, following every branch

Go-deeper logic

When a prefix returns 3 or more suggestions that start with that prefix, we expand it by appending a-z and a space (27 new prefixes). This lets us discover the long-tail: "phone case cute kawaii" lives 4 levels deep in the tree.

Dead branches self-prune. If "zxy" returns 0 suggestions, we stop. The tree is self-limiting — no arbitrary depth caps needed.

Autocomplete position

Position reflects how Etsy ranks that keyword for a given prefix. Position #1 means Etsy considers it the most relevant/popular suggestion. A keyword appearing at position #1 for a short prefix (like "p") is extremely popular.

We track the best position — the highest rank a keyword achieves across all prefixes where it appears. A keyword with best position #1 is in very high demand.

Opportunity Score

score = (11 - position) × depth_weight / listing_count

where depth_weight = 1 + (depth × 0.3)

The formula rewards:

  • Higher autocomplete position (more demand signal)
  • Greater depth (more specific = less competition)
  • Fewer listings (less supply = more opportunity)

A keyword at position #1, depth 3, with only 200 listings scores much higher than one at position #8, depth 1, with 50,000 listings.

Trend detection

By comparing weekly snapshots, we identify keywords that are new to autocomplete, jumping in position, or fading. A keyword appearing at position #3 this week that wasn't in autocomplete last week is an emerging trend.

What this data is NOT

Let's be clear about limitations:

  • Not search volume. Etsy doesn't publish search volumes. Autocomplete position is a proxy for demand, but it's not a count of searches.
  • Not conversion data. We know what people search for, not what they buy. High search volume doesn't guarantee sales.
  • Not real-time. Data is updated weekly. Intra-week trends won't appear until the next update.

We believe honest data with clear limitations is more valuable than fabricated "search volume" numbers. If you've ever questioned where a tool's data comes from — now you know exactly where ours does.

Listing count data

Listing counts come from Etsy's public API (findAllListingsActive endpoint). This tells you how many active listings match a keyword — a direct measure of competition. We update listing counts regularly but they may lag by a few days.