Skip to content
CodeTrendy
Submit
Back to stats

How we count.

The numbers on the stats page are computed directly from our database. No vanity multipliers, no since-launch totals dressed up as monthly. Here is exactly what each one means.

  1. 01

    Listing impressions

    A listing impression is one server-recorded view of a listing detail page or a homepage card after the visitor stayed on the page for at least 2 seconds. Every event is counted, including repeat views from the same visitor.

  2. 02

    Outbound clicks

    An outbound click is recorded when a visitor clicks a listing link that takes them off-site to the listed product. This is the metric sponsors care about because it represents actual traffic delivered.

  3. 03

    What is included

    The numbers on this page reflect every server-recorded view and click within the 30-day window. We do not subtract crawler traffic, scrapers, or automated browsers - if a request hit our servers and was recorded, it is in the total.

  4. 04

    Test users

    Internal test accounts (User.isTest=true) never appear on the page. They are excluded from builder counts, votes, reviews, and signups.

  5. 05

    CTR (click-through rate)

    For each listing with at least 5 views in the window, CTR is clicks divided by views. The headline figure is the median across those listings, not a global average, so a single viral listing cannot skew the number.

  6. 06

    Refresh cadence

    Every aggregate is cached for 1 hour. The live visitor strip and the last-24h visitor count refresh every ~12 seconds via a small client poll.

  7. 07

    Geography

    Country is captured from the request edge headers at write time. Rows with no country are excluded from the country breakdown but counted in totals.

  8. 08

    Devices and referrers

    Devices are parsed from the user-agent string at read time. Referrers are grouped into search / social / dev / direct buckets based on the source hostname.

  9. 09

    Engagement (votes, reviews, signups)

    Counted from the votes, reviews, and registration_logs tables. All three are filtered to exclude test users and failed signups.

Question we missed?

If something on the stats page does not look right, or you want to verify a number before sponsoring, reach out. We are happy to walk through the SQL.