Metrics
Everything this page shows about pdoom1.com’s traffic is read out of the
daily snapshots the site already keeps in git. No number here is typed by hand,
and none is estimated.
The counts are small. They are also real people, and we would rather publish
them from the start than wait until they flatter us — a metrics page that
appears once the numbers look good is marketing, and one that starts small in
public is evidence.
The last 30 days
Reading the latest snapshot…
—
Visitors
Distinct people, de-duplicated across the whole
window. This is why the daily numbers below cannot be added up to
reach it: one person visiting on two days is one visitor here and
two there.
—
Pageviews
Pages loaded. Higher than visitors when people
read more than one thing.
—
Visits
Sessions. A returning visitor adds a visit, not
a visitor.
—
Average visit
Short. Most people load the homepage and decide
from it, which is what the page is for — but it is not the
number of someone who read the thing.
—
Bounce rate
The share who saw one page and left. Published
because the unflattering numbers are what make the flattering ones
worth believing.
Every day we can prove
Reading the snapshot index…
Pageviews
Visitors
No data recorded
Snapshots disagree
Show the same numbers as a table
Where people came from, and what they read
Same 30-day window as the tiles above. Referrer names come from visitors’
browsers, so they are external text and are escaped before display.
What we track, and why
Analytics is Plausible, self-hosted on
our own box at analytics.pdoom1.com. Not the hosted cloud, and
not Google. It sets no cookies, builds no cross-site profile, and stores no
identifier that can follow you between sessions.
We track it because a small project needs to know whether anything it makes
reaches anyone. That is the whole reason. There is no advertising here, no
list to sell you to, and nothing on this page is personal data — every
figure is a count of events, not a record of a person.
- What is collected: page path, referrer source, country,
and whether a download or outbound link was clicked. Aggregated before it
is ever written down.
- What is not: no cookies, no IP addresses at rest, no
cross-site tracking, no individual profile, no session that survives the
day.
- Do Not Track is honoured, and you can opt out
explicitly from the privacy page. That button
drives the one flag the tracker actually reads, and it persists.
- The archive is public. These snapshots live under
public/ in a public repository, so anyone can check this
page against its source rather than take our word for it.
Where these numbers come from
A GitHub Action asks the Plausible API for the last 30 days once a day and
commits the answer to
public/data/analytics/history/. That exists as a backup —
the analytics box has no other one — and this page is a second reader
of the same files. It fetches
/data/analytics/latest.json and
/data/analytics/history/index.json, and nothing else. There is
no live API call from your browser to the analytics server.
What this page will not do
- It never fills a gap. A day no snapshot recorded is
drawn as a gap and named as one. A recorded zero is a different fact from
a missing day and stays a different fact.
- It never remembers. If a file fails to load you get an
error saying so, not the last number we happened to know.
- It never adds windows together. Consecutive snapshots
overlap by 29 of their 30 days, so summing them would multiply the truth
by about thirty.
The snapshots themselves
Reading the snapshot index…
Show every snapshot taken