What this site measures

Two separate things on this site count anything, and they are not the same thing. One is a small piece of code in your browser that tells our server when a page opened or an episode played. The other is the web server's own access log, which every web server on earth keeps whether any script runs or not, summarised once a night. Both are set out below, field by field, and both lists are built from the same tables the code measures with, so they cannot quietly fall out of date.

1. The beacon in your browser

When a page opens, and when you play an episode or click a link out to a podcast app, your browser sends our server one short message. Every message stores exactly these fields and no others:

• ts The date and time, to the second, in the server's own timezone.
• e Which of the events below it was.
• p The page you were on, as its address on this site.
• r The website you came from, as a bare name like google.com or chatgpt.com. Never the page you were on there, never anything you searched for.
• vw Whether your screen is small, medium or large. Three words, not a measurement.
• nav Whether the page was drawn by a fresh load or swapped in as you clicked, which is how we tell whether that part of the site is working.
• v The daily visitor key, described below.
• id On a play, a listening milestone or a click, which episode, app or button it was. Nothing else carries this field.

These are the events it sends, all of them, under the names it sends them under:

• pageview you opened a page
• play you started an episode
• listen25 the episode you started reached a quarter of the way through
• listen50 it reached halfway
• listen75 it reached three quarters
• listen100 it reached the end
• app_click you clicked through to a podcast app
• guide_cta you clicked one of the buttons on a guide page

2. The web server's access log, summarised once a night

Every request this site answers, from a person or from a robot, is written to the web server's access log along with your network address and your browser's own description of itself. That is how a web server works, it happens before any of our code runs, and it is not something a website can switch off. Our host keeps that log for up to 24 hours and then deletes it.

Once a night, before that deletion, a job on our own server reads the log and writes down two things. The first is a set of daily totals: how many pages were read, which pages, which robots came, and how many of the day's requests were exploit scanners rather than readers. The second is a per-request extract, one row per human pageview, kept for 90 days and then deleted. A row in that extract holds:

• t The clock time of the request, to the second.
• v The same daily visitor key, which cannot be reversed and does not survive the night.
• p The page that was asked for, as its address on this site.
• r The website the request said it came from, as a bare name.
• s The status the server answered with, such as 200 or 404.
• u A browser family: chrome, safari, firefox, edge, opera or other. Never the browser's full description of itself.

Your network address and your browser's full description of itself are in neither the totals nor the extract. They are read from the log to work out the visitor key and to tell a robot from a person, and then they go when the log goes.

To tell one visitor reading five pages from five visitors reading one page each, the server needs some way to say "this is the same reader". It builds one by taking your network address, your browser's name for itself and a long random number, and running the three of them through a one-way hash. What it stores is the result: a string of hex that nothing can be turned back into.

The random number is generated fresh at midnight and the previous one is overwritten. So today's key cannot be worked out from yesterday's data, and yesterday's cannot be recomputed at all, by us or by anybody who ever got hold of the files. The same person visiting on two days is two unconnected keys, and that is on purpose: we want to count people, not follow them.

Your network address and your browser's name for itself go into that hash and are not written down by either measurement above. Where they do appear is the web server's own access log, for up to 24 hours, as described in the second half of the section above. Neither the nightly totals nor the 90-day extract carries them.

• No cookies. This site sets none, which is why you have never seen a cookie banner here.
• No third-party analytics. There is no Google Analytics, no Meta pixel and no tag manager. Nothing on these pages reports to anybody but us.
• No tracking across websites. We cannot see where you went before this site or where you go after it, and we have built nothing that could.
• No profile, no advertising audience, no sale. We do not build a record about you, we do not hand one to an advertiser, and we do not sell or rent anything we collect. There is nothing here worth selling.
• No fingerprinting. We do not read your fonts, your canvas, your battery, your timezone or your list of plugins.

If your browser sends a Do Not Track signal, or Global Privacy Control, the beacon described in part one above does not run. Not a reduced version of it and not an anonymous version of it: the code checks before it does anything and then does nothing at all. No pageview, no play, no click, and it does not even attach the listeners that would notice them.

What that signal cannot switch off is the web server's access log in part two, and we would rather tell you that than let you assume otherwise. Your request has to reach the server for the server to answer it, and the server writes a line about every request it answers. So the nightly summary still counts the visit, as one row that holds a one-day key, a page, a bare referring site name, a status code and a browser family. There is no page setting that changes that, short of us not running a website.

The site works exactly the same either way. Nothing here is behind a measurement.

If you sign up for the weekly email, we keep your address so we can send it, and every one of those emails carries a one-click unsubscribe link that works. If you write to us through the contact form, we keep what you sent so we can answer it. Neither of those is shared with anybody and neither feeds the counting described above.

The show is embedded from and distributed by third parties, and they have their own rules: the audio player, the YouTube video on some pages, and the podcast apps you subscribe in. What they collect when you use them is theirs to describe, not ours.

Questions about any of this?

Write to info@homeownersshow.com and a person will answer. If you want anything we hold about you deleted, ask, and we will.