Privacy Policy

No third-party requests, by default

Every font, image, script and stylesheet on this site is served by us. There are no advertising networks, no social-media widgets and no embedded video players. Opening a page here does not, by itself, tell any other company that you did.

Article cover images are drawn by the site itself rather than fetched from an image host, for the same reason.

The one exception is described in full below, under “Analytics cookies”: if you accept optional cookies, Google Analytics loads from Google’s own servers, and your reading of this page is then reported there too. Nothing loads before you choose, and nothing loads at all if you decline.

What your browser stores

Two small things are stored on every visit, before any choice you make, and neither identifies you:

• Your cookie choice itself (essential, analytics), so we do not ask again on every page.
• A single calendar date, compared against itself the next time you open a page, used only to tell whether today is the first time this browser has been counted today — see “Counting page views” below. It never leaves your browser as anything but a yes/no answer, and it never becomes an id.

A session cookie is issued only to signed-in staff using the editorial panel, and only for the duration of that session.

If you accept analytics cookies, Google then sets its own cookies in your browser under its own policy — see “Analytics cookies” below.

Server logs

Our web server keeps ordinary access logs: the address requested, the time, the HTTP status, the browser’s user-agent string and the IP address the request came from. These exist to keep the site running and to investigate abuse. They are not used to build reader profiles and are not shared with anyone.

Newsletters

If you sign up for a newsletter we store your email address. That is the one thing on this site that we keep about a reader, and only if you ask us to.

Alongside the address we store: which desk letters you chose, whether the subscription is active, the times it was created and cancelled, which form you used (the front page or the sign-up page), and a random token that unsubscribes the address. We do not store your IP address, the pages you read, whether you opened a letter, or whether you clicked anything in it. Our letters contain no images and no tracking links, so there is nothing in them that could report back.

What you receive. Everyone who signs up gets the daily letter: each evening, at about 18:00 East Africa Time, the stories we published that day. On a day with no new stories we send nothing. Desk letters, which you can choose on the sign-up page, are sent separately and only when a desk has something to say.

Signing up is immediate. There is no confirmation step: you are subscribed the moment you submit your address, and we send one welcome message with a link to leave. That message is also the safeguard for a mistake or for someone typing an address that is not theirs — it reaches the real owner of the address, with the way out in it. If you have already left, signing up again does not re-subscribe you until you open a confirmation link we send, so nobody else can put you back on the list.

Leaving takes one click, from a link in every letter, with no password and no account. We keep the record that the address was subscribed and asked to stop — that is how we can show we had permission, and it stops the address being added again by accident. Ask the editors and we will erase it outright.

The list lives on our own server, in the same database as the articles. It is not shared, sold, or given to an advertising or analytics company. The letters are sent through a mail provider, Resend, which necessarily handles your address in order to deliver to it and keeps its own delivery records, as any mail service does. We do not turn on its open or click tracking.

Counting page views

We count how often each story and each page is opened, on our own server. This is the one thing we measure about reading, and it is deliberately the crudest possible version of it.

When a page opens, your browser sends us three things: whether it was a story or another page, which one, and the address of the site you followed a link from, if you came from one — the site only, such as example.com, never the page on it, the search you typed or the message it was in. Arriving directly, or from within this site, sends nothing at all.

We also note two things about the request itself: the country it came from, and whether it was a phone, a tablet or a computer. The country is supplied by the network that sits in front of this site; we do not read or keep your IP address in order to get it. The device is read from the line your browser announces itself with, which we use and then discard.

That is added to a running total and then it is over. What we keep is a tally per day: this story, that day, this many views. There is nothing in it that is about you. No IP address, no browser fingerprint, no identifier of any kind, not even the time of day — and no row for your visit that could later be linked to another. Two people reading one story and one person reading it twice produce exactly the same record, because the only thing recorded is that the number went up by two.

The counts are kept apart, and that is the important part. The country tally, the device tally and the story tally are separate running totals of the same day. They are never written down together, so we cannot ask which story was read from which country, or what device read a particular article. Those questions have no answer here — not because we decline to ask them, but because the answer was never recorded.

So we cannot say how many people read us, only how many times a page was opened; we cannot tell whether you came back; and we cannot follow anyone from one story to the next. Those are not gaps we intend to close later. They are what is left when you remove everything that would have made it possible.

We do read your browser's user-agent string on the way in, to drop obvious crawlers from the count. It is not stored.

Readers, added 2026-09-24. We also keep one more number: roughly how many different browsers opened the site each day. Your browser keeps today’s date to itself and compares it on every page it opens; only the answer — already counted today, or not — is sent to us, once, the first time each day. We call this “readers” because it is the closest thing to one we have, and it is still not a person count: clearing your browser’s storage, reading in a private window, or reading from a second device each look like a new reader to us, and nobody who returns after more than a day is recognised as the same one.

So this counter still cannot say how many people read us, only how many times a page was opened and roughly how many browsers opened it; it cannot tell whether a specific person came back, and it cannot follow anyone from one story to the next. Those are not gaps we intend to close in this counter. If you accept analytics cookies, Google Analytics measures some of this instead, under its own rules — see “Analytics cookies” below.

Analytics cookies

One more tool can run on this site, but only if you say yes in the cookie banner shown on your first visit. Nothing below loads before that, and you can change your answer at any time from the “Privacy choices” button in the page footer.

If accepted, Google Analytics loads from Google’s own servers and reports your visit there: pages viewed, session length, device, and location at roughly city level (from your IP address, which Google receives directly — unlike the view counter above, which never reads it). This is a separate, more detailed measurement than the counter above, and it is governed by Google’s privacy policy, not this one.

Why this exists. We use it to understand which stories and sections readers actually find, so editorial and technical decisions are informed by more than guesswork. We would rather tell you plainly what it does than bury it, so the choice you make is a real one.

If reading this site could carry risk for you — because of where you are or what you read here — rejecting this is the safer choice, and the site works exactly the same either way. The view counter above still runs regardless of your answer here; it is the one built specifically so that saying no costs you nothing.

To withdraw consent you already gave: open “Cookie settings” and choose Reject. That stops new data from being sent, but does not reach into Google’s own systems to erase what it already has — that request goes to Google directly, using the link above.

No advertising or marketing tool runs on this site today. If that changes, this section changes in the same update, not afterwards — the rule this page states about itself.

What we do not have, without your consent

By default, and always for the counter described above, we have no record of how long you stayed on a page, how far you scrolled, what you clicked, or which stories the same reader opened. If you accept analytics cookies, Google Analytics separately measures some of this — session length, and by default the scroll, outbound-link and search events its automatic “enhanced measurement” feature collects — under Google’s own policy, not ours.

We have no accounts for readers, and no feature on this site links a reading history to a specific person. There is nothing to register for and nothing to log in to — a newsletter subscription is an address on a list, not an account.

If you write to us

Messages sent to the newsroom are kept in the mailbox they arrive in, for as long as the matter is live and afterwards as a record of corrections and right-of-reply requests. We do not add correspondents to any list. See the contact page.

Your rights

You can ask what we hold about you, ask for it to be corrected, or ask for it to be deleted. Because we collect so little ourselves, the answer is usually that we hold nothing beyond server logs and, if you have written to us, your message. Requests go to the editors via the contact page.

Data held by Google because you accepted analytics cookies is theirs to answer for, not ours — use the link under “Analytics cookies” above to reach them directly.

Changes to this policy

When what we collect changes, this page changes with it, in the same update — not afterwards.