The journal is where this magazine reacts in public — to results pages that answer more of the question themselves, to machines that cite sources inside generated answers, and to the quiet renaissance of channels publishers actually own. Where the guides are the method, the journal is the weather report.
Recent notes
- Zero-click search, and why it isn’t the end — the click economy is shrinking at the margin and concentrating everywhere else; what that changes for publishers.
- AI Overviews: traffic when the answer is generated — being the cited source is the new blue link. How citations are won, and how to measure a visit you cannot fully see.
- The newsletter as a traffic engine — the one channel where the algorithm works for you, run as infrastructure rather than as a hobby.
How the journal differs from the guides
A word on addresses, since this magazine is strict about them: every note keeps its URL for life, and updates land on the same address rather than spawning a successor. If a note is superseded by events, it says so at its own head rather than vanishing — an archive that quietly rewrites itself is an archive nobody can cite. It is a small discipline with a large compounding effect: the notes below are safe to link, safe to quote and safe to revisit.
Three differences of register. Timeliness — a note responds to something that moved this year, a guide aspires to be true for five. Length — notes run a single sitting, guides run an evening. Confidence — a note is allowed to reason in public with incomplete data, which is what a journal is for; the reference shelf holds back until the evidence is boring. Corrections to either are handled the same way, per the policy on the about page.
Notes on the desk
Two notes are owed and herewith declared. First, on measurement honesty: when a note cites an industry pattern rather than our own observation, it says so — “across the datasets we have examined” means exactly that, no more. Second, on absence: the journal has not yet written the pieces this desk most wants to read — a proper treatment of image and video search as traffic carriers, and a reckoning with the economics of programmatic page generation. They are in progress, they will arrive when they are honest enough to publish, and declaring them here is our way of ensuring they do.
From the desk
A note on register, since the journal licenses itself to reason with incomplete data: every observation here is dated by nature, and we would rather publish a careful “this is what we see, this is what would change our mind” than wait for a certainty that arrives after it is useful. When a note’s reasoning is later confirmed or contradicted, the update lands on the same address — our version of the corrections policy on the about page. Nothing is quietly rewritten; that is the whole difference between a journal and a content farm.
The standing questions
Every note in this section circles three questions that will define the decade for anyone who grows websites. What is a click worth when the results page answers more of the query itself? What is a citation worth when the answer is generated by a machine that must name its sources? And what is an owned reader worth when every rented audience gets more expensive? The current answers live in the notes above; the honest position of this desk is that the questions will outlast every answer we publish — including these.
The journal feeds the weekly briefing; everything that appears in the email lands here afterwards. Frequency, for the curious: roughly a note a month, each one written when the observation has ripened rather than when the calendar insists. The British market notes live with the UK edition, which keeps its own rhythm around the same standards — and the field notes that anchor the whole operation sit at their long-standing addresses: activities, vehicles and costs.