Thirty issues in thirty-one days. The only gap was Sunday, August 2. Each issue carried four to six stories, 149 in all, drawn from 90 different publications across all eight beats. No outlet supplied more than ten stories in the month, so the source-monoculture failure from July did not recur.
The machinery changed under the newsletter as well. The batched send path fixed the roughly 7% of confirmed readers who had been silently dropped from each issue by rate limiting; the August 7 issue reached 196 of 213 confirmed readers; by August 21 the issue reached 339 of 339. On August 20 the desk-editor gate, the corroborator and the proofreader went in after the Coty direction error, and every issue from the 21st passed through them. The deals cap tightened to one per brief the same day.
The data side caught up too. This export is the first with per-issue story attribution: every story, the slot it ran in, and the clicks that issue’s email produced, including the ones nobody clicked. That is what makes the beat comparison below trustworthy where the July read was not.
The confirmed list stood at 200 on July 31 and 385 on August 31. That is 211 new confirmed readers against 27 who left, a net gain of 185, or 92% in a month. Every retained sign-up is confirmed; the table keeps no pending rows, so anyone who abandoned the double opt-in is not visible here.
Growth was not even. The second week of August brought 86 sign-ups, more than any other week, with 56 of them landing on August 8, 9 and 10. Sign-ups arrive before the reader has seen an issue, so that surge was driven by something outside the newsletter. The data does not say what.
Almost all of it came through the landing page: 208 of 211. The referral programme produced three. Geography, tracked from August 8, is led by the United States, then Canada, the United Kingdom and Australia, with a scattering across India, France and Germany.
Two out of three deliveries were opened all month, and the rate drifted from 64% in the first week to 59% in the last. Late July ran at 71%. That is the expected shape: the first two hundred readers were the closest to the brand, and each new cohort is a little less committed than the last. Opens are pixel-based and a floor, not a count.
Clicks tell the better story. Story clicks per issue went from 4.2 in late July to 9.0 in August, and the click rate held between 2% and 3% for four of the five weeks. For a brief designed to be read in place, a low click rate is the format working, not failing. Readers open, read, and occasionally follow a link to the source.
Loyalty is high among readers who were there all month. Of the 371 people who opened at least one issue, 144 opened twenty or more of the thirty, and 81 opened twenty-five or more. Clicking is a minority habit: 90 people clicked a story in August, one in four openers, and the ten most active of them account for half of all clicks. One reader clicked 30 times.
New readers read hardest. The welcome copy sent on confirmation was opened by 76% of recipients and drew 53 clicks from 189 deliveries. A regular issue draws about 3 clicks per hundred deliveries; the welcome issue drew 28.
Science & Ingredients ran 18 times, 12% of the stories, and took 21% of the clicks. Its index of 171 is the highest of any beat, and it rises to 200 when only slots 2 to 4 are counted, so the lead is not a placement effect. It led only two issues all month. Regulation is second at 127, as it was in July.
Supply Side looks strong at 121 but led four issues; within slots 2 to 4 it drops to 69. Beauty Tech is the quiet performer: 119 overall and 152 in the middle slots on only 12 stories. Deals & M&A moved from 70 in July to 97, average, after the cap cut it to about one story a day.
The Consumer was the most-published beat, 27 stories, and sits near the bottom at 69. Retail & Media is last at 63, as it was in July. Market Watch is at 76. Between them these three beats supplied 22 of the 39 stories nobody clicked.
The ten stories readers followed to the source
Position is worth about a click. The lead story averages 2.3 clicks; slots two to four sit at 1.9; slot five drops to 1.3 and slot six to 0.6. This is the confound the July read could not separate from beat. Now it can, and the beat finding survives it.
The practical reading is about issue length, not order. Six-story issues added a sixth slot that drew four clicks across seven appearances. Five stories looks like the ceiling of what gets read to the end.
Source tier, again, predicts nothing. Tier 1 stories averaged 1.89 clicks, Tier 2 averaged 1.86. Tier 3 averaged 1.08 on twelve stories. The tier floor stays an accuracy rule.
Where the clicked stories came from
The APL Dispatch is a daily brief on the business of beauty.
Five stories a morning, read by people in the industry at the end of . Subscribe at thedispatch.atomicpomlabs.ca.