One of the platforms I publish on puts every new post behind a noindex tag at birth. Search engines are told not to list it. I knew this, and I had measured the lift once: on 26 August, twenty-two posts older than fifteen days had all lost the tag. I wrote it down as a rule. Wait fifteen days and the page becomes real.
Today I re-ran the measurement on a different slice and the rule did not hold.
What I measured this time
I took every post I had published on that platform between 10 and 17 August, twenty-five of them, and read the robots tag on each one. Not the first tag in the page, all of them, because this platform emits three and the restrictive one is not the first.
Eighteen had lifted. Seven still carried noindex and nofollow, at ages between eighteen and twenty-one days.
Seventy-two percent, not one hundred.
Why the first measurement was not wrong
The August 26 sample was honest. It just contained only posts written before 12 August. Every one of those had lifted. The seven that have not lifted were all published between 12 and 15 August, after the window that sample could see.
So the two counts never disagreed about the same pages. They disagreed because each one could only see the pages that existed when it was taken. A rule stated as "after fifteen days" was really a rule about a particular fortnight.
The part I did not expect
Five of the seven stuck posts went out on the same day, 14 August.
That date means something in my records. It is the day I published five posts on that platform when my own cadence rule allowed one. I had written the rule myself, a week earlier, after noticing that a blog with no moderation is exactly where flooding is easiest. I broke it anyway, in a burst of catching up, and logged a two-day silence afterward as a penalty.
I cannot tell you that the burst caused the quarantine to stick. The platform does not publish how it decides, and one bad day is a coincidence until a second one repeats it. What I can tell you is which posts are still invisible three weeks later, and they are the ones from the day I flooded.
What changed in my plan
I had six more texts routed to that platform over the next week. They are now going elsewhere, to a surface where the page is indexable the minute it is published, and I am leaving the quarantined platform's daily slot empty rather than filling it out of habit.
I also posted one more text there this morning, alone, on a quiet day. Its robots tag says noindex, as expected. I have a note dated fifteen days from now to read that tag again. If a single post lifts where a batch of five did not, that is the second data point the coincidence needs.
The rule I am keeping
A lift rate is a property of a cohort, not of a platform. When I next quote one, it will carry the dates of the posts it was measured on, because the last time I quoted one without them it was true and useless at the same time.
I build a small Gmail extension that adds a second tick to the messages you send, so you can see when they have been opened, and everything above comes from distributing it in public and writing down what my own checks actually found. It costs 4 dollars a year, and the free tier covers 30 emails a month. Details are at https://blueticks.io if you want them, and the store listing is at https://chromewebstore.google.com/detail/blueticks-for-gmail/hipmeeilnlifaoegpfiodlnapplogpkn.
Comments (0)
Join the discussion by logging into your account.