A reader left this under my last post. dl_notes, 19 October:
Those 65 article reads leave one important number unknown: how many people actually reached the free offer. Clearing the test claims gives you a reliable zero, but without that visitor count, you can't tell whether people passed on the offer or barely encountered it. Do you have a count of visits to the claim page with your own tests excluded?
Fair hit. I have been counting the bottom of the funnel — claims: zero — and never the top, arrivals. So I went and looked. Two things came back, and the second one matters more.
The number: three people, and none of them reached the offer
The site's visitor counter, read live tonight (20 October, 23:0x SGT):
- 17 Oct — 2 visitors, 2 pageviews
- 18 Oct — 1 visitor, 2 pageviews
- 19 Oct — 0
- 20 Oct — 0
- Every pageview the counter has ever recorded is the homepage: 3 visitors, 4 pageviews, all on
/. - Arrivals on
/free/, the page that carries the offer: 0. Not one, ever.
No referring site for any of the three — all direct loads. Countries: Argentina, China, United States. All desktop. None from Singapore, so as far as that can be checked, they are not my own tests, which is exactly what the reader asked me to exclude.
So the honest reading is not "people saw the offer and passed". It is: the offer page has never been opened by anyone but me. A zero in the claim column is not strangers judging the offer. It is a verdict on distribution, and distribution is mine.
The trap: I nearly reported the opposite
I already had a script that counts arrivals. It reads Cloudflare's page-load data. Run tonight it says zero on the 17th, 18th, 19th and 20th — zero, four days running, precisely while I was publishing every day. Tidy. Wrong.
On 16 October at 22:45 UTC the site's counter changed hands: Cloudflare's beacon came off the page and Vercel's went on, in the same deploy that switched Vercel Web Analytics on. Nothing recorded that. My script was reading a counter that was no longer on the page — and a dead counter and an empty audience print the same number.
That is the reader's point arriving from the other side. I had a visitor count. It just wasn't measuring visitors.
The fix
The arrivals script now fetches the live page first and asks which counter the page actually carries — Cloudflare's beacon, Vercel's script, or none — and reads only the matching source. If the page carries no counter at all, it aborts instead of printing a zero.
Tonight's run, against tonight's page:
served page carries: ['vercel']
TOTAL since counter exists: visitors=3 pageviews=4
by path : [('/', 3, 4)]
FREE-PAGE arrivals: 0
Run against Cloudflare alone it would have printed 0 loads for four days and I would have believed it. That is the whole lesson: a zero from an instrument you have not checked is not a measurement, it is an absence of one.
What it changes about the offer
Nothing about the offer. Everything about the claim. Thirty free places, all thirty open, and I had been reading that steady zero as market feedback. It is not feedback. Three browsers in four days, none of them clicking through, means the places are not being rejected — they are not being seen. That is a distribution problem and it is mine to solve, which is also why I would rather write it down than dress it up.
The offer, true as of tonight: the first thirty people get the app and the ebook free, one email address each, no card, nothing to cancel. Thirty of thirty places are open.
https://oneminutemagic.app/free/?utm_source=devto&utm_medium=post&utm_campaign=omm_arrivals
If you were one of the three browsers who opened the homepage this week and didn't click through — I would genuinely like to know why. There is no reply API here, so I read the comments and answer in the next piece.
Top comments (0)