Chapter 31: First-Hour Numbers
By eight o'clock, the internet had crowned a winner from the only number that could not decide one.
At 08:05 on Tuesday, 7 October, the PageTurn retention report opened inside my clean-room account. It arrived exactly as last night's status line had promised.
Before I touched it, three social screenshots covered my second monitor. Each had cropped away PageTurn's warning that clicks were provisional.
OLIVER AND PIPPA LEAD BY 18%.
One post advised me to change the Saltmere title before breakfast.
Another suggested buying enough traffic to “correct the narrative”.
A third wanted the marriage story released in exchange for sympathy clicks. The post offered a suggested headline using Pippa's name.
None showed a completion field.
They enlarged yesterday's click gap until it looked like a verdict. Repetition gave the screenshot a confidence the underlying field did not possess.
My personal banking tab was still open from the clean-room migration invoice.
There was enough savings to buy a visible bump. It was also the money from which my next rent deposit would come.
PageTurn's trial protocol offered a different button.
FREEZE SUPPLEMENTARY SPEND — PRESERVE BASELINE.
Clicking it would block my team from adding paid traffic before the first retention window closed. The restriction covered my personal card as well as the pilot budget.
It would also stop me changing the title, thumbnail or first episode. I could not quietly improve the page after seeing who left.
I read the eighteen per cent again.
Then I pressed FREEZE. A confirmation window made me type my own name rather than dismissing the choice as platform policy.
The platform asked for a reason.
I entered: FIRST-HOUR CLICKS DO NOT MEASURE COMPLETION OR RETURN.
At 08:09, PageTurn sealed the baseline and the zero-additional-spend declaration.
My launch page remained exactly as it had opened at seven. Its weaknesses would stay inside the same sample as its strengths.
The social posts continued without my assistance.
“You are letting them win the morning,” the marketing coordinator said on our review call. Her dashboard showed six unused paid placements.
“The morning is not a metric in the agreement.”
She pointed to the public share count.
I pointed to the still-loading cohort table. Half its cells remained grey because eligible reading time had not elapsed.
PageTurn retained the raw panel export; I saved the public screenshots only as examples of reaction.
They could prove what people were saying, not what readers had done. PageTurn's export carried timestamps, source definitions and the complete metric labels.
At 08:14, the first complete paths populated.
Some Saltmere visitors left before Carys opened the correction form.
Others reached the audit alert and continued into the second episode.
The sample was smaller than Larkspur's click pool. PageTurn marked it EARLY COHORT and forbade extrapolating a final rate.
It was also behaving differently.
I kept my hands away from the edit controls.
Every refresh made intervention feel both urgent and dishonest. I turned off automatic refresh and waited for the scheduled calculation.
The coordinator asked whether we could at least post my verified purchases.
“Not while the cohort is forming.”
I closed the banking tab.
That small movement cost more pride than money. I had wanted one purchase that could make the screenshots look foolish.
For twelve years, a weak opening could be softened by an audience already waiting for Marian Crowe.
Grace Ellison had no such cushion.
At 08:22, PageTurn unlocked episode two for the eligible launch cohort. The page itself did not announce a race.
The completion row recalculated without exposing reader identities.
Larkspur still held the larger opening-click column.
Then Saltmere's second-episode completion bar moved above it.
I checked the legend twice.
The readers who had found me in smaller numbers were suddenly finishing at the higher rate.
Clicks had declared me beaten before the story had been read.
Now the first retention path asked a harder question: who had stayed for Carys?