On November 1, the person on my fan site changes. Through October, people came to an independent fan reference I own with Halloween searches. From November they ask what to give. The Halloween guide carried the first peak: 5.5K of the site's 45.9K page views in the four weeks to October 27, about one in eight. (Google Analytics 4, pages, September 30 to October 27, 2026.) This post is about handing that attention to the next season's page without starting over. It's the third outcome in my system, get chosen, planned for a reader who hasn't arrived yet.
The last handover on this site has a result: when I rebuilt its seasonal guide around the reader's deadline, it paid off, which why content that ranks still fails to move people closes on. This one's results aren't in. This is the method and what I'll measure; a dated note with the real numbers follows in January.
Why keep the page that already ranks?
Because a new address starts at zero, and this one didn't. The site already had a gift guide. Search Console listed 14 different searches that showed it between July 9 and October 23, 848 showings in all, at an average position of 8.6. The two biggest were the series' name plus "gifts" (333 showings) and the series' name plus "gift ideas" (222). (Google Search Console, query and page rows for the guide, data through October 23, 2026; position weighted by showings.)
A page called something like "Christmas gifts 2026" would have competed with that guide for the same searches and thrown away the history. So the Christmas rebuild went onto the address that ranked, on October 30. I keep URL structure boring on purpose. The season goes on the page: the title, the look and the picks. The address stays the same all year.
What changes on the page when the season does?
The reader. A Halloween shopper is a fan buying for themselves. A gift shopper is often someone who has never read the books, buying for someone who has. So the rebuilt guide opens with two ways in, one for fans and one for people who have never read the books, and every pick says who it's best for, why it beats the obvious alternative, and when to skip it. That structure is the one in what a good commercial page looks like. Each pick is also marked by how far into the series it's safe to look, because a gift shouldn't spoil the story for the person opening it.
Where does the answer go on a phone?
Where phones actually get to, measured in pixels. Clarity's scroll reports give a share of page height, and page heights had changed, so the readings were converted to pixels first. On the site's home page, an estimated 42% of phone views got 1.2K pixels down and 16% got 2K. On its most-read page, an estimated 59% and 24%. Clarity reports scroll in 5% steps, so these are interpolated, not counts. (Microsoft Clarity, phone views, October 19 to 25, 2026, converted from page heights measured at 390 pixels wide.)
The guide itself had too few views that week to read on its own (299 in four weeks, all devices), so I placed it by the site's other pages. Three picks a reader shouldn't miss now sit right under the title.
How do readers get from one season's pages to the next?
Through links where they already are, with a banner as a reminder. On October 30 the guide got a card in the phone menu, a homepage tile and one line each on three busy pages, beside the Halloween links rather than in place of them.
I didn't lead with the banner because of what readers did with the last one. On the home page in the week of October 19 to 25, the Halloween banner's close button was the second most-tapped thing on phones, 11 taps. (Microsoft Clarity, home page heatmap, phones.) That's one week and one page, a lead rather than a test. The banner still switches to the gift message on November 1, on each reader's own date, with no deploy. Which kind of link sends more people is a January question.
How do you research questions a season hasn't asked yet?
Search Console only knows what already happened. None of the guide's 14 listed searches was phrased as a question, and none asked about Christmas. That's what the report shows, not why: it only lists what already happened. So the questions came from where the season shows up first.
On October 30 the guide's questions went from 7 to 16. Six of the nine new ones had no Search Console data behind them at all. They came from Google and Bing autocomplete, from the kind of prompts people type into chatbots, and from a gap no other gift guide covered, each written down with where it came from and the date it was checked. Some of the new questions ask whether something exists at all. When I couldn't find it, the answer says so and gives the date of the check, and that line carries an expiry date so it can't go stale.
One of the 14 listed searches already looked like the future: a full sentence, written the way people talk to a chatbot. It showed the guide 21 times.
The site's check for reader questions changed with it. Each question now needs either Search Console showings or a named, dated signal: an autocomplete suggestion, a chatbot prompt or a competitor gap, with where and when it was checked. "People ask this" no longer counts as evidence.
How does the season switch itself on and off?
By date, not by deploy. Each page checks the reader's own date: before November 1 it carries the Halloween look, and through January 1 the Christmas one. The banner swaps its message the same way.
Dated words get expiry dates of their own. Twelve Halloween phrases across the site fail the build from November 1, and one more raises a warning, so the first deploy after Halloween has to take them down. Every line that says "Christmas 2026" fails from January 2, and then the guide drops the season and stays a year-round gift guide on the same address. (The site's expiry list, read October 30, 2026.) The expiry habit has its own miss: a sweep on October 7 found Halloween lines that had none, and they got one. How I keep my sites accurate covers this kind of check across my sites.
What did I get wrong?
A check that couldn't see a shopping page
The question check couldn't see a shopping page. The rule that decided which pages needed reader questions counted only question-shaped searches. The guide's 848 showings were all shopping searches, so to that rule the guide had no demand at all. Its seven questions dated from October 9 and none was written for the holidays. The fix is a second rule: any shopping page with 300 or more showings must show at least three reader questions. That's a floor, and the old guide's seven questions would have cleared it. What found the gap was reading the searches by hand; the rule only makes sure the next shopping page isn't invisible to the check.
Working words on a shopping page
The rebuilt guide shipped our working language. The same day it went live, I caught four problems on the live page. Three were words meant for us, not for shoppers: a book's catalog code in a comparison table, a note saying when store listings were checked, and a label written in our own shorthand. The fourth was a contents list pinned to the side of the page on wide screens, which on real screens sat over the header. All four were fixed the same evening. The words became a rule: a shopping page's visible text can't use the vocabulary I build and check the site with. The rule was proven by planting the three lines that shipped and confirming each one fails, while a plain affiliate disclosure still passes. The pinned list got a fix and no rule. That one I'll catch by looking at the page on real screens before I call it done.
How will I know if it worked?
By comparing November and December with what I wrote down before the season started.
| What I'll compare | Before the rebuild | Source |
|---|---|---|
| Showings in the searches Search Console lists | 848, July 9 to October 23 | Search Console, query and page rows |
| Average position | 8.6 | Same rows, weighted by showings |
| Store-link click events on the guide | 11 events across 299 page views (events, not a conversion rate) | Google Analytics 4 notes, September 30 to October 27 |
| Which link sends people to the guide | Not measured yet | Clarity and Google Analytics 4, from November 1 |
In January I'll add a dated note to this post with complete months only: what moved, what didn't, and which of these decisions I'd make again.
What can't this tell you yet?
- Whether any of it worked. Every number above is from before the season. None of it is a result.
- What caused a change, when one comes. Holiday demand rises on its own. A bigger December is the season until it beats what the same page did in the weeks before, and even then growth can't isolate what the rewrite did; it's a lead.
- Every search. Google says Search Console leaves some queries out, so 848 is the listed total, and the real one may be higher.
- How the guide itself is read. The phone reach figures come from other pages on the site, in a week full of Halloween traffic.
- Whether it repeats. One season on one site is one trial. Next October is the second.
The season changes who's asking. The page that already earned the search should be the one that answers them.
Written for this site and first published October 30, 2026. Every number in it comes from my own exports and is dated where it appears.