Game Science might — or might not — release new material during two August windows. We built a four-URL search structure before either date arrived. One event page reached an average position of 5.3; one established page became the site's traffic workhorse; and two hubs exposed problems that the aggregate numbers had hidden.
The Problem: SEO for Something That Hasn't Happened
Black Myth: Zhong Kui is Game Science's second title in the Black Myth series, announced at Gamescom 2025. There is no release date, no platform SKU list, and no confirmed announcement schedule. But the studio has an August publishing pattern spanning 2020–2025: mid-to-late August has repeatedly brought gameplay demos, engine upgrades, CG teasers, and, in 2024, the full launch of Black Myth: Wukong.
In 2026, there are two dates to watch:
- August 20 — the anniversary of Wukong's first gameplay reveal (2020) and its global launch (2024). Game Science has not confirmed anything for this date.
- August 25 — Gamescom Opening Night Live, the same event where Zhong Kui was announced in 2025. Game Science's participation is unconfirmed.
Two observation windows. Zero confirmed announcements. Meanwhile, the query "black myth zhong kui release date" generated 10,368 impressions during the three-month export window.
The question: how do you build SEO pages for events that may not happen?
The Strategy: 4 URLs, 4 Roles in the Funnel
Instead of publishing new articles for each rumor cycle, we pre-embedded four existing URLs into a layered funnel — each page serving a distinct user intent:
| URL | Role in funnel | User intent matched |
|---|---|---|
| Release Date page | Bottom-of-funnel answer | "When does it come out?" |
| Aug 20 observation page | Mid-funnel event tracking | "Will we see it on [date]?" |
| Trailer hub | Top-of-funnel discovery | "What footage exists?" |
| Project hub | Top-of-funnel overview | "What is this game?" |
Each page cross-links to the other three. The Release Date page points to the next observation window; the trailer hub points to the release timeline; and the project hub provides the confirmed baseline. This creates a handoff path without forcing four intents onto one page. Whether users follow that path — and whether the pages avoid query cannibalization — still needs navigation and query-overlap data.
The Data: 85 Days of GSC, Pre-Event
Here's what Google Search Console shows for May 8 to August 7, 2026 — before either observation window has opened:
Overall growth
| Month | Daily clicks | Daily impressions | CTR | Avg position |
|---|---|---|---|---|
| May (24 days) | 4.3 | 975 | 0.44% | 15.7 |
| June (30 days) | 10.7 | 2,276 | 0.47% | 14.7 |
| July (31 days) | 24.2 | 3,786 | 0.64% | 11.3 |
| August (7 days) | 36.1 | 6,136 | 0.59% | 9.9 |
Daily clicks grew 5.6× from May to July, from 4.3 to 24.2. By the first seven days of August, the daily average had reached 36.1. Average position moved from 15.7 in May to 9.9 in early August. That August acceleration came from expanding impression coverage, not better click-through efficiency: daily impressions rose 62% from July while CTR dipped from 0.64% to 0.59%.
### Per-URL performance (85-day totals)
| URL | Clicks | Impressions | CTR | Avg pos | Verdict |
|---|---|---|---|---|---|
| Release Date (en) | 274 | 40,249 | 0.68% | 7.5 | Workhorse — 19% of all site traffic |
| Release Date (ko) | 93 | 1,986 | 4.68% | 13.0 | Korean CTR is 7× the English rate |
| Release Date (zh-cn) | 31 | 845 | 3.67% | 8.5 | High-intent, small volume |
| Aug 20 page (en) | 21 | 454 | 4.63% | 5.3 | The event hasn't happened yet and it's already position 5 |
| Project Hub (en) | 0 | 774 | 0.00% | 10.2 | Position 10, zero clicks — rewrite hypothesis triggered |
| Trailer Hub (en) | 0 | 43 | 0.00% | 7.6 | Sample too small to judge (95% CI upper bound ~7%) |
Three findings jumped out:
The event page produced a strong early signal. The Aug 20 observation page hadn't been promoted or linked from anywhere except the other three funnel pages. It was published July 21 with a status module, a six-year timeline table, and FAQ structured data. By August 7 — two weeks before the observation window — it had reached an average position of 5.3 with a 4.63% CTR. Queries including "black myth zhong kui august 20" and "gamescom zhong kui 2026" showed that the intent existed before the event. This supports the pre-embedding hypothesis; it does not yet prove the performance of the full four-page funnel.
Volume and traffic efficiency were not the same thing. The English Release Date page received 40,249 impressions at a 0.68% CTR. The Korean version received 1,986 impressions at a 4.68% CTR — seven times the English rate. In a separate country-level comparison, clicks from Japan (147 from 12,109 impressions) surpassed clicks from the United States (135 from 56,481 impressions). Page locale and searcher country are different dimensions, so we analyze them separately. The practical lesson is to find where impressions turn into qualified visits most efficiently, not simply where impression volume is largest.
Position 10 with zero clicks made value expression the cheapest next test. The Project Hub sat at an average position of 10.2 with 774 impressions and zero clicks for three months. Its title was a generic "Black Myth Zhong Kui Hub," its meta description was a checklist of column names, and its first screen used phrases like "Use this page as the main English hub for..." — editorial meta-language that answered no user question. Rank alone cannot rule out query mix, SERP features, device, country, or authority as contributing factors. But the page already had meaningful visibility, while its first screen had an obvious defect, so rewriting the value expression came before any backlink campaign.
What Broke: The "Value Expression" Rewrite
The Project Hub's 0/774 was the clearest signal in the dataset. Google's own documentation says it may generate snippets from page content rather than meta descriptions. Our page's first screen was dominated by boundary-setting language ("This page will not claim...") and SEO routing tables ("Game Science Black Myth Zhong Kui → Official FAQ + this hub → Safe answer: Use first-party status wording before any claim").
We rewrote the entire first screen on August 9:
Title: Black Myth: Zhong Kui — What's Confirmed So Far
H1: Black Myth: Zhong Kui: What's Confirmed So Far
First paragraph (before):
"Black Myth Zhong Kui, officially styled Black Myth: Zhong Kui, is the next Black Myth title centered on Zhong Kui, the Chinese demon-quelling protector. Use this page as the main English hub for Game Science status searches, confirmed facts, release-date boundaries, platforms..."
First paragraph (after):
"Black Myth: Zhong Kui is Game Science's second Black Myth title and is currently in development. No release date has been announced. This page tracks official trailers, platform wording, developer statements, and verified updates, while keeping rumors clearly separated."
We also:
- Removed an 80-line SEO query-to-route mapping table from the user-facing body
- Added
data-nosnippetto prevent the editorial boundary section from being used as snippet text - Updated the fact baseline review date from April 20 to August 9
The rewrite is too recent for GSC data, but the hypothesis is testable. After the page reaches 300 impressions from the same query cluster, we'll compare CTR within a similar country, device, and position mix. Our internal decision bands are <0.4% (rewrite again), 0.4%–1% (weak pass), and >1% (pass). These are experiment thresholds for this page, not universal CTR benchmarks.
The Trailer Hub: "Every Trailer, Explained"
The trailer hub had a different problem: not enough data. With zero clicks from 43 impressions, the rough 95% upper bound is about 7%, so the sample cannot distinguish a genuinely weak snippet from ordinary variance. We changed the title from Black Myth: Zhong Kui Trailers & Videos (a directory label) to Every Black Myth: Zhong Kui Trailer, Explained (a value promise) and added a reference table showing each trailer's date, type, canon status, and core finding on the first screen. Judgment is deferred until 200 impressions from the same intent cluster.
The Reproducible Framework
If you're doing SEO for an uncertain event — a product launch with no date, a conference appearance that's not confirmed, a feature release that might slip — here's the playbook:
1. Never build a page that can only say "we don't know."
The Release Date page doesn't just say "no date." It has a development-cycle analysis — clearly labeled as an estimate, with 2027 presented as the earliest realistic window rather than a confirmed date — plus a platform-status table, a "what a delay claim requires" explainer, and 10 FAQ entries with structured data. Confirmed facts and analytical estimates remain visibly separate. The absence of a date is context, not the page's only value.
2. Pre-embed event pages before the event exists.
The Aug 20 page went live July 21 with a six-year timeline, a status module, FAQ schema, and event-day replacement copy prepared in HTML comments. The comments were an operational shortcut for a fast update, not ranking content. By August 7, the page had reached an average position of 5.3 because the search intent ("will zhong kui be at gamescom 2026") already existed even though the announcement did not. Google needs time to crawl, index, and rank; publishing only when the event begins gives away that lead time.
3. Layer the funnel, don't flatten it.
Four pages serve four intents. A single "everything page" would have to answer release-date, trailer, event, and project-overview queries at once. Separating those jobs gives each page a clearer first-screen answer, while cross-links preserve discovery. We still monitor query overlap rather than assuming the separation automatically prevents cannibalization.
4. Track per-URL, per-query-cluster, per-country — never site averages.
"Average position 9.9" is useless for debugging. The Project Hub at position 10 with zero clicks and the Release Date page at position 7.5 with 274 clicks average to "okay," which masks a page that needs a rewrite. Break everything down by URL, query cluster, country, and device.
5. When a visible page earns no clicks, test value expression before chasing more backlinks.
The Project Hub was already appearing around position 10. Its first screen answered "how to use this page" instead of "what is this game and what do we know?" That made title, snippet inputs, and above-the-fold copy the fastest and cheapest variables to test. Ranking does not prove that authority is sufficient forever; it tells us that acquiring more links should not be the automatic first response to a click problem.
What We're Not Doing
During the two-week window before August 20, we've frozen all non-event content work: no new mythology deep-dives, no bestiary infrastructure, no entity-card expansion. The reasoning is simple — unbuilt pages are "not yet tested," not "underperforming," and event traffic is the highest-ROI use of limited pre-window time. Post-August 26, we'll review each frozen item against GSC demand data.
What the Data Supports So Far
The acquisition trend is real, and the release-date page is a proven traffic asset. The event page provides promising evidence that pre-embedding can win visibility before an unconfirmed window opens. The Korean locale and traffic from Japan justify further market-specific work, but neither result should be generalized across all multilingual pages. The Project Hub rewrite is awaiting 300 comparable impressions; the Trailer Hub still lacks a usable sample; and commercial monetization remains unproven because there is no meaningful RPM history yet.
Uncertainty became a design constraint rather than a reason to publish an empty page. The useful outcome was not a blanket claim that all four URLs succeeded. It was a system that separated confirmed facts from estimates, surfaced one strong early result, and made two failures measurable enough to guide the next test.
All data from Google Search Console, May 8–August 7, 2026. Site: blackmyth.game — a non-official fan reference for the Black Myth series. Built with Astro 5, deployed on Cloudflare Pages.
Top comments (0)