Case study: openwhenitstime.com (Sealed)
SEO Case Study: A New Domain to 574 Clicks a Month
How our own product site grew from 45 to 574 Google clicks a month in four months: the technical fixes, alternatives pages, Search Console loop and what failed.
Published
Our own product site, openwhenitstime.com, grew from 45 Google clicks in April 2026 to 574 in August, and 690 in the 28 days to September 23. What worked: a clean technical base, honest competitor alternatives pages, and a Search Console loop of retitling pages to the exact queries people type. What we got wrong: one unsourced claim about a competitor.
openwhenitstime.com is the home of Sealed, a digital time capsule we built and run: you write a letter, add photos or a voice note, and it is delivered on a date you choose. Each capsule is encrypted with its own AES-256-GCM key. We launched on a brand-new domain in March 2026, with no links and no audience.
How fast did traffic grow?
| Month (2026) | Google clicks |
|---|---|
| April | 45 |
| May | 55 |
| June | 97 |
| July | 289 |
| August | 574 |
| Aug 27 to Sep 23 (28 days) | 690 |
In that last 28 day window the site had 29,612 impressions at an average position of 7.0. Three pages did most of the work:
| Page | Clicks (28 days) | Impressions | Avg. position |
|---|---|---|---|
| FutureMe alternatives | 276 | 6,634 | 5.3 |
| Open when letters | 137 | 14,548 | 7.7 |
| Time capsule prompts | 118 | 1,956 | 4.4 |
Source: Google Search Console, pulled September 26, 2026.
What did we fix first on the new domain?
The SEO foundation shipped on March 27, before any content push: robots rules, a sitemap, per-page metadata, Open Graph images, JSON-LD (Organization, WebSite, Article, FAQPage, BreadcrumbList and others), and llms.txt files. Robots rules explicitly allow the AI crawlers.
Two weeks later we found the problem that mattered most. A theme component loaded with server rendering switched off, and it wrapped the whole app, so every page reached crawlers as empty HTML. The April 12 fix restored server rendering and made www the single canonical host everywhere. On a new domain, nothing else counts until a crawler can read the page without running JavaScript.
Do competitor alternatives pages work for a small brand?
For us they were the best pages on the site. The category leader in digital letters to the future is FutureMe. We wrote an alternatives page on March 30 and a page on whether FutureMe is free on April 9.
The searches that brought clicks were questions like "is futureme not free anymore" (19 clicks in 28 days at position 2.8), "is future me free" and "does futureme cost money now". People with a pricing worry about a competitor are ready to look at another option. In the 28 days to September 23 the alternatives page alone earned 276 clicks.
This only works if every claim about the competitor is true. We learned that the hard way (see below).
How did we use Search Console to grow?
The loop we ran, and now run on every site:
- Find what a page already gets impressions for. On April 9 the FutureMe page had 313 impressions at a 0.32% click-through rate. People were seeing it and not clicking.
- Retitle to the exact phrasing. Titles and headings were rewritten to mirror the top queries word for word.
- One page per unserved phrasing. Three new posts went out the same day for queries no page answered well.
- Make sure every page gets crawled. In July, 16 of 23 sitemap URLs had never been crawled. On July 23 we linked every guide from the homepage.
- Request indexing by hand, then wait. Judge impressions after two to three weeks and positions after four to six, not the next morning.
July is where the curve bends, from 97 to 289 clicks. The homepage linking change shipped that month, but we cannot prove it caused the jump, so we will not claim it did.
What did we get wrong?
We published a competitor claim we had not verified. An early version said FutureMe had moved to paid plans at a specific price. It was not sourced. Searching for the claim then returned our own pages as its source. On September 3 we removed it, corrected three other competitor statements, and rebuilt the affected pages with sourced facts and a "checked on" date, knowing it could cost us clicks. If your content gets cited by AI engines, your mistakes get cited too.
What would we do the same way for a client site?
- Get the basics right before content: server-rendered HTML, one canonical host, a working sitemap.
- Write for exact questions with a buyer behind them, and be scrupulously fair to competitors.
- Run the Search Console loop monthly and wait the full judgment windows.
The same approach now runs on our other products, Taperlark and Synopt, where the lesson so far is a different one. On Taperlark's new domain, a wave of articles earned almost no impressions: without authority, content alone does not move.
Frequently asked questions
How long does it take a new website to get Google traffic?
For this site, the first real month was April with 45 clicks, about a month after the SEO foundation shipped. It took until July, four months in, to pass 250 clicks a month. Pages that match an exact, underserved search moved first; broad topics took longest.
Do competitor alternatives pages work for a small brand?
They were our best page. The alternatives page for the category leader earned 276 clicks in 28 days at an average position of 5.3, and questions about whether that competitor is still free brought many of its clicks. It only works if every claim about the competitor is true and sourced.
Can you rank a new site without backlinks?
Partly. This site grew mostly on pages that answered specific questions better than the existing results. But its broad head term sits at position 7 to 8 behind established sites, which is where authority starts to matter.
What should a new website fix before writing content?
Make sure crawlers get server-rendered HTML, pick one canonical host and point every other host at it, and ship a working sitemap. On this site a theme component had switched server rendering off, so every page reached crawlers as empty HTML until the fix on April 12.
What is the Search Console loop?
Every few weeks, pull the queries a page already gets impressions for, rewrite its title and headings to the exact phrasing people use, and write one new page for each phrasing no page serves well. Then request indexing by hand and wait two to three weeks before judging.