A case study on our own domain, thearsenalreview.com. An internal test, not a client job. Part of the case studies.
Most restores we write about were small blogs: a few dozen posts, one theme, one era. thearsenalreview.com was a different animal. It was an Arsenal fan site where readers wrote and published their own analysis, and between October 2019 and January 2022 it gathered hundreds of pieces from close to a hundred named writers. Then the site went quiet and the domain lapsed. We registered it on 26 October 2026 and ran our WordPress restore on it the same day.
It took eleven runs over two days to get it right. This page is about why, and what came back.
Why this one was hard
The original was not a WordPress site, or a classic blog of any kind. It was a JavaScript app: the server sent a nearly empty page, and the browser drew the cards, pictures and footer after it arrived. The Wayback Machine saves what the server sends. It does not keep what the browser draws. On the archived copy even the clicks were handled by script, so there were almost no ordinary links to follow.
The Internet Archive says as much in its own help pages: simple HTML is the easiest thing to archive, and JavaScript is hard. A site built like this one is the hard case on purpose.
The first pass looked finished, and was not
The fourth run finished with most of its checks green: the kept articles sat at their old addresses with their original dates. Then a human opened the pages. Every article showed the same body text, borrowed from one archived page. Pictures were missing, the footer was gone and the links did not click.
The checks we had then looked at addresses and status codes, and all of those were fine. That is exactly the gap a human review is there to catch, and it caught it. We did not patch this one site. We changed the restore itself: it now refuses to deliver when it cannot find each page’s own content, and the final check confirms that different articles really show different text. The remaining runs went into that fix and a handful of smaller faults, and every restore since then runs on the corrected version.
What came back
We restored the site as it looked in early 2022, its last and most complete design, using the 2 February 2022 capture of the home page as the reference.
- 514 articles at their original addresses (
/article/<slug>/), with their original publication dates. - 94 original bylines, credited on the articles exactly as the writers signed them.
- 472 of the 514 articles with their original header picture.
- The About and FAQ pages, and the original sections, such as the women’s football and columnist strips on the home page.
- 1,452 permanent (301) redirects. Older address variants of kept articles point to the article itself; addresses we chose not to restore point to the home page.
The archive held 1,933 article addresses. We kept the analysis, columns, tributes and interviews, 532 in all, and left out 1,401 pieces of dated transfer and match news that nobody needs back in 2026. That was our scope choice for this domain, not a limit of the archive. Of the 532, eighteen turned out to be duplicate addresses of the same article or author profile pages, which became redirects. That leaves 514.
What did not come back
- Writer photos. Each writer’s avatar is a plain grey circle on the restored site. We did not put a guessed or borrowed face in its place.
- Moving parts. The home page sliders are static now; the first cards show, the rest stay in place without the script that scrolled them.
- 42 header pictures. Those articles show no image rather than a borrowed one.
- Two of the newest posts on the February 2022 home page, a January transfer-window piece among them, were classed as dated news and left out, so the restored home page opens on the next article.
Where the restored copy lives
The restored site is kept, unchanged, at thearsenalreview.waybackpress.com as a frozen snapshot for this case study. It is marked noindex, so search engines are asked to keep it out of their results. The writing on it belongs to its original authors; we show it here because the domain is ours and the restore is the subject of this page.
What the domain does now
A restore is a foundation, not a museum. The main domain has moved on: the same 514 article addresses, the same redirects, a new design, and an Arsenal data section on top, with results, fixtures, the Premier League table, every player and manager, season records and head-to-heads. Old links still land on the article they always pointed to. That is the whole reason to keep the addresses.
The short of it: an archive keeps what a server said, not what a browser saw. For an app-built site, getting the second back is the job, and a check that looks only at status codes will tell you it is done long before it is.