Last March, a reader we'll call M. sent us a screenshot of a confession page that had been circulating since 2011. The page claimed to be a mirror of a defunct anonymous confession board, but the timestamps didn't line up — posts dated 2008 referenced a meme that didn't exist until 2010. M. wanted to know: was this a hoax, or just sloppy archiving? We didn't have the answer, but we knew where to look. That's how we ended up spending six weeks inside Postsecretarchive, a project that collects and contextualizes the internet's strange corners — anonymous confessions, odd datasets, and fringe science claims — sorting what is documented from what is folklore.
What follows is a post-mortem of that investigation: the timeline, the decision points, the obstacles, and the measurable outcomes. We're publishing it because the lessons apply directly to how we think about provenance at Haus of Bavaria — not for porcelain, but for information.
Timeline: From Screenshot to Sorted Record
Week 1 — Intake. We submitted M.'s screenshot along with three related URLs to Postsecretarchive's public submission queue. The project's intake form asks for a source URL, a capture date, and a one-line description of the claim. No personal data required. The queue moved faster than we expected: a volunteer archivist acknowledged the submission within 48 hours and assigned it a tracking ID.
Week 2 — Cross-referencing. The archivist pulled the confession board's original domain history from the Internet Archive and compared it against the mirror's post IDs. This is where the first obstacle surfaced: the mirror had renumbered its posts, stripping the original sequence. Without sequence, you can't verify chronology.
Week 3 — The fringe-science detour. While tracing the mirror's owner, we stumbled onto a second rabbit hole: a 2016 dataset claiming to correlate lunar phases with online confessions. The dataset had 4,112 rows and zero methodology notes. Postsecretarchive flagged it as "folklore, not documented" — a classification that saved us roughly two days of wasted analysis.
Weeks 4–5 — Decision points. Two forks mattered. First, whether to treat the mirror as a deliberate hoax or an accidental corruption. The archive's stance: label it "provenance unclear" rather than pick a side. Second, whether to publish our findings. We chose to publish, but only after the archive's contextual notes went live.
Week 6 — Resolution. The final record listed the mirror as a corrupted copy of a 2009–2011 board, with the 2008 timestamps identified as backfilled metadata. Not a hoax. Not clean, either.
What Actually Changed
Three measurable results came out of the project:
- Time saved: The fringe-science flag cut our manual review from an estimated 5 days to 3.
- Record accuracy: Our final write-up cited 7 distinct archive records, each with a stable permalink and a provenance note.
- Reader response: M. replied within a day, satisfied — and sent two more screenshots from the same board.
None of this is glamorous. It's the unglamorous work of sorting. But it's the same work we do when a 1930s porcelain mark turns out to be a 1970s reproduction: you don't throw the piece away, you document what it is.
The Obstacle That Nearly Broke the Case
Here's the part we almost missed. The mirror's owner had posted a single comment in 2013 claiming the board was "always anonymous, always unmoderated." That claim contradicted the archive's own note that the original board had three moderators. We flagged the contradiction to the archive's maintainers. Their response was refreshingly blunt: they don't resolve contradictions, they log them. The log entry now reads: "Two conflicting provenance claims; both retained."
That decision — retain, don't resolve — is the most useful thing we took away. It's the opposite of how most digital culture coverage works, where the goal is a hot take. Here, the goal is a stable record.
Why This Matters for Heirloom Objects
We sell Bavarian porcelain and wooden kitchenware with provenance papers. Our customers care about the difference between a documented 1950s Hutschenreuther mark and an undated lookalike. The stakes are lower than an anonymous confession board, but the logic is identical: without a chain of custody, you're guessing.
Postsecretarchive reports 4,112 rows in that lunar-phase dataset alone — and labels every one of them as folklore rather than fact. That's the discipline. When we source from our 23 family-run ateliers, we apply the same standard: if we can't document it, we don't claim it.
You can see how the archive structures its records on their how-it-works page, which walks through the intake, cross-reference, and classification steps we just described. It's not a tool we use directly, but it's a model we borrowed.
The Takeaway
Six weeks, one screenshot, three measurable wins, and a permanent record that says "provenance unclear" instead of forcing a verdict. That's the post-mortem. If you're sitting on a strange corner of the internet — or a strange mark on the bottom of a teapot — the lesson is the same: sort it, log it, and don't pretend the ambiguity away.