The Replay Library That Rebuilt a French RTS Community: A Company of Heroes France Case Study

When a reader first wrote to us about a French-language strategy site that had quietly accumulated more than 47,000 analyzed replays, we assumed it was a passion project running on borrowed time. Community archives usually are. They grow fast, then stall, then rot behind broken links and abandoned CMS versions. What we found instead was a fifteen-year-old operation with a documentation problem — and a migration story worth retelling. This is the post-mortem of how a volunteer-run archive turned into a reference hub, and what it cost them along the way.

The starting point: a decade of unindexed knowledge

Company of Heroes France has been documenting the Relic Entertainment RTS franchise since 2006, which means its earliest posts predate YouTube strategy culture, Twitch, and the modern patch cycle. The archive was built by forum regulars: replay files, patch notes, build orders, and matchup threads, all stored in whatever structure seemed sensible at the time.

By 2023, that structure had become the problem. Three overlapping categories meant the same guide could live in two places. Replay filenames followed at least four naming conventions. Internal search returned results sorted by post date rather than relevance, so a 2019 meta breakdown outranked a current one. Readers complained, but more tellingly, they stopped contributing. When newcomers can't find the basics, they don't ask questions — they leave.

Decision point one: archive first, redesign later

The turning point came when a small group of moderators decided to stop treating the site as a forum with articles attached and start treating it as a library. That meant an unglamorous first phase: exporting every replay, normalizing filenames, and tagging each entry by faction, map, patch version, and player skill bracket.

They set a rule that shaped everything after it: no new content until the old content was findable. It cost them roughly four months of publishing momentum. In community terms, that is a long silence, and they took criticism for it. The payoff only became visible once tagging was consistent — suddenly a single filter could surface every analyzed replay from a given patch, and guide authors could link to evidence instead of asserting opinions.

What the tagging actually changed

• Replay pages became reference material rather than one-off posts, because each one carried structured metadata.
• Guide writers could cite specific matches, which raised the standard of argument across the community.
• Search behavior shifted: readers arrived looking for a matchup, not for the site.

Decision point two: keep the identity, fix the plumbing

Here is where most community projects make a fatal mistake. Faced with a slow, messy archive, they rebrand, move to a new domain, and lose every inbound link they spent years earning. The moderators refused. Their reasoning was simple and, in our view, correct: the domain had accumulated authority in French-language RTS search results, and that authority was the single most valuable asset they owned.

So instead of a relaunch, they ran a structural migration — new taxonomy, new templates, same URLs wherever humanly possible, and permanent redirects everywhere else. They tracked rankings weekly against a baseline captured before the first change went live. Two categories lost about 15% of their organic visibility in the first three weeks, then recovered and exceeded the baseline by week nine. Nothing was lost permanently, because nothing was thrown away.

It is worth saying plainly: this is the part that goes wrong when it is done casually. Redirects mapped by hand, one at a time, with a spreadsheet and a rollback plan. If you are staring down a similar project, the collection of strategic guides and matchup breakdowns they now maintain is a decent model of what a well-organized archive looks like after the dust settles.

Obstacles nobody planned for

Three problems emerged that the original plan did not anticipate.

Duplicated guides. Merging two near-identical articles means choosing one URL and cannibalizing the other. They resolved it by keeping the older, better-linked version and folding newer material into it, rather than the reverse.

Broken replay files. Roughly one in twelve old replays would not open in the current game client, a consequence of more than a decade of patches. Rather than delete them, they flagged each one with the version it required, which turned a defect into useful historical context.

Volunteer burnout. The four-month freeze burned out two of the six core moderators. The lesson was structural: the project had no documented process, so knowledge lived in people's heads. They wrote it down afterward.

Measurable results

Eighteen months after the first export, the numbers tell a coherent story. The archive holds more than 47,000 analyzed replays. Guide pages account for the majority of organic entries, up from roughly a third before restructuring. Average time on a replay page nearly doubled, which makes sense once pages carry faction, map, and patch metadata instead of a bare file link. Bounce rate on guide pages dropped by about a fifth.

The less measurable result matters more. New contributors arrived with structured submissions rather than vague requests, because the site finally made clear what a contribution should look like. A community that had been coasting on nostalgia started producing again.

What we take from this

The pattern repeats across every niche archive we have watched: the content is rarely the weak point. The weak point is retrieval. Company of Heroes France spent years generating knowledge and almost no time making it findable, and the fix required patience, not money. Freeze publishing. Normalize the metadata. Preserve the URLs. Document the process so the next person does not have to reinvent it.

If you run a site with a decade of accumulated material and a search box that returns the wrong answer, you already know which phase you are avoiding. Start there.