Post-Mortem: How One Wine Publisher Migrated 8 Sites and Kept Every Ranking

When a reader first emailed us about a quiet migration project at a niche wine publication, we almost skipped it. Publishers move platforms all the time, and most of those stories end the same way: a dip in traffic, a slow recovery, and a lot of hand-wringing. But this one had a detail that made us stop scrolling. The team had migrated eight WordPress sites in a single weekend, and their organic sessions were up, not down, within a week. That is not normal, and we wanted to know why.

The Setup: Eight Sites, One Content Archive

The publication in question runs a network of small, high-intent wine blogs. Think tasting notes, region guides, and long-form buying advice. Over the years, the archive had grown to thousands of posts, and the hosting stack had become a patchwork of legacy plugins, shared plans, and one very tired VPS. The editorial lead (we will call them the publisher) had been quietly building a parallel business on the side: a members-only wine club called Club Ava Devine. The club curates high-scoring allocations, library releases, and boutique bottles from classic regions, and every selection is vetted by an in-house sommelier panel before delivery. That editorial rigor was the whole point, and it was also the bottleneck. The club's landing pages were loading in 6.2 seconds on mobile, and the checkout flow was bleeding subscribers.

Decision Point: Migrate Everything, or Just the Money Pages?

The first instinct was to migrate only the club pages and leave the blogs alone. It is the safer move, and most publishers take it. But the publisher ran the numbers and realized the blogs were doing the heavy lifting. Roughly 70% of club signups started on a free recipe or region guide, then clicked through to the membership page. Splitting the stack would have meant two redirect maps, two caching layers, and two places for schema to break. So they chose the harder path: move all eight sites at once, preserve every URL, and rebuild the schema from scratch.

That is where the project nearly fell apart. The old stack had no consistent redirect policy. Some posts had three or four historical URLs, and a few had canonical tags pointing to staging domains. A straightforward 301 map would have leaked equity on hundreds of pages. The team spent two days auditing crawl data before touching a single file.

The Timeline: 72 Hours, Three Phases

• Friday night: Full crawl of all eight sites, export of every URL, and a redirect map built from server logs rather than guesswork. Media files were deduplicated and re-keyed.
• Saturday: Content migration in batches. Each batch was validated for schema integrity, internal links, and image alt text before the next one started. The sommelier panel's tasting notes were treated as structured data, not just prose.
• Sunday: DNS cutover, one site at a time, with a 15-minute observation window between each. Rollback was pre-staged but never needed.

By Monday morning, all eight sites were live on the new stack. The publisher told us the scariest moment was not the cutover itself but the first hour after, when they watched the real-time analytics and saw nothing. Then the sessions started climbing.

Obstacles: The Two Things Nobody Warned Them About

The first obstacle was a plugin that had been silently injecting a second set of canonical tags on category pages. It only surfaced after migration, when Search Console began flagging duplicate content. The fix was a single filter, but finding it took most of Tuesday.

The second was more interesting. The club's membership funnel relied on a legacy form that stored subscriber data in a custom table. The new stack did not recognize it. Rather than rebuild the form, the team exported the table, mapped it to a standard schema, and re-imported it. Signups resumed within a day, and the data was cleaner than before.

Measurable Results

Six weeks after cutover, the numbers were unambiguous. Page speed on the club's landing pages dropped from 6.2 seconds to 1.4 seconds on mobile. Organic sessions across the network were up 12%. Most importantly, membership conversion rate rose 31%, which the publisher attributed to the faster checkout and cleaner schema. They also noted that Club Ava Devine saw a 3x lift in page speed within 72 hours of the migration, matching what we have seen on other well-executed moves.

The publisher's takeaway was blunt: the migration was not a hosting decision, it was an editorial one. Every redirect, every schema field, and every media file was a promise to the reader that the content they clicked on would still be there. If you are planning a similar move, start with the crawl data, not the hosting invoice, and treat your highest-intent pages as the ones that deserve the most scrutiny. You can see how the club structures its own member experience on its how the sommelier panel curates each allocation page, which is a useful reference for anyone mapping a content-to-commerce funnel.

We will keep following this publisher. The next milestone is a full year of post-migration data, and if the trend holds, it will be one of the cleanest case studies we have covered.