I thought MyCozyTrove would be the straightforward one. It covers fishing tackle and camping gear — things I actually use, categories I know well. I had the domain, the workflow, and the content ready to go. But while SafeHarborPrep and OpsForge Labs started showing search impressions within weeks, MyCozyTrove sat in a near-complete dead zone for about four months.
It wasn’t a content volume problem. I was pushing articles consistently. It wasn’t a quality problem. The reviews were based on gear I’d actually used in the field. The problem was a set of structural and technical issues that, stacked together, were enough to tell search engines the site wasn’t organized enough to be worth surfacing.
Here’s what actually happened.
The Gear Category Problem
The first mistake was structural, and I’ve covered the general version of it in How I Chose Topics and Categories. The MyCozyTrove version was a catch-all Gear category that had ultralight tents sitting next to saltwater lures with nothing connecting them except the fact that you could pack both. To a search engine, that’s not a category — it’s a signal that the site doesn’t have a defined area of expertise.
What I was missing wasn’t just the articles. It was the category infrastructure underneath them: the index pages, the collection definitions in the Eleventy config, the nav entries. Without those, articles aren’t indexed under a coherent topic — they’re just floating. The internal link equity has nowhere to flow because there’s no structure to flow through.
I had to stop publishing for about a week and rebuild the taxonomy from scratch. Fishing got its own category. Camping got its own. Backpacks got their own cluster. The moment those articles were indexed under proper category pages instead of a generic Gear bucket, the pattern changed. The articles were the same. The context around them finally wasn’t.
Publishing Reviews Before the Hub Existed
The second mistake was sequencing. I was eager to get product reviews live, so I published ten L4 reviews before writing any of the hub or decision pages that give those reviews topical context.
The cluster model I use across all five sites — L1 hub, L2 decision pages, L3 trigger pages, L4 reviews — only works if the structure exists before the bottom-of-funnel content goes live. A review of a specific backpack with no parent guide on how to size a pack, no comparison of frame types, no decision page explaining who needs a 50-liter versus a 70-liter — that review has no gravity. It exists in isolation and search engines treat it accordingly.
I had to go back and write the hub content after the fact, then update the reviews to link into the new structure. That retrofit cost more time than building it correctly upfront would have, and the reviews sat underperforming for the entire period between when they went live and when the cluster was complete.
The full model is in The Content Cluster Strategy That Gets AI Search to Notice You.
Affiliate Approvals Took Longer Than Expected
Traffic is one part of traction. Monetization is the other, and I underestimated how long the approval process would take for the programs MyCozyTrove needed.
Scheels through Rakuten is a strong program for outdoor and camping gear, but they do manual content reviews. When I first applied, the site was still in the fragmented Gear-category phase — no clear structure, no coherent cluster, just a pile of reviews with no context around them. The application stalled. By the time the site architecture was fixed and the application moved forward, I’d already missed weeks of having proper CTAs in live articles.
That meant a second pass through dozens of articles once the approvals came through — retrofitting the affiliate links and CTA sections that should have been there from the start. Time I could have spent building out HomesAndGardenDecor or TheCoffeeCan Camp.
The lesson is in The Gear and Outdoor Affiliate Programs That Actually Approved Me: get to minimum viable content and start the affiliate verification clock early. Don’t wait until the site is “ready.” Ready takes longer than you think, and some of these approvals take weeks regardless of when you apply.
Technical Issues That Made It Worse
Two technical problems compounded everything else.
The first was encoding corruption in the template files. Because I write and edit across different environments, Windows UTF-8 artifacts had gotten into some of my Markdown headers. The site looked fine to a human reader. To a crawler, the headers were throwing silent errors that made the site look structurally broken at the code level.
The second was a broken mobile nav. The menu toggle wasn’t responding correctly on certain Android browsers. Given that most people browsing camping and fishing gear reviews are doing it on a phone — often while they’re actually outside or standing in a shop comparing options — a non-functional mobile menu means losing nearly every mobile visitor within the first ten seconds. I have no idea how long it was broken before I caught it. That’s the part that still bothers me.
Both required a full cleanup pass of the template files and a CSS refactor before the bounce rate stabilized.
What I’d Do Differently
I wouldn’t write a single review until the category infrastructure existed. The index pages, the nav entries, the collection definitions — all of it before the first article goes live. Articles published into an incomplete scaffold don’t get the benefit of the structure once it’s eventually built. They get a second crawl cycle, which on a low-authority new site can take months.
I’d apply for affiliate programs at 10 to 15 articles, not 50. Start the verification clock early. Some programs take weeks regardless of your content quality, so there’s no reason to delay the application while you keep writing.
And I’d build the hub before the reviews. Every L4 needs an L3 to feed it and an L2 to categorize it. Publishing in reverse order meant every early review I wrote had to be retrofitted once the cluster existed. That’s a solvable problem, but it’s a waste of time you don’t have to create for yourself.
MyCozyTrove eventually found its footing. It’s getting clicks now, not yet converting at meaningful volume, which is honest and about where you’d expect a site at this stage to be. The four months of slower-than-expected progress came down to architecture, sequencing, and two technical issues that should have been caught earlier. None of it was mysterious once I found it.
For the full picture of how this fit into running five sites at once, see What I Actually Did to Build Five Affiliate Sites.