Blog Refresh: What I Changed and Why It Matters
Every few years, a blog needs attention. Not a redesign — those are expensive in ways that have nothing to do with money. But a refresh: the things readers don’t see that determine whether they find you at all.
I’ve just finished one. Here’s what changed and why.
The Content
I archived all 153 posts from the old Jekyll site and migrated them into this Astro build. Dates intact. URLs preserved where possible. Most held up. A few didn’t — outdated technical advice, dead links, topics I no longer work in. Those earned a quiet retirement.
The bigger addition: 24 new posts tracing the 13-year story of WriteDaily — a side project that’s been running since 2013, processing over 350,000 writing sessions — and introducing WordFloor, the product that grew out of everything WriteDaily taught me.
These aren’t filler. Each post has a reason to exist. The arc runs from launch through iteration, migration, security, AI exploration, user research, and a waitlist reveal. They link to each other where it makes sense, which search engines notice and readers benefit from.
The Publishing Machinery
The visible content is only half the job. The other half is whether the right systems can find, understand, and correctly represent it.
Search
Every post now has a canonical URL, an accurate meta description drawn from its home preview, and structured data (Article + BreadcrumbList) that tells search engines what kind of page they’re looking at. The sitemap language tag is set correctly — a small thing that matters when you’re writing a British English blog and Google is deciding which audience to show it to.
Pagination links — previous and next — are now proper <link rel="prev"> and <link rel="next"> tags in the page head, not just navigation links in the body. The blog listing pages have unique titles and their own canonicals.
Social
Every post now carries Open Graph and Twitter Card metadata, including a default image, a proper canonical URL, and the correct page type — article for posts, website for everything else. Without these, sharing a post on LinkedIn or Twitter renders as a bare link. With them, the platform pulls the title, description, and image. Small machinery, large effect.
Structured Data
I fixed a bug where the Article schema was pulling from the wrong field — posts without subtitles were shipping structured data with an empty description. It now falls back through home_preview to a sensible default. Search engines that render rich results will see a real description, not a blank.
The author schema (Person) correctly links to this domain and to LinkedIn via sameAs. It’s subtle, but entity clarity is one of the few SEO signals where doing the work actually matters.
Why This Matters
Most blog infrastructure advice falls into two categories: things that don’t matter (keyword density, exact-match domains, secret ranking tricks) and things that do (crawlability, canonical URLs, accurate structured data, clear entity information, content that deserves to exist).
I focused on the second category. None of it is clever. All of it is deliberate.
A blog in 2026 sits at the intersection of three discovery surfaces: classic search, social platforms, and AI answer systems. Each has different expectations, but they all reward the same underlying behaviours: crawlable pages, specific claims, visible evidence, accurate metadata, and content that couldn’t belong to anyone else.
What’s Next
The WriteDaily narrative arc is complete — 13 years of product history, told in posts that link together. WordFloor’s waitlist is open at wordfloor.com. The rest of the blog continues to accumulate whatever I’m working on, thinking about, or building.
If you’re doing a similar refresh, the checklist is short: make your pages crawlable, make your metadata accurate, make your structured data honest, and make your content worth finding. Everything else is theatre.
If you’re curious about the WriteDaily posts, the research findings from 13 years of data are a good place to start. The WordFloor introduction explains where that research led.
