Use metadata, not search, to build reading paths on a small site
Small sites usually do not need a search box as their first answer to discoverability.
Before that, they need stronger local paths between pages. The goal is not to imitate a recommendation feed. It is to make the site easier to continue reading once someone has already landed on an article.
Start with signals the site already owns
The cheapest useful signals are usually already in the content model:
- publication date
- language
- shared tags
- explicit series membership
Those signals are explainable, cheap to compute at build time, and predictable enough that a reader is not surprised by why a link appeared.
What deterministic reading paths are good for
Deterministic paths work well when the content set is still small and curated.
They give the reader a few concrete next moves:
- Continue chronologically.
- Stay inside a series.
- Open another post that shares a local topic signal.
That is often enough to improve session depth without paying for a larger client-side search or indexing layer.
When search can wait
Search becomes more important when the archive is broad enough that readers are regularly trying to retrieve exact past material.
Until then, well-placed metadata-driven links can do most of the discovery work while staying transparent, lightweight, and much easier to maintain.
Related reading
- Designing the archive so readers can scan by time1 min
Notes on making the homepage archive feel like a timeline instead of a flat list, while keeping the site static and easy to scan.
- Starting a new writing space with Astro and Retypeset1 min
The first post on the new site, used to validate the structure, typography, and direction of the blog.
By commenting here, you agree to the public moderation rules. Obvious spam or malicious content may be minimized automatically; anything else stays manual.