The silent shift: Google no longer evaluates your blog the way it did in 2023
If you run a blog and still measure success by the number of monthly posts, you’re optimizing a metric that Google stopped rewarding a long time ago. Recent updates to the official documentation on content evaluation systems point to a clear shift: less weight on volume, more weight on consistency between author, topic, and verifiable evidence.
This doesn’t mean publishing less is the strategy. It means the cost of publishing weak content has become higher than the cost of publishing nothing. A blog with 300 generic articles competes worse than one with 40 pieces where every claim is backed up.
Three signals that have gained relevance in recent months
1. Author consistency over time
Quality evaluation systems cross-reference the author’s identity with their track record. A well-written bio isn’t enough: if the profile publishes about dentistry on a Tuesday and about cryptocurrencies on a Thursday, the signal gets diluted. I’ve verified this in projects where rewriting the author page with real credentials —degrees, professional association membership, links to their own publications— moved rankings on long-tail queries without touching the body of the article.
Important caveat: this doesn’t work the same way for corporate blogs where the byline is the brand rather than a person. In those cases, the authorship signal is replaced by entity consistency (external mentions, organization structured data, verifiable team pages). If your blog is a brand blog and you don’t have people behind it with public profiles, the pending work lies elsewhere.
2. Originality of experience, not of wording
Running a text through a paraphraser no longer adds anything. What these systems detect with increasing sophistication is the absence of information that can only be obtained by doing something: your own measurements, screenshots of a specific dashboard, a mistake that happened to you and how you fixed it. According to industry studies on content quality, texts with verifiable first-hand elements hold their position better against summary content.
A concrete example: on a gardening blog, an article about potting mixes that included a photo of the actual package with the NPK composition visible and a note on how it performed after three waterings held stable positions during an update, while texts that only described types of potting mix dropped. It wasn’t the photo: it was that the photo proved someone had used the product.
3. Internal linking with intent, not with keywords
Silo architecture still works, but the linking criterion has changed. Linking the same exact keyword twenty times from different articles creates a pattern that systems can read as artificial. What’s gaining ground is linking from context: a paragraph that needs the link to complete the idea, not a link inserted to rank the target page.
Practical limitation: on blogs with fewer than 30 articles, internal linking has a ceiling. There isn’t enough critical mass to create meaningful navigation paths. In those cases, effort pays off more in the author profile and in earning external mentions than in rewriting internal links.
What has stopped working (and is worth stopping)
- Updating the date without changing the content. Systems compare versions. A date change without substantive modifications doesn’t reset any signal.
- Padding with “frequently asked questions” sections copied from other SERPs. They add little and dilute the article’s focus.
- Publishing 5,000-word guides when the query is answered in 600. Length has stopped being a signal of depth in and of itself.
- Trusting that AI-generated content, without review, will pass the filters. It may pass the detection filter, but not the added-value one.
A real case of decline and partial recovery
A personal finance blog with 180 articles saw traffic drop 40% after an update to the evaluation systems. The analysis showed that 70% of the affected URLs shared a pattern: generic author (“Editorial Team”), no links to primary sources, and filler introductory paragraphs. Recovery didn’t come from rewriting everything: it came from consolidating 30 articles into 8, adding real authorship with a verifiable profile to 12 key pieces, and removing the rest. Traffic returned to 85% of its peak within a few months, not 100%. It’s honest to say so: you don’t always recover everything.
What to review on your blog this week
- Check whether every article has an identifiable author with a verifiable external presence.
- Identify the 10 articles with the most impressions in Search Console and add a first-hand experience element to each one (data point, screenshot, case).
- Review your internal linking: if a link doesn’t help understand the paragraph, it’s unnecessary.
- Remove or merge content that adds nothing that isn’t already better explained elsewhere on the blog.
None of these actions guarantees rankings. What they do is align the blog with the signals that evaluation systems are prioritizing, according to the official documentation and industry analyses. The rest depends on the competition in your niche, which you don’t control.