The State of UK Blogging is an annual measurement of what actually ranks on google.co.uk for the questions British bloggers ask. The same 39 questions, the same search settings and the same definitions every September, so the numbers become a line rather than a dot. Edition one was published on 8 September 2026.
Instrument v1.0, frozen 8 September 2026. Next edition: September 2027.
Cite this series
Blogging Titan (2026). State of UK Blogging 2026. Published 8 September 2026. https://bloggingtitan.com/case-studies/state-of-uk-blogging-2026/
Editions
| Edition | Published | Corpus | Dataset |
|---|---|---|---|
| State of UK Blogging 2026 | 8 September 2026 | 174 pages, 133 sites, 39 questions | CSV, 174 rows |
| State of UK Blogging 2027 | September 2027 | Same 39 questions | Due with the edition |
The question set, fixed
Thirty-nine questions, unchanged between editions and unchanged from the United States page-one census that this series extends. No UK-flavoured variants are added, deliberately: the variable under test is the search engine, not the question.
affiliate marketing for bloggers; best ai writing tools for blogging; best blogging platforms; best wordpress plugins for blogs; blog analytics metrics to track; blog comment strategy; blog design best practices; blog images best practices; blog niche ideas; blog post ideas; blog seo tips; blogging income report; blogging mistakes to avoid; content marketing strategy; evergreen content ideas; guest posting strategy; how long should a blog post be; how much do bloggers make; how often should you post on a blog; how to add video to blog posts; how to build backlinks to your blog; how to get traffic to your blog; how to grow an email list blog; how to increase blog engagement; how to make money blogging; how to monetize a blog; how to name a blog; how to promote a blog; how to rank on google; how to start a blog; how to update old blog posts; how to use ai for blog writing; how to write a blog post; how to write blog headlines; internal linking strategy; is blogging still worth it; keyword research for bloggers; newsletter vs blog; wordpress vs wix
How the corpus is built
- Each question is run on google.co.uk with gl=gb, hl=en-GB and personalisation off, from a UK connection, on a single day.
- Organic results only. Platform threads and listings are excluded: Reddit, Quora, YouTube, Facebook, LinkedIn, X, Pinterest, TikTok, Instagram, Medium, Substack, Hacker News, Amazon and Google’s own properties. These are threads, not articles.
- Results are deduplicated on domain and path, ignoring query strings and fragments. The best rank is kept, along with every question the URL appeared under.
- Each page is read from its most recent Internet Archive capture, so every measurement points at a permanent, checkable URL.
- Pages returning under 200 words of recoverable body copy are dropped as extraction failures, and the count is reported.
What is measured
Source density. The body is split into its smallest text blocks: paragraphs, list items, table cells, headings, blockquotes and captions. In each block, percentages and multiplier claims are counted. A claim is sourced if the same block, or its list or table parent, holds a link to another domain. The test is deliberately generous, so the sourced share is a ceiling and the unsourced share is a floor.
Body copy. Words, images and subheadings are counted inside the article container, with navigation, headers, footers, sidebars, scripts and forms removed. Subheadings are h2, h3 and h4.
Dates. Structured data first, then article meta tags, then the first time element carrying a datetime.
Money. Affiliate links by network and tracking-parameter patterns. Display advertising by ad-network fingerprints in the source. Email capture by an email input or a named platform embed. Own product by a same-domain link to a course, shop, pricing, membership, service or checkout path.
UK-specific measures. Whether the host is a .uk domain, whether the page names a price in pounds, whether it names a price in dollars, and whether a page carrying affiliate links also carries disclosure wording.
Calibration, and why it is published
An annual series is only worth reading if this year’s instrument measures what last year’s measured. So each wave re-reads the previous corpus with the current instrument, at the same archive captures, and publishes both readings side by side.
Instrument v1.0 against the September 2026 United States corpus, 160 pages, same captures:
| Measure | Published September 2026 | Re-read with v1.0 |
|---|---|---|
| Claims with no source | 71.2 percent | 70.5 percent |
| Pages sourcing none of their statistics | 54.1 percent | 52.3 percent |
| Median body words | 2,395 | 2,407 |
| Third quartile words | 3,619 | 3,618 |
| In the 1,500 to 2,500 band | 27.3 percent | 28.1 percent |
| WordPress share | 51.6 percent | 51.3 percent |
| Display advertising | 16.1 percent | 16.3 percent |
| Email capture | 55.9 percent | 54.4 percent |
| Own product | 5.6 percent | 18.8 percent |
Every measure reproduces within about two points except own product, where v1.0 counts any same-domain commercial path and the September pass counted a narrower set. So own-product figures and the any-revenue roll-up are not comparable across those two readings, and every other figure is. Comparisons between waves always use one instrument reading both corpora, never a new reading against an old published number.
Changelog
- v1.0, 8 September 2026. First frozen version. Question set, exclusion list, capture rule, 200-word floor and all measurement definitions set as above. Calibrated against the September 2026 United States corpus.
Using the data
The dataset for each edition is published as CSV with one row per page, carrying the live URL, the archive capture link and every measurement. Quote any figure you like with a link back to the edition it came from. If you find an error in a row, the capture link is there so you can show us.