RUNTIMEWIRE IN PUBLIC
Building a newsroom in the open.
Most media companies keep their numbers behind glass. We don't. This page shows our real readership, how it's growing, what people are reading, and who's citing us — updated continuously, with traffic anomalies stripped out.
The numbers
Every metric here comes from our own first-party pageview logs — the same data our newsroom runs on. A "read" is a single article view, bucketed by Central Time days to match how we report internally.
How we're growing
We publish our daily and monthly readership trend, with milestone annotations marking launches and press moments along the way.
What drove the growth
We list the stories that pulled the most readers, on a delay, so this page can never double as a live assignment feed. Roughly 99% of our pageviews arrive as in-app navigation, which loses the original referrer, so the traffic channel is intentionally coarse.
Who's citing us
When other outlets, aggregators, and communities cite our reporting, we log the verified mentions here — no "As seen in" theatre, just the receipts.
Same-story reporting audit · Aug. 1-15, 2026
RuntimeWire compared its reporting with coverage from 11 competing technology newsrooms. Articles covering the same underlying story were evaluated as clean text without publication branding. The result: 98 RuntimeWire wins, 1 tie, 21 losses across 120 clean-text comparisons — an 81.7% win rate. VentureBeat was the strongest large-sample challenger and won its direct comparison with RuntimeWire: 17 VentureBeat wins, four RuntimeWire wins and one tie.
Audit period: Aug. 1 through Aug. 15, 2026 · Eligible population: 1,996 articles · RuntimeWire articles: 476 · Articles from 11 competing newsrooms: 1,520 · Topic clusters: 1,531 · Stories covered by multiple publications: 199 · Blind-graded pairs: 205 · Clean-text comparison set used for the headline result: 120.
Candidate matches were produced using Gemini embeddings and entity-assisted matching. GPT-5 mini verified comparison edges, and GPT-5.5 reviewed cluster splits. The headline result uses the 120-pair clean-text subset. Publication names, branding and surrounding page presentation were excluded from those comparisons.
The full blind-graded set contains 205 pairs. The 98-1-21 headline result applies to the 120-pair clean-text subset — pairs where the competitor's feed delivered complete article text rather than stubs or truncated copy.
Download audit workbook (.xlsx) (XLSX, 608 KB) · Download raw audit files (.zip) (ZIP, 1.0 MB)
How we count
We strip out windows of obvious bot or non-human traffic before computing these figures, so the trend reflects real readers. If a number here ever looks wrong, tell us: tips@runtimewire.com.
← Back to the wire