I’m seeing a 62% completion rate on about 950-word pieces versus 60% on about 700-word ones over the last 30 days (n=38), which feels like noise. What benchmarks do you use for length vs. read-through, and where do you draw the line for a “good” delta — 1%, 3%, 5%?
From my stuff, the lever that moved read‑through more than length was adding a two‑sentence “TL;DR” under the headline; it lifted completion 3–5 pts on about 1k‑word pieces and about 2–3 on 600–800, regardless of length. I’d call your 2‑pt gap noise until you A/B that and segment by entry source, since social skews harsher on longer reads. Are these mostly email clicks or homepage traffic?
‘n=38’ feels tiny — I’d treat 2 pts as noise; add estimated read-time label.
Building on @badger55, I’d segment by source/device and normalize for read-time, then track “engaged minutes per 100 words” plus “reached final H2” — apples to slightly longer apples — and treat 3%+ as meaningful, anything smaller as drift. If you can share, are the 700 vs 950 posts using the same template and intro length, or is structure varying?