Every figure in this report comes from one data object at the top of the file. Change a number there and every chart redraws. All of it is sample data for this demo deck.
Prepared by the analytics team · Period 1 July to 30 September 2026 · Extracted 12 October 2026
Sessions grew 18% while time per session fell 9%: more people, reading less each.
0%
Sessions, quarter on quarter
Source: web analytics, 12 months to Sep 2026
Weekly sessions
Thousands · zero-based axis
Source: web analytics · band is ±1 standard error of the weekly mean
Sessions by surface and device
Thousands
Source: web analytics, Q3 2026
Share of reading time by content type
Percent of total
Source: content pipeline, six quarters to Sep 2026
Article length against completion rate
24 articles · r = 0.41
Source: content pipeline · r = 0.41, which explains about a sixth of the variation
Total reading time by section
Thousand hours
Source: web analytics, Q3 2026
Reading time by day and time of day
Index, 100 = weekly mean
Source: web analytics · local time of the reader
| Row | Expected | Observed | Why |
|---|---|---|---|
| Sat 18–22, week 6 | 124 | 211 | Live event coverage |
| Weather section, Sep | 11.9 | 26.4 | Two storms |
| Tablet sessions, week 9 | 44 | 12 | Tagging outage, 31 hours |
| Long reads share, Q3 | 31% | 38% | Three series published together |
| Completion, 4,000 words | 52% | 71% | One article drove it; not a pattern |
The tagging outage is a data fault, not a reading fault. It is excluded from the trend chart and noted here instead of being quietly smoothed.
Growth is real and steady, but it is phone growth. Anything that only works on desktop is reaching a shrinking share.
Long reads are taking share from news, not from video. The trade-off being managed is editorial, not format.
Weekday evening is the only block above index 130 every day. Publishing before 17:00 local matters more than publishing often.
| Metric | Numerator | Denominator | Excludes |
|---|---|---|---|
| Session | Visits | — | Bots, 30 min idle split |
| Reading time | Foreground seconds | — | Background tabs |
| Completion rate | Reached 90% scroll | Article opens | Sessions under 5 s |
| Share of time | Type reading time | All reading time | Untyped pages |
| Index | Cell reading time | Weekly mean | Outage window |
| Dataset | Owner | Period | Pulled |
|---|---|---|---|
| Web analytics events | Data platform | Jul–Sep 2026 | 12 Oct 2026 |
| Content pipeline | Editorial systems | Jul–Sep 2026 | 12 Oct 2026 |
| Device registry | Data platform | Snapshot | 11 Oct 2026 |
| Outage log | Reliability | Jul–Sep 2026 | 10 Oct 2026 |
Everything in this sample report is invented. Replace the DATA object at the top of the file with your own and every chart on these twelve slides redraws.