2023-H1 is excluded — it's the backlog-inflated first period, and including it would dwarf every real period on this axis.
% of that period's new titles carrying each genre. Full attribution per genre (a 3-genre title counts fully toward all 3), so bars can add up to more than 100% — same convention as the rest of the genre charts.
Same top 8 genres as the mix chart above, but as period-over-period growth in raw title count rather than share of the mix — a genre can hold steady share while still shrinking in absolute volume. No bar for a period a genre had zero titles in previously (shown as "new" in the tooltip instead of a meaningless infinite %). 2023-H1 is excluded from the period axis — it's the backlog-inflated first period, so "change since 2023-H1" isn't a real momentum signal.
Red = share of that period's production (runtime hours). Blue = share of that period's viewership (each title's launch-period views only, not lifetime — so recently released titles aren't penalized for having had less time to accumulate views). Sorted by the gap between the two: top rows over-invested, bottom rows under-invested.
Same gap as the chart above (production share minus viewership share), trended across every release period instead of frozen on one — shows whether a genre runs consistently over- or under-invested or whether it's a one-period blip. Above 0 = over-invested that period, below 0 = under-invested. Top 8 genres by total runtime produced, same set as "Runtime Produced by Genre." 2023-H1 is excluded — it's the backlog-inflated first period, and its production/viewership mix is an artifact of that, not a real investment signal.
Launch-period hours watched per runtime hour produced, by genre, across each release cohort. Titles with missing or unreadable runtime data are excluded rather than counted at zero, so they can't inflate the ratio. 2023-H1 is excluded entirely: Netflix's first engagement report only published hours viewed, runtime wasn't added until the next report. A declining line means Netflix is getting less return per hour invested in that genre over time; a rising line means the opposite.
X = average efficiency across every release cohort this genre/country appears in (same hours-watched-per-runtime-hour metric as the chart above). Y = a 0-100 consistency score built from how much that efficiency swings cohort to cohort: 100 means it delivers almost the exact same return every single time, lower means the number bounces around, some cohorts are hits and others flop. Bubble size = total runtime hours produced across qualifying cohorts. Only genres/countries with releases in at least 3 separate cohorts are shown, otherwise there isn't enough history to call something "consistent." Top-right is the safest bet: strong average return, and it shows up reliably. Bottom-right is a boom-or-bust category: high average masking wide swings. Top-left is a dependable but modest performer. Bottom-left is both weak and unpredictable.
TV only, Season 1 to Season 2 specifically. Each returning season is its own row in Netflix's report, so this matches every show's seasons to each other by name (parsing "Season N," "Part N," "Volume N," "Book N," and similar Netflix naming conventions, plus bare numbering like "Stranger Things 4"), then compares Season 2's launch-period views back to Season 1's. Views, not hours watched, since Netflix's views figure is already runtime-normalized, so a shorter or longer season doesn't distort the comparison. Transitions where Season 1 had under 1M views are excluded, Netflix rounds small titles to the nearest 100K, which turns ordinary noise into headline-looking swings for anything that small. Shows whose Season 1 first charted in 2023-H1 are also dropped, that report never published a views column. Each bar is a views-weighted retention rate across every qualifying show in that genre (total Season 2 views vs. total Season 1 views), not a simple average of percentages, so a handful of tiny shows can't outvote the genre's actual hits. Below 0% means that genre's shows lose audience by Season 2 on the whole; above 0% means they grow. Only genres with at least 3 qualifying shows are shown.
Indexes each genre's Season 1 audience at 100, then walks it forward through Season 2, Season 3, and Season 4 using the same views-weighted retention rate and 1M-view floor as the chart above, so the two charts agree with each other. The gray bar is the Season 1 starting point; each later bar shows where the index lands after that transition, so a shorter bar means a genre keeps bleeding audience further into its run. A genre only requires 3 qualifying shows at Season 1-to-2 to appear at all; if it doesn't have enough shows reaching Season 3 or Season 4 to say anything reliable, its bars just stop there rather than guessing at a stage. Action, Adventure, and Western are left off entirely, all three only clear the minimum with a 3-show Season 1-to-2 sample and never reach a qualifying Season 3, too thin to sit next to genres backed by dozens of shows and a full multi-season staircase. Sorted worst-to-best by wherever the index ends up on its last available season.
Picked algorithmically, not a fixed list: every show whose Season 1 released after 2023-H2 (so its entire run sits inside Netflix's actual reporting window, no pre-report catalog titles), ranked by Season 1 launch views, top 5 taken. Each show is indexed to its own Season 1 at 100, then walked through Season 2 and Season 3 where a third season exists. A season only counts if its release date lines up with the report period crediting it with views, the same safeguard used above, so a stale catalog number can't sneak in and masquerade as a launch. The ordinal above each bar is that season's rank among every TV title's views for that same report period, Netflix's own per-period leaderboard, the same basis as its public Top 10. None of this year's top 5 have reached a Season 3 yet, so every show here currently shows just the one Season 1 to 2 step.
Bubble size = number of titles carrying that genre. Top-left (low runtime, high views) = underserved — room to make more. Bottom-right (high runtime, low views) = heavy investment without matching payoff.