The MindShare Index ranks what America is watching, listening to, and collectively experiencing across every major platform — streaming, music, podcasts, theaters, and live national events — using a single comparable score. The goal is simple: one chart that captures the full landscape of American cultural attention.
Every entry on the chart is scored using the same formula, regardless of category. A Netflix series competes directly with a Spotify track, a podcast, a theatrical film, and a NASA moon launch. The math decides the ranking.
Each entry shows three things at a glance.
Rank sits in the left column — the large numeral, with a movement indicator (▲ / ▼ / — / NEW) directly underneath.
Category sits as a small uppercase label above the title. One of five, mutually exclusive: Live Event · Show · Movie · Music · Podcast. This tells you what kind of thing the entry is — not where you receive it.
Channels sit as colored pills below the blurb. They cover both reception (Broadcast, Cable, Theatrical, Netflix, HBO Max, Disney+, Prime, Hulu, Apple TV+, Peacock, Paramount+, Spotify) and attention (YouTube, TikTok, Instagram, Twitch, Podcast). The attention channels are intentional: TikTok appears on entries that have viral momentum on TikTok regardless of where they originate; Podcast appears on movies and shows that are driving heavy podcast discussion. Channels answer two questions at once: where can I receive this, and where is the cultural conversation happening.
The entry with the highest raw score in the last 52 weeks defines 100.0. Everything else is relative to that peak. In a big week, the #1 might hit 100. In a quiet week, it might score in the 60s. This is intentional — scores are directly comparable across weeks. A MindShare of 50 means the same thing in March as it does in August.
What changed in v5.3 (July 5, 2026): the score is now the square root of share-of-peak rather than the linear ratio. Cultural attention is heavy-tailed: the 2026 World Cup’s group-stage-finale week measured several times anything else in the Index’s history, and under a linear ratio that one entry read 100 while the entire rest of the chart compressed below 15. The concave curve preserves everything that matters — rank order is unchanged, 100 still means the biggest thing in the last 52 weeks, and scores remain comparable across weeks — while keeping mega-event weeks legible. Raw scores and the benchmark itself are unchanged. Pre-v5.3 MindShare scores are not directly comparable to v5.3+ scores. Historical charts are not rescored, per the standing rule against retroactive changes.
The base metric is estimated total hours of active American attention directed at a piece of content in the measurement week. This is calculated differently for each platform because each publishes different data.
Netflix is scored from its US Top 10 page (netflix.com/tudum/top10/united-states), which ranks the ten most-watched Films and Shows in the United States each week. We don’t get reported viewing hours from this page — only rankings — so attention-hours are derived from a rank-decay curve against a content-type baseline: 20M US attention-hours for US #1 TV Show, 15M for US #1 Film, decaying to roughly 4M at #10 Show and 3M at #10 Film. The decay curve mirrors the framework used for Disney+, HBO/Max, Prime, and other streamers where actual hours are also unavailable.
Library titles (catalog films riding an algorithmic resurgence, like a 2017 film re-entering the Top 10) receive a −30% discount off the rank-decay output, because library surges are less engaged than new-release viewing. Genuine cultural phenomena (a Stranger Things-tier premiere week) can receive a 1.5–3× editorial multiplier, documented in the entry’s blurb.
What changed from v5.1.3: Previously, Netflix scoring used the global TSV’s reported hours multiplied by a content-type US-share percentage (ranging from 5% to 45%). That multiplier was the single largest judgment call in Netflix scoring, introducing roughly ±50% error on any one title. The US Top 10 page replaces that guess with an actual US measurement. We lose reported-hours precision but gain geographic precision — the correct trade for a US-focused cultural attention chart. The global TSV is now used only as a validation check. Pre-v5.2 Netflix scores are not directly comparable.
None of these platforms publish viewing hours publicly. We estimate US weekly hours by combining FlixPatrol US rankings with Nielsen Gauge share ratios. For example, Warner Bros. Discovery streaming commands roughly 19% of Netflix’s US viewing share per Nielsen’s April 2026 Gauge (Netflix 7.8% of TV, WBD 1.5%), so a #1 HBO Max title is estimated at approximately 19% of Netflix’s equivalent baseline, with rank decay applied to lower positions. Nielsen’s calibration anchor for April: The Pitt, the month’s No. 1 streaming title, totaled 4.5 billion viewing minutes, roughly 17M US hours per week. One disclosure: as of April 2026 Nielsen reports Disney+ and Hulu as a single bundled line (5.0% of TV, including ESPN+); we split it 60/40 per the last separately published reading (January 2026) and will revisit if Nielsen resumes splitting them. Ratios refreshed each time Nielsen publishes a new Gauge report. The same rank-decay framework is now used for Netflix as of v5.2, bringing Netflix’s scoring into line with the approach used everywhere else.
Domestic box office grosses are converted to attention-hours: weekend gross divided by average ticket price ($11.50), multiplied by runtime. An $80M opening weekend for a 2.5-hour film produces roughly 18M attention-hours.
Spotify podcast charts and YouTube view counts are global, not US. We apply a US-share estimate by show type, just as we do for Netflix. Available data indicates approximately 50% of JRE’s listeners are outside the United States; US-culture-heavy shows like Call Her Daddy skew higher (~65% US). Podcast attention-hours combine Spotify plays and YouTube full-episode views, adjusted for partial listening. A 3-hour podcast episode doesn’t generate 3 hours of attention per play — completion rates are typically 50–60% for long-form content. We apply a 1.25-hour average for Spotify listens, a 30-minute average for YouTube views (where a “view” counts at 30 seconds), deduct 15% for cross-platform overlap, then apply the US-share percentage.
What changed in v5.3.3 (July 13, 2026): hour estimates for non-benchmark shows are now assigned from a fixed tier table keyed to each show’s blended standing across the Spotify and Apple/download charts that week. The tiers are hard bands: an estimate outside a show’s band requires a documented reason in the entry itself (a verifiable viral episode, pulled view counts). This exists to prevent estimate drift — the failure mode where week-over-week precedent quietly inflates above the documented method. The chart’s largest podcast is always scored on pulled per-episode actuals, never a default. Chart standing itself is read across three sources — Spotify’s podcast charts, Apple/download charts, and Edison’s quarterly US reach study — because any single chart under-indexes entire categories (Spotify under-ranks news shows; download charts over-rank radio simulcasts).
Beginning with Vol. 1 No. 10, music is scored at the artist level, not the individual track level. The cultural unit for music is the artist — just as TV is scored by show and podcasts by show. All of an artist’s charting tracks on the Spotify US Top 200 are combined into a single entry, with individual tracks listed in the entry details for transparency.
Spotify publishes exact weekly US stream counts but represents approximately 31% of US music streaming. We apply a Platform Coverage Correction once per artist that accounts for all-platform streaming (3.2× base), radio rotation intensity, TikTok sound usage, and ambient environmental play. A major radio artist with heavy TikTok presence might receive a 7–10× correction. A streaming-only artist with no radio gets the base 3.2×. Radio intensity and ambient levels are editorial judgments, assigned consistently and documented for each entry.
The candidate pool is rule-based, not curated: every primary artist with three or more charting tracks in the Spotify US Top 200, or any single track at 4M+ weekly streams, enters the scoring pool automatically. Nobody gets skipped for editorial reasons — if an artist at the floor doesn’t chart, it’s because the math said so.
Beginning with Vol. 1 No. 9, broadcast television series are scored using Nielsen overnight (Live+Same Day) ratings combined with next-day streaming estimates from Hulu and Peacock. A show drawing 5 million L+SD viewers for a 1-hour episode generates approximately 5M live attention-hours, plus an estimated 1–3M additional hours from next-day streaming depending on its Hulu/Peacock trending position. Notable broadcast entries include SNL (which also generates significant YouTube clip attention), prime-time dramas (Grey’s Anatomy, 9-1-1, Matlock), and live entertainment events.
National television news entered the Index with Vol. 1 No. 26 (July 27 – August 2, 2026). Until then it was excluded by omission rather than by rule: there was simply no collection step for it, while the written exclusions covered only local news, on data-reliability grounds. That was not a defensible position, and the gap is now closed in the direction the arithmetic points.
Three tiers are in scope. Network morning shows (Today, Good Morning America, CBS Mornings) on the 7–9am block average. Network evening newscasts (ABC World News Tonight, NBC Nightly News, CBS Evening News). And cable news primetime, meaning the daily anchored programmes on Fox News, MS NOW and CNN inside the 8–11pm block. The 8–11pm boundary is a coherent daypart, and it excludes The Five, Special Report and The Ingraham Angle; that is worth stating plainly rather than leaving implied.
Attention-hours are average total viewers × episodes aired × programme duration — two hours for morning shows, half an hour for evening newscasts, one hour for cable flagships. Nielsen’s average-audience figure is already duration-corrected, so multiplying it by the length of the block gives person-hours directly rather than double-counting.
Morning shows outscore evening newscasts, which is the least intuitive result the category produces and it is arithmetic rather than judgment: ABC World News Tonight draws roughly three times Today’s audience and still finishes behind it, because two hours across five mornings simply accumulates more time than a half-hour bulletin. Any definition of national news that left the morning shows out would be cherry-picking.
One structural note. Nielsen permanently excludes retitled telecasts from weekly averages and never backfills them, so a published evening-news average routinely covers fewer nights than the newscast actually aired. We take the published average as the best estimate of a typical night and multiply by the nights the newscast actually broadcast. A retitled telecast is a measurement gap, not an absent broadcast. This applies only to that artifact: a newscast genuinely pre-empted for sport did not air, and those nights are not counted.
If a tier’s numbers have not been published for the measurement week, that tier does not chart that week. It does not fall back to a monthly or quarterly average. A thirteen-week mean is a different measurement wearing a weekly number’s clothes, and a chart about attention shifts cannot mix the two. When a tier is missing it is disclosed on the chart itself, not quietly omitted.
Beginning with Vol. 1 No. 9, the chart includes Major National Events — live broadcasts, ceremonies, and historic moments that command measurable, intentional American attention at scale. If tens of millions of Americans chose to watch it, it belongs on the chart alongside streaming premieres and podcast episodes.
Events must pass a Three-Gate Eligibility Test to chart:
| Gate | Requirement |
|---|---|
| 1. Measurable | Must have verifiable live viewership data (Nielsen ratings, platform concurrent viewers, press-reported numbers) |
| 2. Sustained | Must generate multi-day attention across multiple platforms beyond the live moment |
| 3. Intentional | Audience must be actively choosing to watch, not passively exposed via feed algorithms |
This lets in: the Super Bowl, championship games, NASA missions, the Oscars/Grammys/Emmys, presidential debates, and historic trials. It keeps out: the daily news cycle, viral social media moments, political drama without a broadcast event, and ambient cultural chatter.
Event viewership data is typically stronger than streaming estimates — live broadcasts generate next-day Nielsen ratings, and platforms report concurrent viewer peaks. For multi-day events, each day is estimated separately and summed.
When estimating attention-hours from viewers × duration (box office, live events, broadcast TV), per-viewer duration is capped at 3 hours per content per broadcast window. Nobody watches every second of an 18-hour golf tournament or a 6-hour awards pre-show end to end. This cap prevents long-duration events from producing inflated scores. It does not affect platform-reported hours (Netflix, HBO, etc.), which already reflect actual viewing behavior.
Beginning with Vol. 1 No. 11, the chart includes a sidebar called The Periphery — a separate view tracking five always-on platforms: TikTok, Facebook, Instagram, ChatGPT, and Claude. The main chart measures discrete weekly content events (shows, songs, episodes, films). The Periphery measures the ambient attention infrastructure that runs underneath them — where attention lives between events.
Periphery scores are computed as aggregate US weekly attention-hours per platform, indexed against the same rolling 52-week peak as the main chart. Sharing the benchmark is the design choice: it makes the two views directly comparable, which lets the Periphery answer a question the main chart cannot — how big are these platforms relative to a single cultural event? The answer is large. On the v5.3 scale the major social platforms read in the 70s and 80s: week in, week out, their ambient attention rivals the biggest single cultural moment of the year. The 2026 World Cup group-stage week was the first time a content event out-scored every platform on the sidebar.
Refresh cadence is monthly, not weekly. Major-platform usage shifts on month-or-longer timescales; weekly refreshes would be noise. The sidebar shows a "Next refresh" label.
Deltas reflect researched platform-usage trends, not peak movement. The arrow indicators next to each platform are hand-curated month-over-month estimates from public usage data — they describe whether the platform's underlying attention is growing, flat, or declining. They are deliberately not mechanical recalculations against the previous month's score, which would swing whenever the 52-week peak moved without representing any real change in user behavior.
Why "Prototype": Per-platform attention-hour estimates currently combine public usage stats (eMarketer, Data.ai, SimilarWeb) and trade-press disclosures (TechCrunch, Backlinko). That is honest enough to publish but not to claim as a fully sourced production methodology. The Periphery will graduate from "Prototype" to "Production" once each of the five platforms has a locked, citable canonical source. Until then the sidebar carries the prototype label and a "values estimated" disclosure.
Not all hours are equal. A consumed podcast hour represents deeper engagement than a scrolled-past TikTok. The depth multiplier weights attention by focus and intentionality.
| Content Type | Multiplier |
|---|---|
| Long-form podcast (1hr+) | 3.0× |
| Theatrical film | 2.5× |
| Historic national event | 2.3× |
| Awards ceremony | 2.0× |
| Documentary | 2.0× |
| Political event (debate / State of the Union) | 1.9× |
| Scripted TV — drama, comedy, animation, any platform | 1.8× |
| Streaming film | 1.7× |
| Long-form YouTube (15+ min) | 1.7× |
| Championship sporting event | 1.6× |
| Reality TV / competition | 1.5× |
| Live sports/entertainment | 1.3× |
| Music (active album campaign) | 1.3× |
| National TV news — daily news-cycle programming | 1.25× |
| Music (playlist-driven / catalog) | 1.1× |
| Short-form YouTube (Shorts) | 0.7× |
| TikTok / Reels | 0.6× |
What changed in v5.9 (August 11, 2026): every depth pair without a written criterion was collapsed into a single row, and the prestige TV multiplier was removed entirely. The table used to carry seven high/low pairs — prestige versus general TV, two reality rows, three streaming-film rows, and so on — with nothing written down to say which side a given title belonged on. That is not a measurement, it is a slider, and two builds could defensibly disagree about the same show.
Prestige TV was the clearest case and it is now gone. The row was defined by platform — its own label named HBO and Apple TV+ — so in practice it credited a drama 22% more for the same hour watched based on which service carried it. It was also a double-count: prestige positioning shows up as discourse and awards presence, and Cultural Reach already pays for exactly those. And it was untestable, because there is no measurable definition of prestige. Scripted television is now one row at 1.8× regardless of platform, budget or critical standing.
The principle, stated so it does not drift back: depth encodes a form of consumption, not a tier of merit. The range that carries the argument is the one between forms — a podcast hour at 3.0× against a TikTok hour at 0.6×, a fivefold span standing for single-tasking audio versus thumb-scrolling, a dark room versus a screen left on in the background. Gradations within a form were never a claim about behaviour. A new row has to name a way of consuming something, not a judgment about how good it is.
What changed in v5.8 (August 11, 2026): reality and competition programming rose from 1.2–1.3× to 1.5×, and national TV news fell from 1.4× to 1.25×, crossing the two over. The old ordering discounted reality for being “social, often second-screen” viewing — but being talked about is how reality generates cultural attention, not a reason to pay it less. News moved the other way for the opposite reason: the bulk of its hours accumulate in the 7–9am block, the most background-heavy daypart on television.
Depth is not a size lever, and we tested that before changing it. Because a MindShare score is the square root of share-of-peak, taking a third off a multiplier takes roughly a tenth off a score. Every national news entry that charted at 1.4× still charted at 1.2×. Pushing a category off the chart with depth alone would require a multiplier below TikTok’s, which would be an exclusion wearing a multiplier’s clothes. Depth adjusts the weight of an hour. It is not a way to decide what belongs.
Depth follows what a title is, not the platform it arrived on (v5.8). Platform charts return titles with no notion of form, so every title is classified before it is scored: a reality or competition series takes reality depth wherever it streams, a feature film released to a streaming service is scored as a film and never as television, and a documentary takes documentary depth. This is enforced by a script rather than by memory — an entry that fails to declare its form stops the build rather than quietly defaulting.
New content receives a temporary boost to ensure fresh entries can compete with established titles. Week 1: 1.25×. Week 2: 1.12×. Weeks 3–4: 1.05×. Week 5+: no bonus. Weekly-recurring content (a podcast, a live weekly show) is always treated as Week 1, because each episode is new content.
What changed in v5.3.4 (July 13, 2026): in-season daily reality — a series airing five or more new episodes in the measurement week — is now classified as weekly-recurring content: Week 1 velocity, no recency decay, for exactly the same reason as podcasts. An audience watching six fresh episodes in a week is not paying residual attention to old content, and the old continuing-entry treatment was discounting these shows hardest at precisely the point in a season when they peak. Applied from Vol. 1 No. 23; published charts are not rescored.
Measures how many distinct platforms and contexts a piece of content appears in. A Netflix show that’s also trending on TikTok, driving Spotify streams, and generating news discourse scores higher than one that lives only on Netflix. Each secondary platform and cultural context adds 0.15 to the multiplier, capped at 2.0×.
The Cultural Velocity bar on each chart entry is a composite score (0–100) measuring how much cultural momentum a title carries this week. It combines three components:
| Component | Weight | What It Measures |
|---|---|---|
| Newness | 0–40 | How recently the title entered the chart. New entries score highest; established titles score lower. |
| Cross-Platform | 0–30 | How many distinct platforms the content appears on. More platforms = more cultural surface area. |
| Momentum | 0–30 | Whether the title is rising, falling, or holding steady. New debuts and big rank climbers score highest. |
Cultural Velocity is a display metric only — it does not affect the MindShare score calculation. It’s designed to help readers quickly spot which entries are surging this week versus which are established and holding.
Content loses relevance over time. Entries in weeks 1–2 have no decay. By weeks 9–20, the decay factor is 0.65×. Evergreen catalog content (21+ weeks) decays to 0.50×, which is intentional — these tracks persist through ambient cultural presence, but at reduced scores reflecting that they are not “of the moment.” In-season daily reality carries no decay while airing at five-plus episodes a week (v5.3.4, see Velocity above).
| Source | Data | Frequency |
|---|---|---|
| Netflix US Top 10 page | US-specific rankings (Films + Shows), primary Netflix source as of v5.2 | Tuesdays |
| Netflix Top 10 global TSV | Global hours, used for validation (pre-v5.2 was primary) | Tuesdays |
| kworb.net / Spotify Charts | US weekly stream counts, Top 200 | Fridays |
| Box Office Mojo / The Numbers | Domestic weekend grosses — two-source agreement required before any theatrical entry is scored | Mondays |
| FlixPatrol | US Top 10 rankings per platform | Daily |
| Nielsen Gauge | Platform share ratios, validation | Monthly (~2 wk delay) |
| Spotify Podcast Charts | Podcast rankings (US) | Weekly |
| Apple Podcasts / download charts | US podcast chart cross-check | Weekly |
| Edison Podcast Metrics | US podcast reach benchmark — the must-include list | Quarterly |
| YouTube Data API (v5.3.5) | Creator watchlist uploads, late-night/SNL clip counts, podcast episode views, US most-popular chart — pulled directly from the API at build time, with dated snapshots for true views-in-period | At build + daily snapshots |
| Nielsen Overnight Ratings | Live event total viewers (via Deadline, TVLine) | Next-day |
| Streams Charts | Live event concurrent viewers (YouTube, Twitch) | Next-day |
| Nielsen L+SD / Deadline / TVLine | Broadcast TV overnight ratings, total viewers | Next-day |
| Hulu / Peacock Trending | Next-day streaming rankings for broadcast shows | Daily |
Every edition of The MindShare Index is published as open data. There is a manifest of all editions at /data/index.json, a single cumulative file containing every entry from every edition at /data/all-weeks.csv, and a per-edition JSON at /data/2026-wNN.json that includes the underlying raw scores as well as the published MindShare figures.
The data is released under CC BY 4.0. You are free to use it, including commercially, with attribution to The MindShare Index (mindshareindex.com). If you are citing a specific week, name the volume and number: for example, "The MindShare Index, Vol. 1 No. 24, July 13–19, 2026."
Every field is defined in a machine-readable data dictionary at /data/schema.json: types, units, enumerated values, and a list of known issues. All files under /data/ are served with Access-Control-Allow-Origin: *, so you can fetch them straight from a browser without a proxy.
One note on parsing. The weekly_change column is a display string and its encoding varies by era, so do not parse it. Use rank, prev_rank and the status column, which carries a clean RANKED / NEW / RE / UNKNOWN enum.
Two things to know before you use it. Editions published before v5.2.1 carry a null category field, because the category taxonomy did not exist yet. And backfilled editions predating the scoring script do not carry a raw score, only the published MindShare value. Both gaps are flagged in the manifest.
The MindShare Index updates every Tuesday, aligned with Netflix’s weekly data publication. Each edition covers the prior Monday–Sunday measurement week. The chart is labeled with that week (e.g., “Week of March 23–29, 2026”).
Weekend Nielsen figures sometimes publish after the edition does. When final numbers move an entry, the current edition is trued up mid-week and the correction is noted in that edition’s intro. Prior editions are never altered — the true-up window closes when the next edition publishes.
Questions about the methodology? Reach out to Conor Sullivan at Sullivan Brothers.