CLICK ON ANY FINDING TO EXPAND
"What to know" gets Featured 1.6× more often—but organic views trend lower (padj=0.16, not significant at α=0.05, n=17).
→ Use "What to know" when targeting Featured specifically. Don't apply it broadly.
Details ↓News brand notification CTR declined 29% over 9 months (1.77% → 1.25%, Jun 2025–Mar 2026). Attribution language (1.32×) is the only consistent headline signal still lifting CTR. Two distinct ecosystems: news brands vs. celebrity content have non-overlapping formula signals.
→ Segment notification strategy by content type. Use attribution language for news brands. Monitor the CTR decline—the channel is maturing.
Details ↓Sports underperforms across Apple News featuring and notifications (lowest CTR topic among all topics). It performs organically well on Apple News and SmartNews—but algorithms don't surface it. Nature/wildlife is the mirror: underperforms on Apple News but overperforms on SmartNews.
→ Write platform-specific sports briefs. Sports is a reader-intent topic, not a distribution topic. Don't rely on algorithmic reach for sports content.
Details ↓Number leads climbed from 0.68× (Q1 2025) to 1.06× (Q1 2026)—the only formula to cross into above-baseline territory. Question format dropped from 0.66× to 0.48×.
→ Lean into number leads; deprioritize question-format headlines. Re-check quarterly as 2026 data accumulates.
Details ↓Formula choice matters—but only within the right topic. Crime + "Here's" = 16% featuring (n=89); Business + "Here's" = 14% (n=72). Sports = 0% featuring regardless of formula—topic opts you out entirely. Weather is a special case driven by automated United Robots content, not editorial headlines.
→ For Crime and Business: use "Here's" format to maximize featuring odds. For Sports: formula doesn't matter—the topic itself limits distribution. Don't spend headline effort chasing featuring on Sports content.
Details ↓Formulas promoted for Apple News specifically hurt SmartNews. Question format: −0.08 rank below baseline (p=3.4e-6, n=918). "What to know": −0.13 below baseline (p=3.0e-6, n=213). "Here's" is the only formula directionally above baseline on BOTH platforms (0.543 SN rank, p=0.038, does not survive Bonferroni correction).
→ Use "Here's" format when syndicating to both platforms simultaneously. Use questions for Apple News only. Never write one Apple News optimized headline and route it unchanged to SmartNews. One-time setup: wire into CSA Apple News and SmartNews persona configurations.
Details ↓Finding 1 · Featured on Apple News
Among the 1175 Featured articles in our dataset, "What to know" headlines are dramatically overrepresented: 14 of 31 (45%) were Featured, versus 27.4% overall. This is a statistically robust formula signal (χ²=4.1, padj=0.07).
Question-format headlines are also Featured more often than expected (44%, 1.61× lift, padj<0.001 ***)—but they significantly underperform other Featured articles once selected. Apple's editors favor questions; the format itself doesn't follow through on views.
Quoted ledes present the inverse pattern: Featured at roughly the baseline rate (34%), but once Featured they deliver among the highest within-Featured percentiles. Questions get into the Featured tier and stall; quoted ledes get in and overperform.
Causal note: The association between "What to know" and Featured placement is observational. The causal direction is ambiguous: editors may independently choose the same stories that writers frame as "What to know," rather than the format itself driving featuring.
| Formula | n | Featured rate | Lift | padj (BH–FDR) | Within-Featured median %ile |
|---|---|---|---|---|---|
| ★ What to know | 31 | 45% | 1.65× | padj=0.07 | 80% (1.27× Featured avg) |
| ★ Question | 199 | 44% | 1.61× | padj<0.001 *** | 44% (0.70× Featured avg) |
| Here's / Here are | 28 | 39% | 1.43× | padj=0.27 | 54% (0.86× Featured avg) |
| Quoted lede | 172 | 34% | 1.25× | padj=0.07 | 86% (1.36× Featured avg) |
| Possessive named entity | 131 | 25% | 0.92× | padj=0.63 | 59% (0.94× Featured avg) |
| ↓ Number lead | 231 | 17% | 0.63× | padj=0.002 ** | 75% (1.19× Featured avg) |
Read this table as: "Featured lift" is how much more often Apple selects this formula for Featured. A high rate means Apple's algorithm rewards it—not that it organically outperforms.
Quote lede breakdown: which type gets featured?
| Quote lede type | n | Featured rate | Lift vs. baseline | p (chi²) | Within-Featured median %ile |
|---|---|---|---|---|---|
| Official/authority quote | 3 | 67% | 2.43× | p=0.38 | 62% |
| Third-party attribution | 151 | 35% | 1.28× | p<0.05 | 62% |
| Subject's own words | 18 | 22% | 0.81× | p=0.82 | 69% |
Subtypes classified by keywords after the closing quote mark. n=172 total quoted-lede articles. p-values are uncorrected for this exploratory breakdown. Interpret as directional guidance, not confirmed findings.
Question format deeper dive: which question types do Apple editors choose?
| Question type | n | Featured rate | Lift vs. baseline | p (chi², vs. other questions) | Organic median %ile |
|---|---|---|---|---|---|
| Is / Are questions | 3 | 67% | 2.43× | p=0.84 | — |
| Will / Can / Could questions | 3 | 67% | 2.43× | p=0.84 | — |
| What questions | 5 | 60% | 2.19× | p=0.79 | — |
| How questions | 4 | 50% | 1.82× | p=1.00 | — |
| Other question types | 178 | 44% | 1.62× | p=1.00 | 29% |
| Who questions | 1 | 0% | 0.00× | p=1.00 | — |
| Why questions | 5 | 0% | 0.00× | p=0.12 | 31% |
n=199 question-format Apple News articles (2025–2026). Featured lift vs. 27.4% baseline. Chi-square p-values uncorrected (exploratory breakdown within a single formula type). Organic median = percentile rank for non-featured articles of each question type vs. cohort.
All 4,283 Apple News articles (2025–2026). Chi-square test: each formula vs. all other articles combined. BH–FDR across all 7 formula tests. Causal direction of "What to know" → Featured is unconfirmed.
Finding 2 · Push Notifications
Across 1923 Apple News push notifications (Jan 2025–Feb 2026), the dataset contains two functionally different content types. Us Weekly (entertainment/celebrity) runs at 3.97% median CTR; the four news brand outlets (Miami Herald, KC Star, Charlotte Observer, Sacramento Bee) run at 1.40%. The formula signals that predict CTR are almost entirely non-overlapping between the two populations.
News brand signals (n=1124, median CTR 1.40%)
| Feature | n (present) | Median CTR (present) | Median CTR (absent) | Lift (95% CI) | Effect size r | padj (BH–FDR) |
|---|---|---|---|---|---|---|
| Attribution (says/told) | 108 | 1.79% | 1.36% | 1.32× [1.21×–1.47×] | -0.31 | padj<0.001 *** |
| Short (≤80 chars) | 820 | 1.42% | 1.33% | 1.07× [0.99×–1.19×] | -0.04 | padj=0.30 |
| Named person + possessive | 137 | 1.35% | 1.40% | 0.96× [0.85×–1.04×] | 0.08 | padj=0.22 |
| Full name present | 578 | 1.36% | 1.42% | 0.96× [0.88×–1.02×] | 0.05 | padj=0.28 |
| Question format | 96 | 1.30% | 1.41% | 0.93× [0.78×–1.02×] | 0.12 | padj=0.17 |
| Contains number | 275 | 1.31% | 1.42% | 0.92× [0.86×–1.02×] | 0.03 | padj=0.46 |
For hard news notifications: "EXCLUSIVE:" earns clicks when the story justifies it. Attribution language ("says"/"told"/"reports") signals source credibility and is associated with higher CTR. Question format consistently hurts. Notification length shows no significant effect when analyzed within this population.
Character length and CTR
Notification character length has a measurable relationship with CTR in news brand content. The data across 1124 news brand notifications shows:
| Length bucket | n (all) | n (news brands) | Median CTR (news brands) | Median chars | p (vs. other bins, news brands) |
|---|---|---|---|---|---|
| <70 chars | 269 | 113 | 1.31% | 64 | p=0.16 |
| 70–89 chars | 1318 | 946 | 1.42% | 79 | p<0.05 |
| 90–109 chars | 296 | 60 | 1.12% | 96 | p<0.05 |
| 110+ chars | 40 | 5 | 1.22% | 112 | p=0.34 |
The serial/escalating story as a content type
The top news brand notifications by CTR are dominated by a single story: Nancy Guthrie's disappearance and its connection to Savannah Guthrie. This defines a content type: the serial/escalating story with a celebrity anchor. The formula: possessive named entity + new development + escalating stakes, published in installments. The structural recipe: "[Celebrity]'s [family member/associate] [new disclosure/development]."
Us Weekly / celebrity signals (n=799, median CTR 3.97%)
| Feature | n (present) | Median CTR (present) | Median CTR (absent) | Lift (95% CI) | Effect size r | padj (BH–FDR) |
|---|---|---|---|---|---|---|
| Full name present | 638 | 4.26% | 2.89% | 1.48× [1.36×–1.56×] | -0.48 | padj<0.001 *** |
| Named person + possessive | 295 | 4.37% | 3.69% | 1.18× [1.12×–1.29×] | -0.24 | padj<0.001 *** |
| Short (≤80 chars) | 395 | 4.13% | 3.82% | 1.08× [1.00×–1.15×] | -0.08 | padj=0.08 |
| Attribution (says/told) | 41 | 4.18% | 3.95% | 1.06× [1.00×–1.15×] | -0.06 | padj=0.60 |
| 'Exclusive' tag | 71 | 3.95% | 3.98% | 0.99× [0.85×–1.13×] | 0.04 | padj=0.60 |
| Question format | 14 | 3.37% | 3.99% | 0.84× [0.55×–1.07×] | 0.33 | padj=0.06 |
| ↓ Contains number | 264 | 3.14% | 4.31% | 0.73× [0.67×–0.80×] | 0.37 | padj<0.001 *** |
For celebrity/entertainment notifications: named person + possessive ("Smith's…") is the dominant signal. Numbers in the headline hurt CTR—specific counts and statistics feel out of place in celebrity context. "EXCLUSIVE" shows no significant effect, suggesting the word has different valence in entertainment contexts. Notification length shows no significant effect within this population.
News brand CTR trend (post-activation, Jun 2025–Mar 2026)
News brand Apple News notifications went live in June 2025. Monthly CTR has declined from 1.77% (June 2025) to 1.25% (March 2026)—a 29% drop over 9 months. The pattern shows an initial sharp decline through October 2025 (1.25%), a brief partial recovery in December (1.31%) and January 2026 (1.42%), then a return to trough levels by March 2026 (1.25%). This trajectory is consistent with audience accommodation to a new push channel—initial novelty drives higher engagement, which normalizes as the channel becomes routine.
Attribution language ("says"/"told"/"reveals") remains the only consistent headline signal that still lifts CTR (1.32×, padj<0.001 ***) against this declining baseline. As the overall level drops, having the right signals becomes more important, not less—the population of readers who do click is increasingly self-selected for story-level interest rather than channel novelty.
Monthly CTR data (news brands only): Jun 2025: 1.77% → Jul: 1.70% → Aug: 1.50% → Sep: 1.40% → Oct: 1.25% → Nov: 1.26% → Dec: 1.31% → Jan 2026: 1.42% → Feb: 1.30% → Mar: 1.25%
Outcome language signal (BH-FDR corrected across 10 signals)
After correcting for 10 simultaneous linguistic signals tested across n=1050 pooled news brand notifications (2025+2026), only two survive BH–FDR correction. Crime/death outcome words are the strongest single notification signal in the dataset—stronger than attribution language, stronger than questions, and stronger than any other tested feature.
| Signal | Description | Lift | n | p_raw | p_adj (BH–FDR) |
|---|---|---|---|---|---|
| Crime/death outcome words | dead / died / killed / arrested / charged / convicted / sentenced / shot | 1.26× | 55 | p=0.000151 *** | p=0.0015 *** |
| Attribution language | says / said / told / reports / reveals | 1.18× | 59 | p=0.004 ** | p=0.020 * |
| All other signals tested (questions, urgency words, local geo-anchor, superlatives, named person lead, money amounts, opinion flags, possessives): not significant after FDR correction. | |||||
Apple News push notifications Jan 2025–Mar 2026 (n=1923 with valid CTR; 2025 includes Us Weekly all year, news brands from June 2025 only). Monthly CTR = median within each calendar month for news brand notifications only. Analyses run separately within each brand-type population; BH–FDR correction applied within each set of tests. Mann-Whitney U; effect size = rank-biserial r; 95% CIs via 1,000-iteration bootstrap. Feature classifier unvalidated.
Finding A · Formula × Topic Interaction (Apple News Featuring)
Editorial guidance by topic
Crime: Use "Here's"—16% featuring rate (n=89), vs. Question at 13% (n=94). Both are viable; "Here's" has a modest edge and is the recommended default.
Business: Use "Here's what [X] means for [community]"—14% featuring rate (n=72), the strongest formula with adequate sample size.
Sports: No formula affects featuring odds—Number lead 0% (n=31), What to Know 0% (n=22), Here's 11.5% (n=52). Topic overrides formula. Optimize for organic SmartNews distribution instead (see Finding 3).
Weather: Featuring rates (53–71%) reflect United Robots automated content, not editorial choices. Exclude from any editorial formula guidance.
Pooled Apple News 2025 + ANP 2026 (n>15,000 articles). For editorial content (non-weather), formula choice has a real but moderate effect on featuring odds—the interaction is topic-conditional, not a universal signal. The overall featuring lift attributed to "Here's" in other analyses is substantially inflated by weather content pulling the average upward.
| Formula | Topic | n | Featuring rate |
|---|---|---|---|
| Here's / Here are | Weather | 102 | 71% |
| Question | Weather | 67 | 57% |
| Number lead | Weather | 19 | 53% |
| Question | Education | 45 | 20% |
| Explainer | Food | 38 | 18% |
| Here's / Here are | Real estate | 28 | 18% |
| Here's / Here are | Crime | 89 | 16% |
| Here's / Here are | Business | 72 | 14% |
| Question | Crime | 94 | 13% |
| Here's / Here are | Sports | 52 | 12% |
| Number lead | Sports | 31 | 0% |
| Possessive | Other | 15 | 0% |
| What to know | Sports | 22 | 0% |
Pooled Apple News 2025 full year + ANP 2026 YTD article-level data (n>15,000). Formula classifier unvalidated; topic classifier unvalidated. Featuring rate = share of articles in each formula × topic cell selected for Apple News Local News Featured slot. Cells with n<15 excluded. Kruskal-Wallis across all combinations significant (p<0.05). Interaction interpretation: observational—causal direction unconfirmed. Weather rows included for completeness; exclude from editorial formula guidance.
Finding B · SmartNews Cross-Platform Formula Trap
SmartNews 2025, n=38,251 articles. SmartNews ranks articles by a percentile-within-cohort metric (0.5 = exactly average). The question format drops to 0.423 pct_rank (p=3.4e-6, n=918)—0.08 below baseline. "What to know" drops to 0.371 (p=3.0e-6, n=213)—the worst-performing formula on SmartNews. Meanwhile both are strong for Apple News featuring.
| Formula | Apple News feat% | SmartNews pct_rank | SN baseline | SN p-value | Cross-platform verdict |
|---|---|---|---|---|---|
| Here's / Here are | 46.5% | 0.543 | 0.500 | p=0.038 (dir.) | Safest cross-platform (directional) |
| Number lead | 17.3% | 0.534 | 0.497 | p=0.29 (dir.) | SN-leaning |
| Direct declarative | ~20% (baseline) | 0.500 | 0.500 | — | Neutral |
| Explainer | — | 0.491 | 0.501 | p=0.62 | Neutral |
| Question | 40.9% | 0.423 | 0.502 | p=3.4e-6 *** | Apple only—hurts SN |
| What to know | ~52% (best AN) | 0.371 | 0.501 | p=3.0e-6 *** | Apple only—worst SN |
Practical guidance
- Cross-platform default: Use "Here's / Here are" format—the only formula directionally above baseline on both Apple News featuring AND SmartNews rank (p=0.038, directional—does not survive Bonferroni correction, but is the best available cross-platform option).
- Apple News-only variants: Use question format or "What to know"—but write a separate SmartNews headline if the story is also going there.
- SmartNews-focused: Number lead format is directionally positive (0.534, p=0.29)—worth testing when volume permits. Direct declarative is the safe floor.
- Never: Route an Apple News "What to know" or question headline unchanged to SmartNews. Both formats are statistically penalized on that platform.
SmartNews 2025 (n=38,251 articles; politics excluded). Metric: percentile-within-cohort (same outlet × month). Mann-Whitney U tests, each formula vs. untagged baseline; BH–FDR correction applied. Apple News featuring rates from pooled 2025+2026 dataset. "What to know" Apple News rate is the best-fit from Q2 analysis. Cross-platform comparison is observational—audiences and algorithmic mechanisms differ across platforms.
SmartNews channel distribution—T1 content (2026, n=217 articles)
| Channel | Total Views | Share % |
|---|---|---|
| Top | 5,120,333 | 86.3% |
| Entertainment | 570,440 | 9.6% |
| Local | 37,236 | 0.6% |
| Health | 23,718 | 0.4% |
| U.S. | 3,473 | 0.1% |
| Lifestyle | 2,769 | 0.0% |
| Business | 1,381 | 0.0% |
| World | 914 | 0.0% |
SmartNews 2026 data, T1 outlets only (excluding Us Weekly). Jan–Mar 2026, n=217 articles. Channel views may not sum to article_view (some views are unattributed to a channel). New in April 2026 data export.
MSN · Formula Divergence
Implication: Write direct, subject-verb-object MSN headlines. Do not repurpose Apple News "Here's" or "What to know" copy for MSN without rewriting.
Across 430 T1 news brand articles on MSN (Jan–Mar 2026, excluding celebrity/entertainment publishers), direct declarative headlines (classified as "untagged"—no formula pattern detected) have a median 3,148 pageviews. Formula-tagged groups with ≥5 articles:
| Formula | n | Median PVs | Lift vs. direct declarative | padj (BH–FDR) |
|---|---|---|---|---|
| Direct declarative (baseline) | 356 | 3,148 | 1.00× (baseline) | — |
| Quoted lede | 30 | 11,745 | 3.73× * | p=0.048 * |
| Possessive named entity | 9 | 6,538 | 2.08× | p=0.295 |
| Number lead | 8 | 2,172 | 0.69× | p=0.710 |
| Question | 22 | 1,970 | 0.63× * | p=0.048 * |
Cross-platform contrast
The inversion is sharpest for structured formulas that Apple News rewards. "Here's" and question formats predict Apple News featuring at 3–4× the baseline rate; those same formats cut MSN traffic vs. direct statements. Possessive named entity (e.g., "Smith's…") is one of the strongest Apple News non-featured signals but underperforms on MSN. The implication is not that one format is universally better—it is that the optimal format is platform-specific.
Top MSN articles are direct declaratives
The highest-volume MSN articles in this dataset are uniformly direct, declarative statements:
- Scandal Erupts Over Texts in Epstein Case (Fort Worth Star-Telegram)—882,938 PVs
- Schiff Responds to Court-Martial Threat (Miami Herald)—514,225 PVs
- Scientists Respond to RFK Jr.’s Aluminum Alarm (Miami Herald)—424,636 PVs
These headlines share a common structure: subject + verb + direct object, no structural signal words, no question mark, no "Here's." The story event does the work; the headline reports it plainly.
MSN Jan–Mar 2026 (n=430 T1 news brand articles; Us Weekly / Woman's World excluded). Formula classification via unvalidated regex classifier. "Untagged" = no formula pattern detected (treated as direct declarative baseline). Mann-Whitney U tests, each formula vs. untagged baseline; BH–FDR correction applied across formula comparisons. Effect sizes not shown—use with caution at this sample size. Pageview minimum: MSN sheet filter applies (≥10k PV threshold in source data).
Finding 3 · Platform Topic Inversion
Sports ranks #2 on Apple News (percentile index 1.10× platform median) but #8—last—on SmartNews (index 0.78×). This is not a small difference: sports sits well above the Apple News median and well below the SmartNews median. The inversion is statistically significant (Mann-Whitney U, p<0.001 ***) across the full year of 2025 data.
Nature/wildlife shows the reverse: it underperforms the Apple News median (0.74×) but outperforms the SmartNews median (1.22×). Among the top 30 most frequent words in top-quartile headlines on each platform, only 9 words appear on both lists (cops, death, found, off, photos, say, shares, star, years)—generic reporting terms rather than shared topical vocabulary, suggesting the platforms reward very different content angles.
Nature/Wildlife: headline formula by platform (dual-headline guide)
| Formula | Apple News (n, pct_rank, feat%, p) | SmartNews 2025 (n, pct_rank, p) | |||||
|---|---|---|---|---|---|---|---|
| n | Pct rank | Feat% | p | n | Pct rank | p | |
| What to know | 2 | 0.839 | 100% | — | 2 | 0.321 | — |
| Question | 4 | 0.796 | 50% | — | 14 | 0.421 | p=0.820 |
| Possessive named entity | 2 | 0.694 | 50% | — | 11 | 0.305 | p=0.097 |
| Quoted lede | 50 | 0.379 | 22% | p=0.677 | 23 | 0.660 | p=0.650 |
| Number lead | 24 | 0.288 | 0% | p=0.409 | 16 | 0.662 | p=0.476 |
| Here's / Here are | 0 | — | — | — | 0 | — | — |
Apple News: pooled 2025+2026 nature/wildlife articles (n=294). SmartNews: 2025 article-level data only (n=475); 2026 SN data is domain-aggregated and invalid for formula analysis. Tests: Mann-Whitney U vs. all other nature/wildlife articles of different formula. Low n per formula cell—treat as directional only.
Sports underperforms across Apple News featuring and notifications
The Apple News/SmartNews sports inversion extends to additional signals, making sports the most consistently platform-penalized topic in this dataset:
- Apple News notifications: Sports notification CTR 1.02%—lowest topic—vs. crime 1.66%, weather 1.59%. Sports drives Apple News organic views (1.10× platform median) but Apple doesn't feature it and it doesn't click-through on push.
The exception remains SmartNews, where sports performs organically (topic ranking #8 from bottom—but at 0.78× median, still below platform average). The pattern is consistent: sports is a reader-intent topic (fans seek it out directly) rather than a distribution topic (algorithms don't surface it). This has direct implications for headline strategy—sports headlines should prioritize depth over reach, with different playbooks for each platform.
Sports subtopic performance by platform
Within the sports inversion: which sports specifically drive Apple News performance, and which are weakest on SmartNews? The table below breaks sports into subtopics (via two-level headline classifier).
| Sport | Apple News n | Apple News median %ile | SmartNews n | SmartNews median %ile |
|---|---|---|---|---|
| hockey | 10 | 72% | 17 | 30% |
| football | 166 | 60% | 540 | 49% |
| baseball | 5 | 55% | 34 | 39% |
| college | 5 | 54% | 37 | 45% |
| sports_other | 85 | 50% | 1404 | 35% |
| basketball | 22 | 30% | 77 | 35% |
| soccer | 0 | — | 6 | 29% |
Politics subtopic performance by platform
Within politics, which story type drives the most engagement on each platform?
| Subtopic | Apple News n | Apple News median %ile | SmartNews n | SmartNews median %ile |
|---|---|---|---|---|
| Federal | 0 | — | 0 | — |
| State | 0 | — | 0 | — |
| Local Govt | 0 | — | 0 | — |
| Election | 0 | — | 0 | — |
| Politics Other | 0 | — | 0 | — |
Top quartile politics headlines
Bottom quartile politics headlines
Topic tagged via unvalidated regex classifier applied to headline text. Coverage: 51% of Apple News articles match a named topic; 49% fall into "other/unclassified" and are excluded from this analysis. Results describe the classified minority—generalizing to all content requires caution. Percentile index = median percentile_within_cohort / platform overall median percentile. Apple News 2025–2026 (n=4,283); SmartNews 2025 (n=37,695); MSN 2025 (n=746). Subtopic classifier unvalidated. No significance testing—treat as descriptive. Subtopics with n<3 show "—".
Finding 5 · Trends Over Time
Chart shows the three highest-volume formulas (≥15 articles/quarter). Here's/Here are and What to know are in the table below—their n=4–9/quarter is too small to distinguish trend from noise.
Lift by formula across periods
Each cell shows the median lift for that formula in that period, relative to unclassified articles published in the same period. A dash means fewer than 5 articles qualified.
| Formula | Q1 2025 | Q2 2025 | Q3 2025 | Q4 2025 | Q1 2026 | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| Number lead | 0.68× | (n=38) | 0.75× | (n=72) | 0.87× | (n=41) | 0.92× | (n=19) | 1.06× | (n=48) |
| Question | 0.66× | (n=45) | 0.64× | (n=20) | 0.96× | (n=31) | 0.68× | (n=28) | 0.48× | (n=54) |
| Possessive named entity | 1.32× | (n=20) | 1.25× | (n=16) | 1.10× | (n=29) | 1.24× | (n=21) | 0.93× | (n=36) |
| Here's / Here are | 1.23× | (n=9) | — | (n=—) | 1.41× | (n=7) | 0.90× | (n=8) | 0.13× | (n=4) |
| What to know | 1.56× | (n=9) | 0.42× | (n=4) | 1.52× | (n=4) | 1.00× | (n=4) | 0.43× | (n=8) |
Quarters: Q1=Jan–Mar, Q2=Apr–Jun, Q3=Jul–Sep, Q4=Oct–Dec. Q1 2026 = Jan–Feb 2026 only. Lift = formula median cohort percentile ÷ untagged baseline median within same quarter. Minimum 3 articles required per cell. Data through April 2026.
Character length sweet spot by formula type
For each formula used organically (non-featured articles), the character-length bin associated with the highest median view percentile. Use as a per-formula length target when writing headlines for Apple News.
| Formula | n (organic) | Best length bin | Median %ile in best bin | n in best bin | p (best vs. rest, one-tailed) |
|---|---|---|---|---|---|
| Possessive named entity | 98 | 90–109 chars | 71% | 18 | p<0.05 |
| Number lead | 191 | 90–109 chars | 56% | 19 | p<0.05 |
| Quoted lede | 113 | 90–109 chars | 50% | 8 | p<0.05 |
| Question | 111 | 90–109 chars | 38% | 12 | p=0.10 |
Apple News non-featured articles only (organic signal). Mann-Whitney U one-tailed: best bin vs. all other bins for that formula. Uncorrected for multiple comparisons across formula types—treat as directional guidance.
What a winning headline looks like—vs. what underperforms
Top and bottom organic Apple News articles by formula (2025–2026). Gives editors concrete examples, not just statistics.
Here's / Here are
| Tier | Headline | Views | Month |
|---|---|---|---|
| Top | Here’s the drama that went on behind the scenes with Melania’s viral inauguration hat | 326,572 | 2025-01 |
| Top | Here’s the facemask call that went against the Bills in AFC Championship Game | 42,626 | 2025-01 |
| Bottom | Here’s where you can carry guns in GA. It may be more places than you think | 208 | 2025-10 |
| Bottom | Here are the places in GA you’re most likely to catch the vomiting disease spreading in the state | 241 | 2025-12 |
Possessive named entity
| Tier | Headline | Views | Month |
|---|---|---|---|
| Top | HGTV’s Erin Napier Claps Back at ‘Sick’ Messages She’s Received for Refusing to ‘Celebrate’ Charlie Kirk’s Death | 483,772 | 2025-09 |
| Top | NFL's Ben Cleveland Files for Divorce 2 Weeks After DUI, Cites 'Adultery' | 373,964 | 2025-03 |
| Bottom | Bradenton’s new Target store near I-75 announces grand opening date | 385 | 2025-07 |
| Bottom | July’s heat and August’s rain are not complying with Donald Trump on climate | Opinion | 417 | 2025-08 |
Question
| Tier | Headline | Views | Month |
|---|---|---|---|
| Top | Fast Times at Ridgemont High’s Phoebe Cates’ Unexpected Success Outside Hollywood: Where Is She Now? | 403,564 | 2025-07 |
| Top | How Did Yellowstone's 'Marshals' Spinoff Write Off Kelsey Asbille's Monica After Offscreen Death? | 181,352 | 2026-03 |
| Bottom | Taco Bell puts new spin on a fan favorite, adds 2 frozen drinks. When can you get them? | 191 | 2025-01 |
| Bottom | Lottery player wins life-changing prize at Florida Publix. Where was lucky ticket sold? | 223 | 2025-02 |
Number lead
| Tier | Headline | Views | Month |
|---|---|---|---|
| Top | 10 of the Most Influential Celebrity Naked Dresses of All Time | 483,735 | 2025-02 |
| Top | 17 Universally-Flattering Wedding Guest Dresses for Women Over 40—Starting at $25 | 483,310 | 2025-04 |
| Bottom | 36-foot emaciated sea creature washes up on California beach. Its death is a mystery | 249 | 2025-04 |
| Bottom | 14-year-old boy stabbed his father to death in middle of the night, Georgia cops say | 252 | 2025-01 |
Quoted lede
| Tier | Headline | Views | Month |
|---|---|---|---|
| Top | ‘Weird-looking’ bug with ‘bulging eyes’ found for first time in 2 states. See it | 22,861 | 2025-05 |
| Top | ‘Monstrous’ sea creature found washed up dead on Oregon beach, video shows | 21,968 | 2025-06 |
| Bottom | ‘Large’-eyed creature with ‘slender’ toes found lurking. It’s a new species | 182 | 2025-05 |
| Bottom | ‘What is that?’ How a local fisherman reeled in a rare snapper in the Gulf of Mexico | 360 | 2025-01 |
Apple News 2025+2026 pooled. "Top" = highest organic views (non-featured where featured flag available). "Bottom" = lowest views among articles with ≥100 total views. Examples are illustrative—individual article performance depends on story quality, timing, and audience. Use to calibrate writing, not as a strict formula → outcome model.