Living index · updated daily
No — not on 28 days of daily data. Across 40 major AI apps tracked every day from July 1, 2026 to August 17, 2026, a rise in an app's Google search interest did not precede a US App Store rank climb at any lag from 0 to 7 days. The mean rank correlation stayed between -0.024 and 0.086 — indistinguishable from zero — across the 10 apps that charted often enough to test. Search demand and store rank behave like two separate markets, and this page re-runs the test every time the panel grows.
No lead detected
at lags of 0, 1, 2, 3, 5 or 7 days
Spearman rank correlation between an app's Google search interest k days ago and its US App Store position today. Positive would mean search leads the store. Apps need at least 14 charting days to qualify.
-0.024 to 0.086
Range of the mean correlation across all six lags tested. Every one is well inside the noise band for a panel this short.
10 of 40
Apps with enough chart coverage to test at all. The other 30 either never chart or never move off #1.
No consistent sign
Per-app coefficients scatter across both signs at every lag. Nothing survives as a pattern once you look app by app.
| Lag | Mean correlation | Apps | Per-app range | Reading |
|---|---|---|---|---|
| 0 days | 0.020 | 10 | -0.40 to 0.28 | No effect |
| 1 day | 0.007 | 10 | -0.35 to 0.26 | No effect |
| 2 days | 0.086 | 9 | -0.41 to 0.53 | No effect |
| 3 days | 0.050 | 9 | -0.26 to 0.28 | No effect |
| 5 days | 0.063 | 9 | -0.27 to 0.35 | No effect |
| 7 days | -0.024 | 9 | -0.38 to 0.35 | No effect |
| App | Charting days | Lag 0d | Lag 1d | Lag 2d | Lag 3d | Lag 5d | Lag 7d |
|---|---|---|---|---|---|---|---|
| Google Gemini | 28/28 | 0.25 | -0.34 | -0.26 | -0.26 | 0.05 | -0.20 |
| Claude | 21/28 | 0.28 | 0.14 | 0.53 | 0.22 | -0.20 | -0.18 |
| Grok | 23/28 | 0.02 | 0.26 | 0.10 | -0.10 | -0.27 | -0.23 |
| Microsoft Copilot | 28/28 | -0.01 | -0.08 | -0.41 | -0.14 | 0.18 | 0.35 |
| Meta AI | 25/28 | -0.40 | -0.35 | -0.09 | 0.13 | -0.14 | -0.38 |
| Doubao | 24/28 | 0.24 | 0.24 | 0.41 | 0.20 | 0.35 | 0.22 |
| Canva | 22/28 | 0.03 | 0.02 | 0.17 | 0.28 | 0.15 | 0.18 |
| Udio | 28/28 | -0.23 | -0.03 | — | — | — | — |
| Gauth | 24/28 | -0.24 | -0.04 | 0.21 | 0.01 | 0.11 | 0.19 |
| Photomath | 28/28 | 0.25 | 0.25 | 0.12 | 0.11 | 0.34 | -0.17 |
Read this as a negative result, not a null law. 28 days is a short panel, the testable sample is 10 apps, and the three inputs run on three different clocks (below). A real search-to-store lead could exist and still be invisible here. What the panel does rule out is a large, consistent, short-horizon effect across major AI apps — the version of the claim that would be useful to act on.
Everything that constrains the analysis above, measured from the panel itself rather than asserted.
| Signal | Value changes | Median observations between changes | Distinct values per app (median, 28 days) | Usable daily? |
|---|---|---|---|---|
| Google Trends interest | 787 | 1 | 13 | Yes, with a weekly artifact |
| App Store chart rank | 151 | 1 | 0 | Yes, where an app charts |
| SimilarWeb web visits | 78 | 19 | 3 | No — monthly estimate |
11 of 40
Apps carrying a US chart rank on an average day (range 6–16). Chart rank is the dependent variable, so everything off the top 100 is censored, and apps that enter and exit produce broken runs.
3 values in 28 days
SimilarWeb reports estimated monthlyvisits, so the median app's traffic figure changes about once a month. It is a strong absolute cross-check and a useless daily variable — we use it for levels and trends, never for lead/lag.
5 apps affected
Google Trends' newest weekly bucket is incomplete, so on rollover the reading collapses and refills through the week. 100% of the detected drops land on a Saturday. Search momentum below is a 7-day mean for exactly this reason.
Sparklines show the shape; every value they draw is printed in the row beside them. Gaps are days the app was off the US top 100.
The biggest climber was Suno (48 → 13, +35 places) and the biggest slider Claude (9 → 14, -5). Positions compare the median of an app's first and last five charting days, which is robust to the daily churn caused by an app appearing on one store chart but not the other.
| App | Charting days | Chart rank | Early → late | Search interest | Web visits |
|---|---|---|---|---|---|
| ChatGPT | 28/28 | #1 → #1 (0) | 5.5B | ||
| Google Gemini | 28/28 | #1 → #1 (0) | 2.7B | ||
| Microsoft Copilot | 28/28 | #3 → #1 (+2) | 82M | ||
| Udio | 28/28 | #1 → #1 (0) | 1M | ||
| Photomath | 28/28 | #1 → #1 (0) | — | ||
| Meta AI | 25/28 | #9 → #7 (+2) | 26M | ||
| Doubao | 24/28 | #15 → #17 (-2) | 153M | ||
| Gauth | 24/28 | #17 → #17 (0) | 3M | ||
| Grok | 23/28 | #19 → #10 (+9) | 210M | ||
| Canva | 22/28 | #26 → #8 (+18) | 822M | ||
| Claude | 21/28 | #9 → #14 (-5) | 969M | ||
| Suno | 11/28 | #48 → #13 (+35) | 71M | ||
| Perplexity | 5/28 | — | 104M | ||
| DeepSeek | 4/28 | — | 346M | ||
| Chai | 4/28 | — | 949K | ||
| Kling AI | 2/28 | — | 2M | ||
| Poe | 1/28 | — | — | 5M | |
| PhotoRoom | 1/28 | — | — | 21M |
| App | Chart rank, daily (28 values) | Search interest, daily (28 values) |
|---|---|---|
| ChatGPT | 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 | 56 65 67 71 71 49 61 61 68 69 68 49 51 59 60 64 69 68 50 54 65 65 72 68 70 47 53 53 |
| Google Gemini | 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 | 43 56 54 55 55 45 53 — 57 57 59 45 47 51 54 54 54 54 44 48 49 49 56 57 58 45 47 54 |
| Microsoft Copilot | 24 23 1 3 1 2 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 | 23 52 46 45 45 18 31 37 42 42 41 20 23 34 38 41 42 40 18 20 28 28 41 42 42 15 21 35 |
| Udio | 11 39 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 | 27 30 30 29 29 26 26 26 — 26 26 35 30 30 29 29 28 27 27 27 25 25 24 24 24 19 19 20 |
| Photomath | 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 | 5 9 9 9 9 8 6 7 8 8 7 7 7 9 10 9 9 9 8 8 10 10 — 13 13 9 10 17 |
| Meta AI | 24 7 12 9 5 7 6 8 17 9 7 5 5 — 7 7 13 4 7 4 3 33 — — 10 5 5 7 | 39 46 32 32 32 28 31 31 35 33 32 29 28 27 27 27 27 28 27 29 32 32 32 32 32 26 27 28 |
| Doubao | 19 17 9 14 15 16 2 21 11 13 15 2 23 — 8 20 27 26 17 17 — 28 16 — 27 13 — 17 | 42 71 43 42 42 22 36 35 39 38 37 30 30 35 37 36 37 34 19 27 29 29 30 31 33 24 29 37 |
| Gauth | 19 17 23 14 15 16 16 21 28 13 15 2 23 — 8 20 27 26 17 17 — 28 16 — 27 13 — 17 | 5 6 7 7 7 4 5 6 6 — 6 4 4 6 6 7 7 7 4 4 9 9 13 14 14 7 11 17 |
| Grok | 26 20 2 17 19 12 4 14 2 14 — 22 14 — 19 16 17 29 10 12 — 18 28 — 1 8 — 10 | 21 24 22 24 24 23 24 24 24 23 23 25 23 21 21 21 21 20 18 17 18 18 19 22 21 20 20 21 |
| Canva | 26 23 33 18 33 — — 33 33 29 — 13 5 9 33 9 33 — 9 33 — 5 32 8 — 32 8 8 | 31 44 51 51 51 32 44 44 — 54 54 33 38 48 54 55 58 56 36 42 55 55 65 66 63 36 43 61 |
| Claude | 11 9 9 7 12 16 13 16 17 16 11 13 13 — 16 — 16 15 — 14 — 14 14 — 14 15 — — | 41 55 54 53 53 33 43 43 53 52 51 32 37 43 43 51 48 48 30 39 40 40 46 49 49 30 40 46 |
| Suno | 69 — 68 36 — — — — — 48 — — 17 16 — 17 — — 15 — — — — 13 — — 13 13 | 46 52 48 47 47 48 50 45 45 47 48 53 50 49 49 47 47 48 54 48 48 48 49 50 51 49 44 45 |
| Perplexity | 70 72 — 69 — — 12 — — — — — 35 — — — — — — — — — — — — — — — | 17 23 17 18 18 11 15 16 16 17 17 12 13 15 16 17 18 17 12 13 15 15 15 17 17 11 15 17 |
| DeepSeek | 4 — — — — — — — — — — — — 24 — — — — — — — 33 — — — — 30 — | 51 62 55 53 53 40 52 49 50 50 55 55 68 72 71 75 74 71 42 48 49 49 57 69 67 58 73 76 |
| Chai | 87 — 86 98 — — — — — — — — 37 — — — — — — — — — — — — — — — | 50 80 82 73 73 76 75 68 73 73 67 65 65 59 56 56 56 50 56 52 50 50 48 53 51 47 43 45 |
| Kling AI | — — — 87 — — — — — — — — — 49 — — — — — — — — — — — — — — | 35 40 34 34 34 29 32 32 — 31 32 27 31 31 31 24 27 29 27 27 29 29 27 29 25 26 26 28 |
| Poe | — — — — — — 28 — — — — — — — — — — — — — — — — — — — — — | 24 28 20 20 20 — 22 22 22 22 21 16 17 17 18 18 18 18 18 19 19 19 19 20 20 15 16 16 |
| PhotoRoom | — — — 95 — — — — — — — — — — — — — — — — — — — — — — — — | 53 66 63 64 64 45 57 57 — 61 60 50 54 66 66 62 60 60 61 56 55 55 59 60 60 39 50 53 |
Chart rank and web visits are absolute and refresh with the panel. The search tier is carried from the anchored cross-sectional snapshot, because Google Trends cannot resolve a precise cross-app comparison across this magnitude spread.
8
Web-first — top half of the field by web traffic, yet never charted once in 28 days.
2
Store-first — currently charting despite sitting in the bottom half by web traffic. Distribution, not demand.
30
Aligned — the two signals agree, or the app is absent from both.
| App | Search tier (snapshot) | Chart rank | Days charted | Web visits / mo | Visit change | Search momentum | Posture |
|---|---|---|---|---|---|---|---|
| ChatGPT | Dominant | #1 | 28/28 | 5.5B | -2% | 0.79× | Aligned |
| Google Gemini | Moderate | #1 | 28/28 | 2.7B | -7% | 0.82× | Aligned |
| Claude | Moderate | Off top 100 | 21/28 | 969M | +2% | 0.96× | Aligned |
| Canva | Not tiered | #8 | 22/28 | 822M | -16% | 0.72× | Aligned |
| DeepSeek | Moderate | Off top 100 | 4/28 | 346M | -20% | 0.98× | Aligned |
| Grok | Moderate | #10 | 23/28 | 210M | -14% | 0.59× | Aligned |
| Character.AI | Not tiered | Off top 100 | 0/28 | 166M | +2% | 0.68× | Web-first |
| Doubao | Not tiered | #17 | 24/28 | 153M | -5% | 0.72× | Aligned |
| Janitor AI | Not tiered | Off top 100 | 0/28 | 143M | +30% | 1.26× | Web-first |
| Perplexity | Moderate | Off top 100 | 5/28 | 104M | -25% | 0.31× | Aligned |
| Microsoft Copilot | Low | #1 | 28/28 | 82M | -38% | 0.46× | Aligned |
| SpicyChat | Minimal | Off top 100 | 0/28 | 74M | -5% | 1.20× | Web-first |
| Suno | Not tiered | #13 | 11/28 | 71M | -13% | 0.69× | Aligned |
| Candy AI | Minimal | Off top 100 | 0/28 | 63M | +26% | 0.83× | Web-first |
| Grammarly | Moderate | Off top 100 | 0/28 | 44M | -18% | 0.51× | Web-first |
| ElevenLabs | Not tiered | Off top 100 | 0/28 | 34M | -2% | 0.71× | Web-first |
| QuillBot | Not tiered | Off top 100 | 0/28 | 31M | -30% | 0.59× | Web-first |
| Meta AI | Moderate | #7 | 25/28 | 26M | -8% | 0.57× | Aligned |
| Cursor | Not tiered | Off top 100 | 0/28 | 22M | +5% | 0.35× | Web-first |
| PhotoRoom | Not tiered | Off top 100 | 1/28 | 21M | +6% | 0.70× | Aligned |
| HeyGen | Not tiered | Off top 100 | 0/28 | 11M | +1% | 1.10× | Aligned |
| Midjourney | Low | Off top 100 | 0/28 | 10M | -8% | 0.54× | Aligned |
| Leonardo AI | Minimal | Off top 100 | 0/28 | 9M | -19% | 0.45× | Aligned |
| Fotor | Not tiered | Off top 100 | 0/28 | 7M | -22% | 0.56× | Aligned |
| Civitai | Minimal | Off top 100 | 0/28 | 7M | -12% | 0.65× | Aligned |
| Talkie | Minimal | Off top 100 | 0/28 | 7M | -3% | 1.20× | Aligned |
| Otter.ai | Minimal | Off top 100 | 0/28 | 6M | -8% | 0.58× | Aligned |
| Le Chat (Mistral) | Not tiered | Off top 100 | 0/28 | 6M | -22% | 0.37× | Aligned |
| Runway | Not tiered | Off top 100 | 0/28 | 6M | -26% | 0.46× | Aligned |
| Poe | Minimal | Off top 100 | 1/28 | 5M | -27% | 0.34× | Aligned |
| Gauth | Not tiered | #17 | 24/28 | 3M | -59% | 0.26× | Store-first |
| Ideogram | Not tiered | Off top 100 | 0/28 | 3M | -6% | 0.73× | Aligned |
| Brainly | Minimal | Off top 100 | 0/28 | 2M | -42% | 0.67× | Aligned |
| Kling AI | Not tiered | Off top 100 | 2/28 | 2M | -35% | 0.67× | Aligned |
| Udio | Not tiered | #1 | 28/28 | 1M | -20% | 0.50× | Store-first |
| Chai | Minimal | Off top 100 | 4/28 | 949K | -25% | 0.73× | Aligned |
| Remini | Not tiered | Off top 100 | 0/28 | 869K | -26% | 0.85× | Aligned |
| Hailuo AI | Not tiered | Off top 100 | 0/28 | 860K | +24% | 0.74× | Aligned |
| Replika | Minimal | Off top 100 | 0/28 | 509K | -32% | 0.55× | Aligned |
| Photomath | Not tiered | #1 | 28/28 | — | — | 0.28× | Aligned |
The original snapshot found apps that are heavily used but absent from the store charts. Over a full month of daily observations, that turns out to be structural rather than a one-day fluke.
These apps carry real, measurable web audiences — several in the tens or hundreds of millions of monthly visits — and still did not appear in the US top 100 on a single one of the 28 observed days. Web-first products, age-gated categories and browser-native tools all land here. It is the cleanest evidence on this page that App Store rank measures distribution, not demand.
| App | Web visits / mo | Visit change | Search momentum | Search tier (snapshot) |
|---|---|---|---|---|
| Character.AI | 166M | +2% | 0.68× | Not tiered |
| Janitor AI | 143M | +30% | 1.26× | Not tiered |
| SpicyChat | 74M | -5% | 1.20× | Minimal |
| Candy AI | 63M | +26% | 0.83× | Minimal |
| Grammarly | 44M | -18% | 0.51× | Moderate |
| ElevenLabs | 34M | -2% | 0.71× | Not tiered |
| QuillBot | 31M | -30% | 0.59× | Not tiered |
| Cursor | 22M | +5% | 0.35× | Not tiered |
| HeyGen | 11M | +1% | 1.10× | Not tiered |
| Midjourney | 10M | -8% | 0.54× | Low |
| Leonardo AI | 9M | -19% | 0.45× | Minimal |
| Fotor | 7M | -22% | 0.56× | Not tiered |
| Civitai | 7M | -12% | 0.65× | Minimal |
| Talkie | 7M | -3% | 1.20× | Minimal |
| Otter.ai | 6M | -8% | 0.58× | Minimal |
| Le Chat (Mistral) | 6M | -22% | 0.37× | Not tiered |
| Runway | 6M | -26% | 0.46× | Not tiered |
| Ideogram | 3M | -6% | 0.73× | Not tiered |
| Brainly | 2M | -42% | 0.67× | Minimal |
| Remini | 869K | -26% | 0.85× | Not tiered |
| Hailuo AI | 860K | +24% | 0.74× | Not tiered |
| Replika | 509K | -32% | 0.55× | Minimal |
| App | Web visits / mo | Change | Chart rank |
|---|---|---|---|
| Gauth | 3M | -59% | #17 |
| Brainly | 2M | -42% | Off top 100 |
| Microsoft Copilot | 82M | -38% | #1 |
| Kling AI | 2M | -35% | Off top 100 |
| Janitor AI | 143M | +30% | Off top 100 |
| QuillBot | 31M | -30% | Off top 100 |
| Poe | 5M | -27% | Off top 100 |
| Candy AI | 63M | +26% | Off top 100 |
| Runway | 6M | -26% | Off top 100 |
| Perplexity | 104M | -25% | Off top 100 |
| Fotor | 7M | -22% | Off top 100 |
| Le Chat (Mistral) | 6M | -22% | Off top 100 |
One row per app per day for 40 hand-picked AI apps: best US chart rank across the App Store and Google Play charts (Top Free and Grossing), SimilarWeb estimated monthly web visits, and Google Trends interest for a hand-picked search term. The panel currently holds 1,120 observations across 28 dates, from July 1, 2026 to August 17, 2026, with a longest near-contiguous run of 27 days.
Trends interest is self-normalized to each app's own trailing 12-month peak. That makes an app comparable to itself over time — which is all a lead/lag test needs — but deliberately does not make apps comparable to each other: Google Trends rescales every request against its largest term, and these apps span several orders of magnitude of search volume, so a precise cross-app ranking would be an artifact of lost precision rather than a finding. Cross-app comparisons on this page therefore use chart rank and web visits only, and the coarse search tiers come from the anchored snapshot behind the original study.
Everything above is recomputed from the panel at build time, so the correlations, tiers, trajectories and counts move as the data accrues. The joined cross-sectional snapshot is published as CSV and JSON under CC BY 4.0 on the dataset page; the same daily charts and Trends endpoints are available through the Crawlora API, and the live store leaderboard it shares data with is Top AI Apps.
Not on this evidence. Across 28 daily observations of 40 AI apps (2026-07-01 to 2026-08-17), the mean rank correlation between an app's search interest k days earlier and its US App Store position stayed between -0.024 and 0.086 for every lag from 0 to 7 days — indistinguishable from zero. Only 10 of 40 apps charted often enough to test, so this is a negative result on a small sample, not proof that no relationship exists.
App Store rank is the dependent variable, and most curated AI apps sit off the US top 100. Only 11 of 40 apps carry a chart rank on an average day (range 6–16), and 22 of 40 never charted once across the whole window. Requiring at least 14 charting days leaves 10 testable apps.
They run on three different clocks, which is the core limitation of a daily lead/lag study. Google Trends interest moves most days (787 changes across the panel). App Store rank moves daily but only where an app charts (151 changes). SimilarWeb visits are a monthly estimate: the median app takes 3 distinct values across all 28 days, with a median of 19 observations between changes — so web traffic cannot serve as a daily variable.
Google Trends returns weekly buckets, and the newest bucket is incomplete. When it rolls over, the reading collapses and then refills through the week. 5 of 40 apps show this as repeated single-day drops of 40% or more, and 100% of those drops land on a Saturday. Search momentum on this page is therefore a 7-day mean, never a single-day reading.
The 2026 study was a single cross-sectional snapshot: who ranks above or below their search demand on one day. This page is the living panel built on top of it — the same 40 apps measured every day, which adds trajectories over time and makes the lead/lag question testable at all. The anchored search tiers still come from the snapshot, because Google Trends cannot resolve a precise cross-app comparison across the magnitude spread these apps span.