Tony Wang10 min readX's App Store Rating Hasn't Moved Since Musk Bought It. The Reviews Have.
We mined 92,088 App Store and Google Play reviews of X. The 4.6-star rating never budged — but written-review sentiment and topics shifted sharply after 2022.
When Elon Musk closed his $44 billion purchase of Twitter on October 27, 2022, it was one of the most scrutinized corporate takeovers in tech history. Nearly four years later, the app's App Store rating sits at 4.6 stars — exactly where it was the week the deal closed. That stability is itself the surprise, and it's misleading. We pulled 92,088 reviews of the X app from the iOS App Store and Google Play, reaching back to 2009, and theme-coded them around the acquisition date. The rating didn't move. Almost everything underneath it did.
To be clear up front about what this is and isn't: this is a measurement of what people write in app reviews, before and after a date. It is not a referendum on whether X is a better or worse product now, and we don't offer one. Reviews are written by a self-selected, opinionated minority. What they're useful for is showing how the composition of that opinion changed — which topics rose, which fell, and which, contrary to the popular story, didn't move at all.
The number that didn't move
Three point-in-time snapshots of the App Store listing — two from the Internet Archive, one live today — tell the same story:
| Date | Rating | Total ratings | Developer |
|---|---|---|---|
| Oct 28, 2022 (deal closed) | 4.6 | 6.6M | Twitter, Inc. |
| May 30, 2024 (post-rebrand) | 4.6 | 8.6M | X Corp. |
| Jul 16, 2026 (today) | 4.6 | 11M | X Corp. |
The aggregate is an all-time average weighted by millions of silent tap-a-star ratings, so it moves at a glacial pace and heavily favors years of accumulated goodwill. It is designed not to move. That's exactly why it hides what changed.
What changed: written-review sentiment fell off a cliff in 2023
When you look only at reviews where someone wrote text — a much smaller, much louder population — the mean star rating tracks a clear break at the acquisition. Across 10,861 English-language App Store reviews:
Stars are language-independent, so the same trajectory holds across all 78 storefronts in the full corpus (2022: 3.56 to 2023: 2.66).
The share of one-star written reviews more than doubled across the acquisition:
Again, the caveat that keeps this honest: these are people motivated enough to write, and the App Store's "most helpful" sort surfaces heavily-upvoted reviews, which skew critical. A doubling of one-star written reviews is a real shift in the vocal population — it is not a claim that 44% of X's users are unhappy. The silent majority is what keeps the aggregate at 4.6.
What changed most: the reviews became about Elon Musk
We theme-coded every English review with a keyword classifier across fourteen categories. The single largest shift, by a wide margin, is that reviews started being about the owner:
Before the deal, 0.8% of reviews mentioned Elon Musk. After, 16.4% did — one in six. A social app's reviews are usually about the app; X's became partly about its owner. The other risers map cleanly onto real product changes: subscription/pricing complaints tripled (3.2% → 9.2%) after Twitter Blue, later X Premium, launched the same week the deal closed; a Grok/AI theme that was essentially absent (0.1%) now appears in 4.6% after xAI's assistant was folded into the app; and free-speech praise more than doubled (3.2% → 7.4%), the clearest signal of the supportive camp.
The full before/after table, including the themes that fell:
| Theme | Before | After | Change |
|---|---|---|---|
| Mentions Elon Musk | 0.8% | 16.4% | +15.6 |
| Subscription / pricing | 3.2% | 9.2% | +6.0 |
| Grok / AI | 0.1% | 4.6% | +4.4 |
| Free-speech praise | 3.2% | 7.4% | +4.3 |
| Bots / spam | 2.3% | 6.4% | +4.1 |
| Trolls / hate speech | 6.7% | 8.0% | +1.4 |
| Censorship / bias complaint | 7.7% | 6.6% | −1.1 |
| Account lock / ban | 9.2% | 8.2% | −0.9 |
| Algorithm / feed | 9.9% | 8.3% | −1.6 |
| Bugs / reliability | 10.5% | 8.7% | −1.9 |
| Quality praise | 12.0% | 7.1% | −4.9 |
The narrative the data doesn't support
The most politically loaded claim about post-Musk Twitter — that it became dramatically more or less censored — is not what the review topics show. Censorship and bias complaints were 7.7% of reviews before the acquisition and 6.6% after: essentially flat, and slightly down. Complaints about moderation, shadow-banning, and political bias were already one of the most common review themes under the old ownership. The oldest example in our set is titled "Locked out of my account for political reasons" — from November 2017, five years before Musk.
This cuts against a tidy story in both directions. Reviewers who expected the data to show censorship complaints vanishing after a "free speech" takeover, and reviewers who expected them to spike, will both be disappointed: the theme barely moved. What moved was everything around it — the owner's name, the subscription, the AI, and the overall sentiment.
Meanwhile, quality praise fell the most of any theme (12.0% → 7.1%). Fewer people writing reviews are there to say they love the app.
What reviewers actually wrote, both sides
The reviews got more polarized, not uniformly negative. Verbatim, kept short and shown from both camps and both eras:
Before the acquisition:
"I'm grateful for Twitter" — 5★, 2021 "Locked out of my account for political reasons" — 1★, 2017
After, supportive:
"I'm so grateful to Elon Musk." — 5★, 2023 "In X we trust" — 5★, 2023
After, critical:
"Bring back Twitter" — 1★, 2024 "Elon's Pathetic Echo Chamber" — 1★, 2025
We're quoting these to show the polarization, not to endorse either side. Both grew louder at once.
Cross-platform: where the comparison holds and where it doesn't
| Surface | Reviews analyzed | Historical reach | Role in this study |
|---|---|---|---|
| iOS App Store (93 storefronts) | 49,723 | 2009 → 2026 | The before/after spine |
| Google Play (deep pagination) | 42,365 | mostly 2026; thin tail to 2018 | Current-state cross-check |
| Chrome Web Store | 0 official | — | No first-party X extension exists |
The before/after comparison rides entirely on the App Store, because its per-country feeds reach back years. Google Play's public feed is overwhelmingly recent — X collects roughly 39,700 Play reviews in 2026 alone, so paginating back to 2022 at scale isn't feasible; we use it only to confirm the current-state picture. And unlike some apps we've studied, X ships no official browser extension — we confirmed this by searching the Chrome Web Store for X Corp. and Twitter, Inc. as developers (zero results). The third-party ecosystem around X (Control Panel for Twitter, ~300K users, and others) is real but not built by the company, so it isn't comparable.
What holds up, and what doesn't
What holds up: the displayed rating is stable while written-review sentiment dropped sharply and topic composition shifted, both breaking cleanly at the 2022–2023 boundary. The Elon-mentions, subscription, and Grok shifts map onto documented product changes. The censorship-complaint theme, contrary to the loudest narratives, did not rise.
What limits it:
- Written reviews are not users. They are a self-selected, opinion-heavy minority, and "most helpful" sorting over-samples highly-upvoted (often critical) reviews. Read the shifts in what this vocal group writes, not the absolute levels as a poll.
- Theme coding is keyword-based, not semantic. It undercounts anyone expressing a theme without its keywords, and it counts a mention, not a verified fact. "Reviewers mention X more" is not "X is more true."
- The corpus spans ~78 storefronts and many languages. Star ratings are language-independent (used for the whole trajectory), but theme coding runs only on the 10,861 English-majority-storefront reviews, a smaller sample.
- Google Play can't corroborate the pre-2022 period at scale, so the before/after rests on the App Store alone.
- Snapshot rating figures come from three archived/live points, not a continuous series; the aggregate's stability is directional.
Sources
Methodology
Every number here was pulled through Crawlora's own APIs on July 16, 2026: App Store reviews fanned across 93 country storefronts at full feed depth (Apple caps its public feed at 10 pages × 2 sort orders per storefront, but each storefront is an independent feed reaching back years), and Google Play reviews via continuation-token deep pagination. Point-in-time rating snapshots were read from the Internet Archive and the live listing. The Chrome Web Store check used the chrome-extensions dataset by developer name.
Corpus: 92,088 reviews — 49,723 App Store (78 storefronts with data, 2009–2026) and 42,365 Google Play. Theme coding used a documented keyword classifier across fourteen categories on the 10,861 English-majority-storefront App Store reviews (2,735 before the acquisition, 8,126 after); a review may carry more than one theme. We publish aggregates only — no raw review text tied to individual reviewers. The full aggregate dataset — theme before/after, star-by-year, and rating snapshots — is open on GitHub under CC BY 4.0.
Reproduce it in the Playground; the how to scrape App Store reviews and how to scrape Google Play guides walk through the calls, and our app review sentiment dataset has millions of pre-scraped reviews. For the same review-mining method applied to the Claude app, see what 1,610 reviews of the Claude app reveal.
Frequently asked questions
Did X's app rating drop after Elon Musk bought Twitter?
The displayed App Store rating did not — it has been 4.6 stars throughout, at 6.6M ratings when the deal closed in October 2022, 8.6M by mid-2024, and 11M today. That aggregate is an all-time average weighted by millions of silent ratings, so it barely moves. But among the smaller population who write text reviews, sentiment fell sharply: across 10,861 English App Store reviews, the mean written-review star dropped from 3.81 before the acquisition to 2.95 in the X era, and the one-star share of written reviews rose from 17.8% to 43.6%.
What do people complain about in X app reviews after the acquisition?
The topics shifted more than the scores. Mentions of Elon Musk went from 0.8% of English reviews before the deal to 16.4% after — the single largest change. Subscription and pricing complaints tripled (3.2% to 9.2%) after Twitter Blue / X Premium launched, a Grok/AI theme rose from near-zero to 4.6%, and free-speech praise more than doubled (3.2% to 7.4%). Quality praise fell the most, from 12.0% to 7.1%.
Did censorship complaints about X increase after Musk took over?
No — this is the narrative the data does not support. Censorship and bias complaints were 7.7% of English App Store reviews before the acquisition and 6.6% after, essentially flat and slightly down. Complaints about moderation and political bias were already common under the old ownership; the oldest example in our corpus, a review about a politically motivated account lockout, is from 2017, five years before Musk.
How many X reviews were analyzed, and how?
92,088 reviews — 49,723 from the iOS App Store (fanned across 93 country storefronts, reaching back to 2009) and 42,365 from Google Play — pulled via Crawlora's APIs on July 16, 2026. Star-rating trends use the full multi-language corpus (stars are language-independent); theme coding uses a documented keyword classifier on the 10,861 English-majority-storefront App Store reviews (2,735 before the acquisition, 8,126 after). We publish aggregates only, no raw review text tied to reviewers.
Does this study say whether X got better or worse under Musk?
No. It measures what a self-selected, opinionated minority writes in app reviews, before and after a date — not overall product quality or user satisfaction. Written reviews skew toward strong opinions and got more polarized in both directions after 2022, with free-speech praise and one-star reviews rising at the same time. We report the composition shift and take no position on whether the app improved.