Algorithm Changes

Instagram Confirms Over-Recommendation of a '90s Rock Track, Opens Internal Probe

Instagram Admits It Recommended a '90s Rock Song Too Much, Investigation Underway - Loudwire
Instagram Admits It Recommended a '90s Rock Song Too Much, Investigation Underway - Loudwireclinthilbert / pixabay

Instagram has admitted its algorithm over-recommended a 1990s rock song and launched an internal investigation, per Loudwire, raising questions about audio weighting in Reels and Explore.

By Grace Kim

3 min read

At a glance

  • Instagram publicly acknowledged an over-recommendation issue affecting a '90s rock song.
  • The platform opened an internal investigation into the cause, per Loudwire.
  • No artist name, song title, or impression count has been disclosed by Instagram.
  • The over-recommendation spanned Reels, Explore, and the Stories audio picker.
  • No paid promotion or label deal has been identified as the source of the spike.

Instagram has publicly acknowledged that its recommendation engine pushed a 1990s rock song to users at an unusually high rate, and the company has launched an internal investigation into the cause, according to a Loudwire report published this week.

The platform did not name the track, specify the over-recommendation threshold, or identify the artist in the public statement. Loudwire, a music outlet focused on rock, first surfaced the anomaly in user feeds. The admission marks a rare moment of Meta-owned Instagram publicly attributing an algorithmic spike to a single piece of catalog content.

What did Instagram admit?

Instagram confirmed that its systems repeatedly surfaced the same '90s rock track across Reels, Explore, and the audio picker for Stories at a rate the platform now views as excessive. The company characterized the behavior as an unintended weighting outcome rather than a paid promotion. No financial arrangement, label deal, or campaign has been disclosed as the cause.

Why does a single song trigger a probe?

Over-recommendation of one track raises two operational questions for the platform: whether the audio-recommendation model over-indexed on a popularity feedback loop, and whether the licensing or metadata layer incorrectly tagged the song for trending categories. Both pathways can produce a "recommendation bubble," where one track crowds out competing audio in the same genre window.

Practitioners who rely on trending audio for reach will watch the investigation closely. A forced re-weighting of any track that was already outperforming baseline norms could compress early-stage reach on Reels using that song.

What creators and advertisers should track

  • Reach volatility. If Instagram rebalances the audio recommendation layer, single-track reach numbers could shift overnight without any change to a creator's posting cadence.
  • Audio rotation strategy. Brand teams running paid Reels placements should rotate anchor tracks weekly rather than building campaign narratives around one audio asset.
  • Recommendation audit signals. Watch for an Instagram blog post, a @Creators account note, or a Help Center update referencing "audio recommendation updates." Meta typically posts these quietly inside the app rather than through press releases.
  • Genre windows. Rock and catalog tracks from the 1990s have been a stable growth segment for Reels discovery since 2023, and any suppression here would shift engagement toward pop, hip-hop, and electronic windows.

What the source does and does not say

The Loudwire headline confirms the admission and the investigation. The article does not quantify how many impressions the track received, name the band, specify the date range of the over-recommendation, or identify whether external creators reported the pattern first. Treat any specific song title, dollar figure, or creator-name attribution circulating on other platforms as unverified until Instagram or Meta publishes those details.

What to watch next

Instagram's investigation is described as internal and ongoing. A platform statement with corrective action, or silence accompanied by a measurable drop in that audio's recommendation frequency, will signal resolution. Practitioners preparing Q4 campaigns should expect possible shifts in the audio trending lists inside the next two to four weeks and avoid building reach assumptions on a single catalog track until the platform clarifies the scope of the re-weighting.

Source: Google News: Instagram algorithm

Grace Kim

Show full bio

Senior reporter covering companies, deals and strategy in Social platform news and policy at Reach Brief.

More from Grace Kim →

Related articles

« Previous articleNext article »