Meta Chose to Bleed Its Best Surfaces to Stop TikTok. It Worked - and It Still Lost.
Reels is sold as the product that beat TikTok. It didn't. Meta spent a ~$500M-per-quarter revenue headwind and openly cannibalized its highest-margin Feed and Stories to keep users from leaving. The bet was containment, not conquest.
Comes with a free Cannibalization Decision Tree template.
In November 2019, months before any American had heard the words 'TikTok ban,' Meta was quietly testing a short-video product in Brazil. By the time Instagram Reels reached the US and 50-plus countries on August 5, 2020, the official story had already calcified: scrappy America-first scramble, Meta rushing to plug a hole TikTok had punched in its wall.1 The truth is less heroic and far more interesting. Reels was not a sprint. The Brazil test predated the US TikTok crisis by the better part of a year, which means the short-video roadmap was already in motion before the political emergency gave Meta cover to accelerate it - and the decision to aim that product directly at the most profitable real estate the company owned was one it ultimately pulled the trigger on regardless.
The popular telling says Reels answered TikTok by out-executing it. The more honest version is that Reels answered TikTok by accepting a cost - bleeding its own highest-margin surfaces, Feed and Stories, to keep users from drifting off the platform entirely. Meta didn't build Reels to win. It built Reels to not lose.
“Reels is additive to time spent. But obviously, it does have a cannibalistic impact as well… There is some engagement that we think is coming out of surfaces like Feed and Stories, and we're getting that engagement on Reels.”3
The bet: lose margin on the time you keep, or lose the time
Here is the choice Meta actually faced, stripped of the press-release gloss. A user's attention is a fixed budget of minutes per day. TikTok had figured out how to capture those minutes with an algorithmic short-video feed, and minutes spent inside TikTok are minutes not spent inside Instagram. The threat was never that TikTok would steal a feature. It was that TikTok would steal the time itself - and time-spent is the raw material Meta converts into ad revenue. Lose the minutes and the entire machine starves.
So Meta did the thing that looks like self-harm and is actually defense. It built a TikTok-shaped surface inside its own apps and fed it engagement - knowing, and saying out loud, that those minutes would come partly out of Feed and Stories.3 The catch is that Feed and Stories were Meta's best-monetized inventory, and Reels was its worst. Zuckerberg put a number on the self-inflicted wound: roughly a $500 million revenue headwind per quarter in Q3 2022, because Reels was running 'negative overall to revenue by about $500 million in the quarter.'11 Meta was, in effect, paying half a billion dollars a quarter to move its own users from a profitable room into a less profitable one - rather than let them walk out the door.
| Feed & Stories | Reels | |
|---|---|---|
| Role in the threat | What TikTok was draining time away from | The surface built to hold that time |
| Monetization per minute | High - Meta's best inventory | 'Much less than feed' |
| Near-term P&L effect | Loses engagement to Reels | ~$500M/quarter revenue headwind (Q3 2022) |
| Why do it anyway | — | Keeps the minutes inside Meta's ecosystem |
The trade only makes sense if retained minutes eventually monetize and if the alternative - losing them to TikTok - is worse. By Q1 2022 Reels had already reached about 20% of time spent on Instagram, and by January 2025 Zuckerberg told investors Reels accounted for more than half of Instagram time spent.6 The minutes were being held. Meta framed the monetization gap as a multi-year ramp, the same slow climb Stories once made.7
This was the second attempt, not the first
The clone-it instinct wasn't new, and the first try was a flop. Before Reels, Meta had shipped a standalone TikTok imitator that went nowhere. Days before Reels' US launch, TikTok's own chief executive wrote in a public blog post that it was 'another copycat product… after their other copycat Lasso failed quickly.'10 He wasn't wrong about the lineage. Zuckerberg himself has conceded that Facebook 'certainly adapted features that others have led in.'8 The strategic upgrade from the failed standalone app to Reels was not the video format - it was the placement. A standalone clone has to win the attention war from zero. A surface bolted inside Instagram inherits billions of existing users and simply redirects the minutes they already spend. The lesson Meta learned between attempts wasn't 'copy TikTok better.' It was 'don't make people leave to do it.'
So didn't Reels win? The honest answer is no
The fair objection is that the numbers look like victory. Reels grew from about a fifth of Instagram time in early 2022 to more than half by January 20256 - and you don't post engagement curves like that while losing. True. But growth inside your own walls measures retention, not conquest. The most direct evidence comes from the man with the most incentive to claim a win. Testifying under oath in the FTC antitrust trial in April 2025, Zuckerberg said adding Reels was 'in large part a response to TikTok's rise' - and, in the same breath, that 'users continue to engage more on TikTok than with his apps.'5 That is not the testimony of a company that beat its rival. It is the testimony of a company that survived it. Reels held the minutes Meta would otherwise have lost; it did not take TikTok's. Containment is a real victory. It is just not the one the narrative claims.
And the containment was expensive in more ways than one. Reels arrived into a brutal year for Meta's ads business - FY2022 revenue actually slipped 1% to $116.6 billion, with average ad price per unit down 16%.4 In a year when each unit of inventory was already worth less, voluntarily shifting users toward the inventory that monetized worst was a bet that demanded real nerve. It was the right bet only if you believe the minutes were leaving anyway.
When a new format threatens to relocate your users' attention to a competitor, the worst outcome isn't cannibalizing your own high-margin surface - it's keeping that surface pristine while the attention leaves entirely. Meta chose to move its own users from a profitable room to a less profitable one rather than let them exit the building, and ate roughly $500M a quarter to do it. But two cautions ride along. First: this only works if you can monetize the cheaper surface later - Meta is betting Reels follows the slow Stories ramp, and that bet is still maturing. Second: don't mistake retention for victory. Holding the minutes you'd have lost is defense, not offense. Reels kept Meta's users home. It never made TikTok's users leave.
The cleanest way to see Reels is not as the product that beat TikTok, but as the most expensive insurance policy Meta ever wrote on its own attention business. It paid the premium in cannibalized margin - half a billion dollars a quarter, openly disclosed - to keep a fixed budget of human minutes from migrating to a feed it didn't own. The minutes stayed. The rival kept winning the head-to-head anyway. Meta didn't out-locate the threat; it absorbed it, on purpose, into the one place it could still tax. That is what a successful defense actually looks like: not a flag planted on enemy ground, but a wall that cost a fortune to hold - and held.
Cannibalization Decision Tree
A decision tree for the moment the new thing threatens the cash cow: is the disruption real, will someone else do it if you don't, and can you afford to bleed your own margin to own the future? Blank to run on your own line; filled as the worked example tracing how the story's incumbent chose to cannibalize — or flinched and got cannibalized.
The worked example unlocks with a subscription. See plans →
Sources
Where this comes from — the filings, records, and reporting behind it.
- 1Instagram Reels was officially announced and launched on August 5, 2020, in over 50 countries including the US, following nine months of Brazil testing and a July 2020 soft-launch in India after India's TikTok ban.
- 2Facebook Reels launched in the US on September 29, 2021, as a separate rollout from Instagram Reels, with global expansion to 150 countries in February 2022.
- 3Meta's Q2 2022 earnings call transcript shows management explicitly acknowledged: 'Reels is additive to time spent. But obviously, it does have a cannibalistic impact as well… There is some engagement that we think is coming out of surfaces like Feed and Stories, and we're getting that engagement on Reels.'
- 4Meta's FY2022 10-K (primary SEC filing) shows total revenue of $116.61 billion, a 1% decrease vs. 2021, with ad price per unit down 16% YoY—the financial context in which Reels' monetization deficit was most acute.
- 5During the FTC v. Meta antitrust trial (April 2025), Zuckerberg testified under oath that Meta's addition of Reels was 'in large part a response to TikTok's rise' and that 'users continue to engage more on TikTok than with his apps.'
- 6Reels reached 20% of time spent on Instagram by Q1 2022 (per Zuckerberg on the Q1 2022 earnings call), and video overall made up 50% of time on Facebook; by January 2025, Zuckerberg told investors Reels accounts for more than 50% of Instagram time spent.
- 7Zuckerberg disclosed that Reels created approximately a $500 million revenue headwind per quarter (Q3 2022) because 'the monetization efficiency of Reels is much less than feed,' and Meta acknowledged this as a multi-year journey analogous to the Stories monetization ramp.
- 8At Reels' August 5, 2020 US launch, TikTok CEO Kevin Mayer publicly called it 'another copycat product… after their other copycat Lasso failed quickly,' and Zuckerberg himself testified to Congress that Facebook had 'certainly adapted features that others have led in.'
- 9Instagram Reels testing began in Brazil in November 2019, roughly nine months before the August 5, 2020 US launch.
- 10TikTok CEO Kevin Mayer wrote in a TikTok blog post on July 29, 2020: 'Facebook is even launching another copycat product, Reels (tied to Instagram), after their other copycat Lasso failed quickly.'
- 11Meta's Q3 2022 earnings call transcript shows Reels was 'negative overall to revenue by about $500 million in the quarter,' with management expecting it to become a tailwind within 12 to 18 months.