On October 8, famed short seller Hindenburg Research released a short report on Roblox (RBLX) stock that contributed to a 9% intraday decline in shares that was quickly reversed. In fact, as of this writing, shares are marginally higher than their October 7 close.
Hindenburg disclosed that they had initiated a short position in the stock, with their core theses being that Roblox 1) inflates the number of users on the Roblox platform, and 2) is intended as a gaming environment for children but lacks adequate safeguards to prevent children from interacting with objectionable content.
To support their claims (Hindenburg’s report is available in its entirety at this link), Hindenburg cited research indicating that the platform’s Daily Active User (DAU) count may be inflated by as much as 25% to 42% due to the use of alt (alternate) accounts by individual users.
[text_ad]
The firm also cited evidence of inappropriate content for children, calling Roblox an “X-rated pedophile hellscape, exposing children to grooming, pornography, violent content and extremely abusive speech.” Further, Hindenburg cited The National Center on Sexual Exploitation, which labeled Roblox “a tool for sexual predators, a threat for children’s safety” earlier this year.
However, much of the criticism leveled against Roblox regarding child safety is also true of the internet writ large, and, aside from Roblox being considered a platform for children, the laundry list of complaints about child safety is little different from what one encounters on X (Twitter), Reddit (RDDT), Discord, Google, or any other large-scale internet platform that faces the question of content moderation.
The “right” solution for content moderation is well outside of the scope of this article, but as a parent, I will say this: communicate with your children about what is and is not appropriate online; do not rely on the platforms to protect them on your behalf.
What was true 25 years ago about being safe online remains true today, we’ve just normalized how much more time (and at what age) we are online.
I suspect that much of the stock’s immediate recovery from the Hindenburg report was the market seeing the criticisms and recognizing them as foundational problems of how we operate online.
Now, setting that element of the short report aside, there were two important takeaways that are more relevant to Roblox stock itself.
2 Elements Hindenburg Research Missed with Roblox (RBLX)
1) Bot Traffic
Because it was core to their thesis, the most important pushback concerns claims of inflated DAU figures. Hindenburg repeatedly states or implies that Roblox’s number of users and the amount of time spent in various Roblox games are skewed higher.
Hindenburg goes to great pains to establish the distinction between individual people and the number of accounts that Roblox reports as DAUs. This seems to be, on its face, an accurate and good-faith argument, but the use of alt accounts is widespread across many platforms, including X, Reddit, Instagram, Meta, etc.
What’s more shortsighted is the added emphasis on engagement time by “bots,” programs that give the appearance of an actual person, may perform limited actions in a game, and often send spam/scam messages targeting actual players.
Roblox reports 54.1 million DAU (per Hindenburg citing a 2022 quarterly report) spending 2.4 “engagement hours” per user (per Hindenburg citing 2023 statistics).
Due to the proliferation of alt accounts, Hindenburg estimates that a third (or more) of that user count may be inflated, and the engagement time may also be massively inflated, saying:
- This level of engagement is 58% higher than the average time spent by U.S. 8-12 year olds playing all mobile games (including tablet), according to a 2021 survey, and 26%-166% more time than leading social media platforms such as YouTube, TikTok and Instagram.
- To better understand the company’s reported engagement, we hired a technical consultant that monitored the top ~7,200 Roblox games across ~2.1 million Roblox servers, collecting 297.7 million rows of real-time player data.
- The data suggests that Roblox could be massively overstating the true level of user engagement across its platform: our review of an average of 30.4 million unique daily users found they spent just ~22 minutes in games per day.
- Further analysis of play times showed numerous games in which obvious bot accounts remained in-game for more than 24 hours straight. Our representative sample flagged millions of ‘zombie’ engagement hours that can majorly skew Roblox’s reported “average” engagement.
What Hindenburg fails to consider is that this is not uniquely a Roblox problem, but rather it is an internet problem.
Highlighting the prevalence of alt accounts and bot traffic simply fails to recognize how the modern internet operates.
The 2024 Imperva Bad Bot Report by Thales (which is referenced in this article from Security Brief U.K.) found that half of all internet traffic (49.6%) in 2023 was generated by bots.
Given that the Gaming segment had the largest share of “bad bot traffic” (at 57.2%), it would be notable if a significant portion of Roblox’s user engagement hours were not bot-driven.
However, beyond just being a casualty of internet trends, Roblox is also structured in such a way as to encourage artificially inflating engagement. Which brings us to our second point.
2) Roblox Relies on an Attention Economy
In 2023, Roblox generated $2.7 billion in revenues. Of that figure, $741 million (or 27%) was paid out to their “Creators,” the group of users who create games and experiences on the platform.
Like any other type of user-generated content that rewards contributors for the engagement (and spending) of platform users (such as YouTube, Instagram or X), creators themselves are incentivized to game the system in any manner that they can.
This creates an economic system in which attention is an asset.
Aside from paid services, the markers for monetizable internet content are a) how many people are consuming your content and b) how long people are spending doing so.
Add in the fact that content receiving attention already is most likely to garner more attention (and be pushed algorithmically to the top of “most popular” charts), and it should come as no surprise that Creators are incentivized to create the impression of attention (such as the aforementioned zombie engagement hours).
The problem with inflated engagement time is a consequence of how user-created content is monetized by creators, and Hindenburg’s reliance on those metrics is misplaced. What is positioned as a takedown of Roblox is more of an indictment of how the internet currently functions across most platforms.
That said, this is not a full-throated endorsement of Roblox as an investment (full disclosure, my kids each own one share, down 37% in the last three-plus years).
Roblox continues to lose money (more and more every year; $1.16 billion net loss in 2023) and has experienced regular insider selling and management turnover (which Hindenburg cites in more detail) while also trading at a sky-high price-to-sales ratio (8.6x, which is more than 50% higher than peers). In other words, Hindenburg is not wrong to question the viability of Roblox, but their mechanism for doing so misses the mark.
I wouldn’t consider Roblox a “screaming buy” at this stage by any means; however, despite all their due diligence, Hindenburg’s criticisms are an overstatement of the flaws in the underlying company and offer one takeaway for investors.
Short sellers’ (and Roblox Creators’, it seems) incentives will not always align with what you believe “should” incentivize them, and they’re often willing to game the elements that they can if it’s likely to generate more profit.
[author_ad]
