When I first looked at Roblox (NYSE:RBLX) in March, I placed a “Sell” rating on the stock, saying that while I thought it had an attractive fly-wheel business model, I thought the stock valuation was well ahead of itself. With the stock down about -30%, I upgraded the stock to “Hold” in October, noting that I liked the introduction of new advertising that the company was introducing, but I thought stock-comp was too high and I wanted to see more progress with free cash flow. With the stock up about 40% since my upgrade and the company holding an Analyst Day and reporting Q3 results since I last looked at the stock, I wanted to catch up on the name.
As a refresher, RBLX operates a digital community and online gaming. Platform that consists of three components: Roblox Client, the Roblox Studio, and the Roblox Cloud
Roblox Client is its front-end user experience where users explore 3-D digital worlds, interact with other users, and play games made by the community’s developers. Users pay for activities and virtual goods using the Robux virtual currency that they purchase.
Developers use the company’s free toolset Roblox Studio to create and operate games and other virtual experiences within the Roblox universe. Developers earn money through the sale of virtual items, while also getting a cut of premium subscriptions based on how much time is spent on their experiences by these types of subscribers. RBLX generally receives 30% of each Robux spent on its platform, while developers get 30% of the transaction and distributors 40%. Developers and distributors are often the same entity. The sales of Robux are recorded as deferred revenue and then as revenue after consumable items are used. Durable virtual items are recognized notably over the expected lifetime of a user.
The third pillar of the RBLX community is the infrastructure and services used to power the platform, which is called Roblox Client.
Progress In Some Areas
In my initial write-up, I said that RBLX’s biggest opportunity was to continue to expand its community, especially with older users. One knock on the company had always been that a large percentage of its users were under 13 years old. This led to negative reports on the safety of the platform and had led the company to pour money into trust and safety initiatives.
Between these initiatives and large investments into its infrastructure, the company has continued to spend a lot on capex, as well as ongoing costs. This in turn has led to a lack of free cash flow, something that concerned me when I last looked at the name.
Through the first nine months of the year, RBLX spent $255.5 million on capex while generating $314.8 million in operating cash flow. The company saw a bit of a shift in Q3, with capex spend of only $53.2 million against $112.7 million. So 47% of its OCF went to CapEx in Q3. That compares to the first half of the year when essentially 100% of its $202.2 million went into capex.
It appears much of its infrastructure spending on adding redundancy through core and edge data centers is largely behind it for now. This should lead to solid free cash flow in the new year.
Leveraging ongoing infrastructure and safety costs going forward will also be important. The company did see a drop in these costs in Q3, with them falling -1% year over year and -6% sequentially.
Right now, those costs are about 15% of bookings and the company hopes to reduce that to about 10% over the next 4-5 years. One way it can do this, especially on the safety front, is through leveraging AI, which it is just beginning to do.
Discussing safety at its Investor Day, Chief Product Officer Manual Bronstein said:
“As you know, safety and stability is core and paramount to what we do on Roblox. It’s part of our vision. So we continue to innovate in categories that make us more accurate and better at safety and stability to promote that on the platform. And while doing that, we’re also becoming more efficient at scale. So this year, we continue to improve our automated content moderation. This makes it easy through machine learning models to detect if there’s an appropriate content on the platform and automate those flows. We launched in Q4 something that we’re calling descriptive safety reporting. So if you’re in an experience and you’re a user and you experience something bad, you can actually capture a picture of what happened, share that with some annotation. What that makes is the process a lot more efficient because if you are a person on the moderation side and you receive that information, it’s a lot easier for you to take action. So it makes our moderators more efficient, and it makes reporting more accurate. And last but not least, we launched something in Q4 called the stability prompt for voice. What this does is using our machine learning models and artificial intelligence to detect if I’m having a voice conversation real time, if I say something that is violates our community guidelines. And what we do is we prompt the user with some messaging so that they know that they’re doing something wrong. This is to inspire stability and actually make their behavior better.”
This appears to be a good first start in leveraging AI, and as the process improves, it should need less headcount dedicated to these endeavors, saving costs.
Stock-based comp was another area of concern for me last time I looked at RBLX, and that has not improved. It was a whopping $220.0 million in Q3, up 36% year over year. That’s against only $81.1 million in adjusted EBITDA in the quarter, so stock comp was over 2.5x the EBITDA it generated. While non-cash, stock comp is still a real expense. As a result, RBLX’s share count has risen by 44.9 million shares, or 7.8%, since the end of 2021.
Now RBLX posted pretty good underlying numbers when it reported its Q3 results in November. Revenue rose 38% to $713.2 million, while bookings climbed 20% to $839.5 million. Active daily users were up 20% to 70.2 million, with monthly active unique payers growing 14% and average bookings per monthly unique payer up 5% to $19.02. Importantly, users 13 years and older grew 25% to 40.0 million, as the company tries to shift its demographics to older users.
Advertising was an opportunity I was excited about for RBLX, and on that front the company has been making progress and it will roll out several new ad types in 2024, including video ads. It also looks like it will test real-world commerce in its virtual world this year and likely roll it out fully sometime in 2025.
At its Investor Day, Chief Product Officer Christina Wooten said:
“We want to take you now through surfaces that are outside of our experiences that get hundreds of millions of impressions per day. So typically, people come to the homepage or Discover page on our platform and we’re going to allow brands to actually bid on these experiences, these premium placements through the ad manager. And we’re going to make this even better. So we’re making it more prominent placement on these pages, and we’re seeing exciting results so far. So through test 2x conversions from this premium placement positioning. We are also doing the work to make targeting better. So today, brands can target through gender, geo, age and device. But we know that advertisers want more precision over who sees their ads to drive ad effectiveness. And that’s exactly why we’re enabling genre-based targeting for all of our ad units going forward, starting in first half 2024. So brands want to target their ads based on genre. This allows them to reach specific genres like action, role play, sports and fashion. Okay. So let’s move into the second pillar of our go-forward strategy, which are video ads. Now brands have been asking us for video ads for many years, and we’re excited that they’re coming in 2024. These are going to make it easier for brands and advertisers to come on to Roblox and for us to scale our advertising business. Portals and sponsor experiences are for brands who have an existing experience already on Roblox. Video ads can be for anyone with or without an experience.”
Ads and e-commerce look like two attractive potential growth drivers for RBLX in the coming years. This is a big opportunity for the firm, although execution will be key. However, given the company’s younger demographic, we’ll also have to see how this plays out.
Delivering ads to under 13-year-olds and having e-commerce available could get a bit tricky. Users under the age of 13 still represent 42% of its DAUs in Q3. North America users, meanwhile, were 22% of its base. Typically, North American ARPUs are much higher when it comes to social media advertising-based businesses. If the percentage of users over 13 years old is similar across geographies, the core ad market targeting NA users over 13 years old probably would be about 9 million, or about 13%, of RBLX’s DAUs. So the opportunity may not be quite as big as hoped.
Revenue estimates being reported on most sites (including Seeking Alpha and FinBox) for RBLX are actually booking estimates. Actual revenue will likely be about 75-77% of booking.
Using the midpoint of that formula, RBLX currently trades around an EV/S multiple of 8.3x based on the 2024 revenue consensus of $3.0 billion and 5,9x the 2025 consensus of $4.2 billion. Billings growth is expected mid-teens in the next two years.
RBLX’s valuation is higher than its entertainment peers on an EV/S basis and considerably higher on an EV/EBITDA basis. While RBLX could be valued on a revenue basis, when taking into account its developer exchange fees and Infrastructure & Safety fees, its margins really don’t justify it being valued by that method. At the same time, given its huge stock comp expenses, valuing it an EV/EBITDA method also isn’t ideal.
As such, I’ll try to value the company based on a DAU basis versus Netflix (NFLX) subs. NFLX is valued at about $880 a sub, with an ARPU that is 3.4x higher than RBLX. However, its all-in gross margin is nearly double (42% vs 22%) that of RBLX, bringing its gross profit per sub 6.4x higher. That would value RBLX on an equivalent DAU basis at about $13.
While I like some of the initiatives at RBLX, such as its advertising opportunity, and the company should see a much-improved cash flow situation next year, it continues to have issues I don’t like. Its stock comp remains out of hand and continues to dilute shareholders. Meanwhile, its ongoing infrastructure and safety costs remain higher. I think it can start to use AI to leverage these costs, but right now they are really a drag on the business.
At the same time, valuation remains a big issue. Given its stock comp and low all-in gross margins, there isn’t a great way to value the stock. A DAU basis isn’t perfect, but it is a decent way to value it. On that front, even if the stock were to trade at a premium given its growth and operating leverage potential, it still appears to be very overvalued. As such, I’m going to downgrade the stock to “Sell” following its recent run-up.