From Steve’s blog, Aim Assist:
Lately, I’ve been thinking of a framework to use to better understand in-game monetization strategies and how they may ultimately apply to the Metaverse. Today, whether games are free-to-play or require a full game purchase, there are two different axis on which games can monetize recurrent consumer spending: identity and utility.
Identity is the degree to which purchases are cosmetic and aimed towards building someone’s online persona in a given world. When making purchases based on identity, players are rewarded with social credit in the game. These purchases offer no competitive advantage, but are important, nonetheless.
Utility is the degree to which the purchases have a competitive benefit in a given world. Utility purchases can give players a pure competitive advantage, but some of the more common utility purchases are those that save players time and advance them further in a game.
Three of the leading games building what appear to be early versions of the Metaverse are Fortnite, Roblox, and Grand Theft Auto, each finding a unique balance between identity and utility.
Fortnite’s monetization is based entirely on identity. The in-game purchases, whether skins, emotes, pickaxes, etc. offer no competitive advantage for players. Instead, anything purchased from the item shop is for social credit alone. Fortnite has done a great job of building scarcity into their item shop and typically offers items only for a limited time. In addition, they’ve been aggressive when it comes to IP partnerships, including items in their store from Marvel, D.C., Star Wars, Halo, and others. The dual nature of scarcity and popular culture crossovers has incentivized players to make purchases before they’re unable to. Fortnite also benefits as a third-person shooter, as users can see the skins they’re playing as, unlike many of the other popular first-person shooters it competes with.
Roblox strikes more of a balance monetizing through both identity and utility. Players can purchase outfits and animations, unique abilities, in-games, weapons, and more. Some purchases are purely cosmetic, while others offer a competitive advantage. In Roblox, monetization decisions are made by the individual developers themselves, rather than the company. Roblox sets a rate of USD to Robux, but developers are able to create their own offerings that may skew towards identity or utility in any given game.
Grand Theft Auto offers items that are primarily based on utility – apartments, vehicles, warehouses – all things that give players an advantage in the game. While you could argue there is an identity component to some of the items you can purchase, it’s typically not the primary reason why users spend real world dollars in the virtual world of GTA. GTA (and other Take-Two games NBA 2K and Red Dead Redemption) has built an incredible monetization engine. Players are able to purchase virtual currency to help them acquire in-game items necessary for advancement faster. It ultimately becomes a question of whether $20 is worth saving X hours of game time.
If the games are approaching monetization from different angles – what’s the best choice? We can look at average bookings per daily active user to give us a sense. Roblox reports the number quarterly and we’ve made estimates for Fortnite and Grand Theft Auto.
From Doug’s blog, The Deload:
The current inflationary environment has prompted tech to rightsize headcount with over 60,000 tech employees being laid off in the past year. While sad for the employees, it’s right for the long-term viability of companies and it’s responsible for investors.
Since June 30, Canoo laid off 58 employees, Niantic laid off 85 and Unity Games laid off 200. OpenSea laid off 20% of its employee base, Twitter laid off 30% of its recruiting team and Tesla just laid off another 229 workers from its AP team. Microsoft cut jobs to “realign business groups,” and Oracle is weighing a $1B cost-cutting strategy that would warrant thousands of layoffs, making it the company’s largest reorg in ten years.
Tech businesses are IP businesses
One of the biggest input costs is for labor to write code and build customer bases. In some ways, the overpayment of tech talent is like a more traditional, capital-intensive business building too many factories. That’s why we have capital market cycles. Under investment in capacity, build phase, over investment in capacity, reduction phase.
Tech shouldn’t be any different. It’s just that we’ve lived in the build/over investment phase for so long that it seems different.
The reality is this:
If we persistently invest in talent beyond reasonable incremental productivity for a business, then the business is never going to show meaningful tFCF to investors.
Extreme case in point: FTX is the third largest crypto trading platform in the world by volume, and they have less than 30 engineers. Coinbase is significantly smaller and probably has 30-40X more engineers, if not more.
At a Sohn interview with Patrick Collison, FTX founder Sam Bankman-Fried said that he believed most tech companies were structurally inefficient with hiring. He even thought that Facebook could be on the order of 5-300X smaller in headcount. That’s not a typo.
Whether his math is right or not, companies and investors all seem to be coming to the uncomfortable realization that tech is wildly overstaffed. While it’ll be messy and take time for companies to reduce headcount, such a trend is great for investors.
I hate to have to note this: I’m not celebrating job loss. I’m recognizing a structural reality that needs to change for healthy market function. The marginal software engineer or product manager is not worth $400-500K in total comp or more because he isn’t adding that amount of incremental value for investors.
Companies aren’t charities despite the emergence of stakeholder capitalism.
Time to raise prices
A vast swath of consumer tech was built on the idea of transforming old world services with the internet and using scale to deliver those services at lower costs to consumers. Many companies succeeded in modernizing necessary services on the internet, but often the lower prices were subsidized by easy money rather than benefits of structural scale.
Uber is a great service. I use it. It deserved to exist, but it’s been around for 13 years and never generated any true free cash flow. The dying traditional cab business has probably made more tFCF than Uber has over the past 13 years.
Imagine Uber were entirely private, bought out 100% by one individual. Now imagine there was no liquid market for the stock. Can’t sell it. How would that individual try to earn a return on his investment?
He’d probably lay off more than half of the company, raise prices to serve a smaller group of customers better, and try to pocket hundreds of millions in free cash flow per year into perpetuity. In the tight money, tFCF-focused world, excess staffing and compensation will be punished.
From Doug’s blog, The Deload:
Sanity is finally coming to private market valuations.
Klarna, a star private fintech company, announced new financing at a 85% discount to its last round in March 2021. Affirm is a close public comp. AFRM stock is down about 83% since last January.
Klarna’s release about the raise stated, “Klarna closes major financing round during worst stock downturn in 50 years.”
Sequoia partner Michael Moritz was quoted in that release: “The shift in Klarna’s valuation is entirely due to investors suddenly voting in the opposite manner to the way they voted for the past few years. The irony is that Klarna’s business, its position in various markets and its popularity with consumers and merchants are all stronger than at any time since Sequoia first invested in 2010. Eventually, after investors emerge from their bunkers, the stocks of Klarna and other first-rate companies will receive the attention they deserve.”
The shift in Klarna’s valuation is because the odds of good investment returns at a $46 billion valuation, nonetheless anything higher, were extremely unlikely. Investors aren’t voting in the opposite manner. They’re now voting in a responsible manner with rational assessment of odds vs an irresponsible manner over the past few years.
We looked at an opportunity to acquire some Klarna stock on secondary markets last year. We couldn’t get comfortable with the valuation.
Here’s the simple model we used at that time:
First, establish the target return, holding period, and what that implies for exit valuation.
- Valuation at time of investment of $45.6 billion.
- Four year holding period at a 20% target return. Investing from mid-2021 to mid-2024 liquidity event, which means the market would value the company on 2025 numbers.
- 3% annual dilution.
- The 20% target return plus 3% annual dilution implies around a $100 billion exit valuation.
Next, given the necessary exit valuation, what does that demand from business performance? Given the business performance demands in our set of assumptions, what were the odds that Klarna could yield a good outcome at $45.6 billion?
Didn’t seem good.
Accelerating revenue growth to over 50% off a $1 billion base, then sustaining it for years happens rarely in my studies of great growth businesses. As in almost never. The odds of also hitting a 25% net margin in four years after coming off a heavy cash burn period also seemed unlikely.
A defender of the prior valuation may challenge some of the assumptions up or down — multiples, margins, etc. Any such challenges don’t change the reality that the odds were uncomfortable at $45.6 billion that Klarna would yield a good outcome.
Simple models, like the one above, are powerful because they frame in broad strokes how the business needs to perform, and the investor can assess probabilities of that performance demand. It doesn’t matter if any of the assumptions are off a little.
When a simple model casts doubt on the potential for strong investment outcomes, a more complex model will often only bias the investor in the wrong direction. Harder work and more knowledge make us want to believe the work and knowledge are worthwhile.