ByteDance – Fundamental break.

ByteDance’s business model has a fundamental problem.

• ByteDance had its best year ever last year in terms of revenue, but because it employs over a 100,000 people, it is still losing a fortune which is a situation that a company of this size should have long left behind it.
• ByteDance grew revenues 111% YoY in 2020 to $34bn with 1.9bn MaU.
• However, despite gross profits of $19bn (56%), the company lost $2bn at the operating level and a massive $45bn when non-cash items such as stock-based compensation and fair value movements of assets were taken into account.
• By contrast, Facebook has 2.85bn MaU, generated $86.0bn in revenues, $68.3bn in gross profit (81%), $32.7bn in operating profit (38%) with just 58,604 employees.
• Furthermore, I have long argued that Facebook’s employee base is bloated as a result of its algorithms not being good enough to detect objectionable content before it goes viral (see here).
• Hence, I think that as its AI improves, the number of employees should grow far more slowly (or even decline) as revenue expands.
• Even in its bloated state, Facebook’s revenue per employee in 2020 was $1.48m which is a number any company would be extremely happy with.
• By contrast, ByteDance only managed to generate $0.3m per employee in 2020 which comes as a big surprise.
• This is because RFM estimates that ByteDance is one of the leaders in AI placing comfortably in the top 5 globally.
• This means that it should not need lots of employees to scan the content that is uploaded to TikTok and Duoyin because its algorithms should be up to the job.
• Consequently, why ByteDance employs 110,000 is something of a mystery as it appears to have no need for such a large employee base.
• In December 2020, ByteDance raised $5bn at a valuation of $180bn but with financial performance like this, I struggle to see how this will be repeated in a rational public market.
• It also indicates that there is a fundamental problem with ByteDance’s business model because by the time Facebook was this size, it was already very profitable and cash generative.
• This problem needs to be fixed before this company will be able to generate any fundamental value for shareholders.
• The ByteDance bulls take the revenue number, apply Facebook’s revenue multiple and arrive at a number of $400bn for ByteDance.
• What they are forgetting is that Facebook trades at 7.8x 2021 EV / Revenues because it is very profitable meaning that because ByteDance still loses money, another measure should be used.
• Facebook trades 13.3x EV / gross profit which if applied to ByteDance gives a valuation of ByteDance of $253bn.
• Some would argue that a premium for higher growth should be applied to ByteDance but I would argue the opposite until the business model is able to produce the same level of profitability.
• This is because a business that grows 1000% per year but never makes money is still worth nothing to the holders of the shares.
• With 110,000 employees, it is going to take some time for ByteDance to become a profit and cash generation machine like Facebook.
• Hence, when the IPO comes to market and asks for $400bn from new shareholders, I am very unlikely to be enthusiastic.