Summary of Key Points
DouBao (an AI assistant under ByteDance) has recently started charging an “independent channel fee” for orders that are completed through its services: the comprehensive fee rate for hotel bookings is approximately 12% (11.4% for software service fees + 0.6% for payment processing fees), while it’s around 18% for some lifestyle service categories. This is different from the previous integration with DouYin’s natural traffic settlement system, as DouBao is now considered a separate transaction channel. The platform can identify which orders were generated by AI and has begun to charge separately for these types of orders. The main controversy arises because merchants feel that the fee rate is much higher than the previous 2.5%-8% charged by DouYin, leading to concerns about increased costs; users, on the other hand, are worried that AI recommendations might become biased due to the fee structure and thus less unbiased. DouBao emphasizes that this fee is not an advertising fee (users cannot buy higher rankings with money) but rather a service fee for transaction processing. The system is still in the testing phase, and whether it will be successful ultimately depends on whether merchants receive additional orders and whether users continue to trust the recommendations.
Explanation in Plain Language: What exactly is the 12% fee that DouBao charges?
Firstly, it’s important to clarify that this fee is paid by merchants to the platform, not by users. The composition of the 12% fee is as follows:
- 11.4% goes towards “software service fees” (reflecting the effort DouBao puts in to recommend and facilitate transactions for merchants).
- 0.6% covers the costs associated with using DouYin’s payment tools.
However, not all products are subject to this 12% fee rate. The fee varies depending on the type of merchant (e.g., hotels vs. restaurants), the category of product (e.g., accommodation vs. entrance tickets for attractions), and the transaction channel used (e.g., DouBao vs. direct DouYin searches). For example, the fee rate for lifestyle services is higher, at around 18%.
Why is there a separate charge for DouBao-generated orders?
Previously, orders brought by DouBao were combined with those from DouYin’s natural traffic (e.g., users searching directly for hotels), and the platform could not distinguish which were generated by AI. By separating these channels, the platform is conducting an experiment to understand several key aspects:
- How many additional orders can DouBao actually generate? Are these orders “new” (users would not have purchased them without the recommendation) or “taken from other channels” (e.g., users intended to buy through DouYin or Ctrip but ended up using DouBao)?
- Only by separately tracking and charging for AI-generated orders can the platform determine whether the investment in AI development is worthwhile. Specifically, it wants to see if the value of the orders obtained through DouBao justifies the fee.
Will merchants accept this new fee structure?
Merchants are primarily concerned with whether the cost is worth it. They will consider three key factors:
1. Whether the additional revenue from these orders covers the 12% fee: For instance, if a merchant sells a hotel for $1000 through DouBao and makes a profit of $200, paying $120 in fees leaves them with a net profit of $80; if the profit is only $100, paying $60 results in a net loss of $40.
2. **Whether these orders are “taken from their own traffic”: If the orders were intended to be purchased through other DouYin channels (e.g., direct hotel searches), then the additional fee means they are paying for orders that they would have obtained anyway, potentially resulting in a loss.
3. The quality of the orders: If the recommendations from DouBao are more accurate (e.g., if it recommends a hotel within the Beijing Third Ring Road with breakfast for less than $800 that meets the user’s criteria), merchants might be willing to pay more.
Should users worry?
Users may wonder if AI recommendations will favor merchants that pay more. DouBao claims that merchant payments do not affect recommendation rankings, but users are concerned that the system might prefer merchants that contribute more to the platform’s revenue. Several key issues need to be addressed:
- Which merchants can use DouBao’s services: Will merchants who don’t use it have no chance of being recommended?
- How recommendations are ranked: Are they based on user needs (e.g., price, reviews), or does the algorithm also take into account the fee amount paid by merchants?
- Transparency: Can users tell which recommendations are sponsored and which are natural results? (For example, ChatGPT in other countries separates sponsored content from regular search results, and Google labels sponsored deals as “Sponsored deal.”)
Domestic regulations require platforms to disclose their fee structures and increase the transparency of their algorithms. To build user trust, DouBao needs to show that its recommendations are based on genuine user needs, not just on merchant payments.
What is DouBao’s goal with this move?
DouBao is part of DouYin’s e-commerce and lifestyle service ecosystem (with existing products, payment, and after-sales support). Its goal is clear: to transform user inquiries into actual transactions. Previously, when users asked for recommendations, DouBao could only provide information; now it offers direct booking options, allowing users to place orders with one click. This move aims to increase platform revenue by reducing unnecessary redirects.
However, the success of this strategy depends on two critical factors:
- For merchants: Whether DouBao can continuously generate additional orders without competing with existing channels.
- For users: Whether the recommendations remain unbiased and truly help them find the best options.
In summary, DouBao aims to become an “AI shopping assistant” that generates revenue. To achieve this, it must prove its value by providing merchants with additional sales and ensuring that its recommendations are reliable and unbiased for users. Both aspects are essential for its success.