Summary of Key Points
This article outlines the evolution of the “taxation” models in the retail industry over the past three decades: from offline “space taxes” (paying to have a physical location to be seen by customers), to online “traffic taxes” (paying to gain platform exposure and attract attention), to the “intelligence taxes” of the AI era (paying for the right to be selected by AI algorithms). It focuses on three hidden forms of these intelligent taxes: data feedback taxes, cognitive filtering taxes, and transaction fulfillment pipeline taxes. The article warns that AI will further centralize information power, posing a risk of merchants’ “right to exist” being monopolized. However, it also suggests that the interaction among three parties (neutral AI, platforms, and merchants) could lead to a new ecological balance.
Detailed Analysis
1. From “Competing for Locations” to “Competing for Minds”: Three Phases of Retail Taxation
- Space Tax Era: In physical stores, prime locations (such as entrances to shopping malls) were extremely expensive, essentially buying the monopoly to ensure customers would notice your business. For example, a transfer fee of 200,000 yuan might not just buy 20 square meters of concrete space but the privilege of attracting passing customers’ attention.
- Traffic Tax Era: Online platforms replaced physical landlords, with exposure spots (such as recommendations on home pages) becoming the new form of “rent.” Merchants competed for customer attention by gaining a place on these platforms.
- Intelligence Tax Era: AI has begun to make decisions on behalf of users, shifting the focus from controlling access to influencing their choices. For instance, AI may suggest baby formula or products to mothers, meaning merchants must compete for the right to be selected by these algorithms, which is more concealed and harder to resist than in previous eras.
In simple terms: In the past, you had to find customers; now, AI finds you—unless you’re chosen by AI, you practically cease to exist as a business option.
2. The Invisible Tax: How Data Feedback Taxes Secretly Exert Control on Merchants
The first form of intelligent tax is the data feedback tax: Merchants are forced to share transaction and customer interaction data with platforms, which use this information to train AI algorithms that then compete against them.
- Example 1: When buying a battery-powered scooter, the store owner may encourage online payment to meet platform requirements, and the transaction data is used by the platform to optimize its supply chain or even develop its own brand of scooters.
- Example 2: Amazon once used sales and pricing data from third-party sellers to develop its own products, which were then prioritized for display on its platform, effectively making those sellers pay a form of “traffic tax” while also creating potential competitors for the platform.
- Characteristics: There are no contracts or invoices; data is deducted instantly after a transaction is completed. It’s like helping someone raise a child for free only to have that child later compete with your business.
3. The Narrow Gate: Cognitive Filtering and Structured Data Optimization
The second form of intelligent tax is the cognitive filtering tax: AI algorithms recommend only a limited range of products, leaving most merchants without the chance to be considered at all.
- Platform AI Filters: On platforms like Taobao, searching for “bluetooth headphones” might yield only 3 options, with additional options hidden behind a “more” button, leading to increased user churn as they navigate further down the search results.
- Third-Party AI Filters: Services like Kimi take away the decision-making power by recommending products; users remember the selected model and then purchase from the platform, effectively making the platform a mere intermediary.
- Merchants’ Response: To get through these filters, merchants are altering their product pages to be more readable by AI algorithms (e.g., using structured data) similar to how SEO optimized websites were designed for search engines in the past.
4. The Pipeline Trap: The Emergence of Transaction Fulfillment Fees
The third form of intelligent tax is the transaction fulfillment pipeline tax: AI has turned transactions, delivery, and after-sales services into closed systems, requiring merchants to pay fees to use these services.
- Example: ChatGPT offers a shopping feature, and if merchants use its platform for checkout, they must pay a 4% commission—this is not a referral fee but a “pipeline fee.” As AI adds one-click purchasing options, platforms like JD.com and Taobao become mere delivery channels, earning profits from logistics services.
- Platforms’ Countermeasures: Platforms like Taobao and Amazon are developing their own AI shopping assistants (e.g., Taobao’s Qianwen, Rufus) to retain control over the transaction fulfillment process and avoid becoming mere extensions of AI algorithms.
5. The Triangle of Interests: The Impossible Conflict Among Neutral AI, Platforms, and Merchants
In the AI era, the interests of these three parties create an “impossible triangle”:
- Neutral AI: Seeking to collect “cognitive taxes” (e.g., through commissions).
- Platforms: Desiring to maintain control over access to user data (e.g., traffic fees).
- Merchants: Looking to protect their profits and avoid additional taxes.
The article does not predict how this conflict will resolve, but it warns that if AI monopolizes the power to influence consumer choices, merchants’ survival space will shrink. However, the author also hopes he might be wrong, as this could lead to a healthier retail ecosystem by breaking algorithmic monopolies.
Conclusion
This article is not meant to doom the retail industry but to serve as a wake-up call for merchants. In the AI era, relying solely on traditional methods (such as spending money to gain visibility) is no longer enough. Merchants must learn to interact with AI algorithms effectively—either by making their products understandable to them or by finding new ways to survive in this new landscape. For consumers, it’s important to be aware of how AI decisions may limit their choices. After all, being selected by an AI algorithm does not necessarily mean the best option for them.