W32 - New Thinking on Merchant Operations

Recently, PMs finished their mid-year research visits to merchants. After reviewing some visit notes and initial conversations with the PMs, I gained some new insights into the business.

Large merchants (B) and small merchants (B) are different and must be considered separately.

Large merchants need office efficiency, standardization, and enterprise process management.Here are some examples from the visits. Multi-location merchants prefer to withdraw pending settlement funds from Meituan and consolidate flows from different channels into a single account for unified management, so large merchants have a stronger demand for withdrawals. Large merchants manage in a more granular way; on the merchant app, operations are often handled by finance or operations roles. Finance focuses only on reconciliation, while operations focuses only on marketing and customer traffic. They are often not the restaurant owners, so interest in interest-bearing services like “Shengyibao” is limited.

Small merchants’ top need is customer acquisition, followed by wealth management.They also have funding and enterprise process needs, but not as strongly as large merchants. We see many small merchants buying small promotional packages, though operating capability varies. With fewer locations, small merchants pay more attention to interest-bearing settlement funds, but feedback shows most people confuse “Shengyidai” and “Shengyibao.” That confusion could be a valuable entry point.

On funding needs, both small and large merchants show demand to varying degrees; compared with large merchants, small merchants’ capital needs are looser. Due to operating cash flow needs and pandemic disruptions—especially for businesses with highly cyclical income—working capital is an absolute necessity. Shengyidai has huge potential; as the flagship fintech product, it should be positioned to capture profit, while other financial services should be positioned to operate and monetize traffic. It’s like paid search for Baidu or games for Tencent. But feedback exposes painful weaknesses in Shengyidai: inability to lend, lending amounts that are too small, and lack of dynamic credit based on merchants’ basic operations and scale. From the outside, its credit model still relies on pure personal information and feels “traditional,” which relates to our current business stage—this is only a snapshot. Below I share some views on how Shengyidai might evolve.

I recently read Professor Zeng Ming’s Intelligent Commerce. His methodology for assessing future commerce fits well with the Shengyidai business model. Ant Microloans is our benchmark product, and it can draw on extensive potential-customer data beyond basic merchant information, including many behavioral signals. All data are dynamic “live data.” For example: which items a Taobao seller is listing and how sales are performing; how diligently a seller runs a store (customer-service reply speed, daily hours of operation, etc.); past dishonest behavior; and so on. Lending, an ancient and complex activity, can be abstracted into a simple human–computer input. We have 7 million merchants and tens of millions of daily orders—massive supply-and-demand data creating huge data pressure. Yet in a product like Shengyidai this has not converted into major momentum; “intelligence” is still nascent.Integrating data, algorithms, and product should be a key iteration direction for Shengyidai.