How a Multi-Platform Seller Built Profitability Reporting by Platform and SKU
The client was founded in 2019 and sells home and household goods across four channels: Taobao, Tmall, Douyin (TikTok China), and Pinduoduo. It runs more than 50 SKUs, books an annual GMV of about RMB 80 million, and employs a team of around 30 people. In the early years the founder managed the money himself: he checked sales in each platform dashboard, glanced at the bank balance, and felt the business was doing fine. As volume grew, the company first hired a traditional bookkeeping firm. The firm kept the books and filed taxes, and handed over one company-wide income statement each month. Later the company hired a full-time accountant, but a single accountant could not keep up with the data coming from four platforms, several stores, and dozens of SKUs. By 2023 the founder could clearly feel that sales were growing but profit was not. Combined monthly sales across the four platforms looked like RMB 6 to 7 million. Yet after platform fees, creator commissions, ad spend, refunds and returns, warehousing and logistics, and packaging costs, nobody could say how much each channel actually earned, which SKU was a real hit, or which product was quietly losing money. The bigger problem was settlement. Each platform settles on a different cycle, and the money moving through Alipay, WeChat Pay, the corporate bank account, and platform wallets did not reconcile. The finance records had drifted away from how the business actually ran.
Before working with Caigeek, the client's core problems came down to three. First, profit was only visible as a single total. The bookkeeping firm's income statement had three blocks: revenue, costs, and expenses. Nothing was split by platform, let alone by store or SKU. The founder could see whether the company as a whole made or lost money in a given month, but not whether Taobao was subsidizing Pinduoduo. Second, expenses could not be allocated consistently. Platform commissions varied widely by category and by campaign period. Creator commissions and fixed placement fees were settled after the fact and often crossed month-ends. Ad spend sat in separate dashboards: Qianchuan (Douyin's ad platform), Kuaishou's ad platform, and Alimama. Products with high refund rates were lumped together with low-refund products. None of these costs were collected by platform or by product, so the cost structure was a fog. Third, inventory and cash records did not connect. The warehouse kept a physical count, finance kept a fapiao ledger (fapiao is China's official tax invoice), and the platforms kept sales records. The three rarely matched. The founder could not see which goods were sitting in the warehouse as dead cost, or which products turned fast and tied up little cash. Purchasing decisions ran on experience and gut feel.
Caigeek did not treat this as simple bookkeeping. We set out to build an outsourced finance department for the client, and the project ran in four steps. Step one: business diagnosis and data inventory. A Caigeek consultant spent two weeks with the client's operations, warehouse, and finance staff, and with the founder. We mapped every data source: orders and statements from the four platforms, three Alipay and WeChat merchant accounts, the corporate bank account, the ERP inventory system, and logistics statements. For each source we listed the fields, update frequency, and gaps, and agreed on what could be captured automatically, what needed manual input, and which definitions had to be aligned first. Step two: a three-level ledger by platform, store, and SKU. Using the order ID as the thread, we booked platform sales, platform commissions, creator commissions, ad spend, refunds, logistics costs, and product costs line by line to the right platform, store, and SKU. Shared costs that could not be matched directly, such as warehouse rent, customer service salaries, and office expenses, were allocated by sales volume or order count under agreed rules. Each cost was assigned using documented allocation rules. Step three: automated reconciliation with RPA and SQL. Our technical team wrote RPA scripts that pull platform statements and payment flows every day. A SQL model then compares orders, cash, inventory, and expenses against each other and flags every difference in red. Work that used to take two accountants two weeks now finishes in two to three days, and the error rate dropped sharply. Step four: monthly analysis and early warnings. Before the 10th of each month, Caigeek delivers a platform-level P&L, a store-level P&L, a SKU margin table, an expense analysis, and an inventory turnover report, then sits down with the client to review the exceptions. In one month we found that a Pinduoduo product had a refund rate as high as 28%, which ate almost all of its gross margin. A new Douyin product, in contrast, had modest sales but strong repeat purchases and margins, and we advised putting more budget behind it.
After three months, the client's finance function moved from after-the-fact bookkeeping to a management-reporting process with exception alerts, in-period monitoring, and month-end review. First, the founder saw the profit contribution of each platform for the first time. Taobao and Tmall were growing more slowly, but higher order values and lower return rates made them a stable profit source. Pinduoduo sales were growing fast, yet platform fees, campaign subsidies, and refunds together consumed most of the margin. Douyin was still ramping: creator commissions and ad spend were high, but the brand exposure had clear value. Second, the SKU-level P&L exposed a group of high-volume but loss-making products. Two products sold several thousand orders a month and looked like winners. Once refunds, after-sales costs, ad spend, and inventory buildup were counted, they were actually losing money. The client adjusted the product mix and pricing in time, and concentrated resources on products with genuinely high margins and repeat purchases. Finally, inventory turnover improved markedly. Reconciling the warehouse ledger, the finance ledger, and platform sales records showed that a batch of slow-moving goods had sat in the warehouse for more than half a year, tying up a large amount of cash. After clearing that stock, inventory days fell from 78 to 52, freeing nearly RMB 1 million in working capital. The founder now opens Caigeek's online reports every week to review profit and inventory by platform, store, and SKU. Decisions on ad spend and purchasing no longer run on instinct. They run on data.
Results are based on client-provided and reconciled operating data for the stated engagement period. Individual outcomes vary by business model, data quality, implementation scope, and management actions.
“I used to think I knew whether the company was making money. In reality, I was guessing where the profit came from and where it leaked. After Caigeek split the books across our four platforms, I learned that Pinduoduo's big sales numbers come with very thin profit, while a few new products on Douyin carry strong margins. Now I open Caigeek's reports before every meeting.”
Client comment translated and lightly edited for clarity; company name withheld for confidentiality.
Facing similar problems in your business?
A Caigeek consultant can review your reporting needs and recommend a practical starting point.