From Platform ROI to Product-Level Profitability for a Douyin Advertiser
This advertising-led women's apparel seller launched on Douyin in 2022, using Qianchuan (Douyin's ad platform) for paid customer acquisition, and later expanded to Xiaohongshu (RED) Spotlight ads and Tencent Ads. The company runs a small-batch, fast-response supply chain: it releases more than a dozen new styles every week, uses ad spend to test demand, and only then decides whether to reorder. Monthly GMV once passed RMB 30 million and the team grew past 50 people, but the owner grew more anxious as sales rose. The bank balance was not growing in proportion, and the profit figure finance produced at each month-end never matched the return-on-ad-spend figures the operations team saw in the ad dashboards. The company had tried several ways to manage its finances. A traditional bookkeeping firm could only book fapiao and bank flows and knew nothing about Qianchuan ad spend. An in-house e-commerce accountant could record expenses that had already happened but could not tie ad spend to specific products or orders. The operations team watched ROI in the Qianchuan dashboard every day, but that ROI only divides sales by ad spend. It does not deduct product costs, refunds, platform commissions, or logistics, and its conclusions often ran opposite to real profitability.
After Caigeek came in, we summarized the client's problems as five blind spots. First, no sight of real ROI. The ROI shown in the Qianchuan dashboard is usually sales divided by ad spend, and it does not deduct refunds. Return rates in women's apparel run as high as 35% to 45%. Measured on pre-refund sales, many campaigns look profitable; after refunds, they can be loss-making. Second, no sight of product-level profit. The client launches many new styles every week, and each style has its own cost, price, refund rate, and return on ad spend. Finance could only produce a company-wide total, not tell the owner which style makes money, which one loses, and which one only looks like a bestseller. Third, no sight of channel comparisons. Douyin, Xiaohongshu, and Tencent Ads differ widely in traffic cost, audience quality, and refund rates. Without one set of books covering revenue, cost, and spend for every channel, there is no way to judge which channel deserves more budget and which should shrink. Fourth, no sight of the link between inventory and cash. An ad-driven seller must book fabric, place orders, and pay for ads ahead of sales, which ties up a great deal of cash. Watching sales and profit without watching inventory turnover and cash flow invites a dangerous situation: sales explode while cash runs out. Fifth, no sight of how campaign rhythm shapes profit. Paid traffic moves through stages: cold start, scaling, and harvesting. Each stage has a different ROI and profit structure. A finance function that only summarizes by month cannot support the daily and weekly adjustments operations needs to make.
Caigeek built the client a fine-grained finance system organized around the product, wired along the channels, and run at a daily rhythm. Step one: connect the four ends — ad spend, orders, inventory, and cash. We pulled spend data from the Qianchuan, Xiaohongshu Spotlight, and Tencent Ads dashboards; order and refund data from the Douyin, Xiaohongshu, and WeChat Channels stores; inventory and purchasing data from the ERP; and cash flows from Alipay, WeChat Pay, and the corporate bank account into one reconciliation platform. Order IDs and product SKUs tie everything together. Step two: a contribution-profit-to-ad-spend model at product level. The metric starts with post-refund sales revenue, deducts product cost, logistics, platform commissions, after-sales costs, and advertising spend, and then compares the resulting contribution profit with advertising spend. For the first time, the client could evaluate the profitability of each product, campaign, and creative on a consistent basis. Step three: reports by channel, by day, and by week. Paid traffic moves fast, so we designed a daily and a weekly report. The daily report tracks spend, post-refund sales, refund rate, contribution profit after advertising, and inventory warnings. The weekly report summarizes contribution profit by channel and product line, so the operations team can review the prior week's campaigns and plan the next. Step four: linked inventory and cash alerts. Caigeek built one table per product covering sales, stock, in-transit quantities, purchasing lead time, and cash tied up. When a product's sales spike and stock runs short, the system warns early to reorder. When a product sits unsold and ties up cash, it flags the need to clear it. This reduces stockouts on high-performing products and excess inventory on underperforming products. Step five: participating in weekly performance reviews. Caigeek consultants join the client's product-selection and campaign-review meetings and use finance data to support decisions about product budgets, bids, and channel allocation. Finance became part of the decision process.
Two months later, the quality of the company's operations had visibly improved. First, the contribution analysis identified several high-volume but low-margin products. Some showed strong Qianchuan return-on-ad-spend figures, but their contribution profit was close to break-even after refunds, product costs, logistics, platform commissions, and advertising spend. The client reduced budgets on those products and reallocated spend to styles with stronger contribution margins. Second, the channel mix improved. With one reporting basis, the team found that Xiaohongshu's traffic cost more than Douyin's, but its stronger audience quality and lower refund rates produced a better contribution margin after advertising. One product line on Tencent Ads was near break-even because of an excessive refund rate. The client reallocated channel budgets, and overall ad spend efficiency rose by about 20%. Third, inventory and cash turned healthier. With restock warnings and slow-mover alerts, the client cut the stockout rate on hit products from 15% to under 5%, cleared backed-up styles, and freed more than RMB 800,000 in working capital. Fabric purchasing cycles became easier to control. Finally, the company's overall gross margin rose by about 6 percentage points. Cutting inefficient campaigns, improving the product mix, reducing refund losses, and turning inventory faster — together these moves improved profit meaningfully while sales stayed stable. The owner now starts every morning with Caigeek's daily operations report: contribution profit, channel performance, and inventory warnings at a glance. As he puts it, paid customer acquisition used to feel like driving blindfolded. Now he has a clearer map.
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.
“We used to rely mainly on the Qianchuan dashboard. Caigeek's contribution analysis showed that some high-volume styles added very little profit after refunds and costs. We now review the daily finance report before adjusting advertising budgets.”
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.