Chief Executive Officer / China
What we funded, changed and still could not prove
- Company revenue
- RMB486m
- 2025 reported sales
- Own retail network
- 42 shops
- Seven Chinese cities, June 2026
- Commercial experience
- 30 years
- Supplier coordination through CEO responsibility
Decisions & consequences
01Keeping twelve patterns useful one piece at a time
In 2022 our service team could often identify a broken plate but could not sell a replacement without opening a complete boxed set. Merchandising colleagues argued that customers bought a coordinated range, not a long-term spare-parts arrangement. The service team heard a different question: why should the loss of one item require another full purchase? I asked for a proposal limited to twelve established patterns rather than a promise covering everything we had ever sold.
The commercial director wanted a broad message about continuing availability. Finance questioned the cash tied up in slow-moving pieces and the cost of packing a single plate. I approved RMB2.8 million of initial stock and handling investment, with a five-year support window measured from each selected pattern’s final main-range sale. The offer listed supported items, alternatives and the possibility of an unavailable piece. It did not promise that every request would be fulfilled. We also separated the service review from ordinary range-sales reporting, so a small replacement order would not disappear inside a larger category result.
During the first twelve-month review, 4,800 requests resulted in 3,200 exact replacements, 1,040 accepted alternatives and 560 unfulfilled requests. Average direct fulfilment cost on completed requests was RMB58 against RMB42 of replacement-item revenue. Those figures excluded the initial investment and did not establish a customer-lifetime return. We continued the offer because the board accepted a defined service expense, while requiring closer attention to stock selection and unresolved requests.
I would involve the packing team earlier next time. The first proposal treated a single piece as a smaller version of a full-set order, overlooking handling steps that did not become cheaper with fewer items. The programme changed our range discussions by giving after-purchase demand an actual record. It also taught us to distinguish a limited, funded commitment from an attractive statement that colleagues could not reliably honour.
Letting a storage container look its actual size
In early 2024 we reviewed 440 fit-related returns among 8,000 orders from a kitchen-storage range. Customers often said an item was larger or smaller than expected even though a dimension appeared somewhere on its product page. Our studio had produced attractive photographs, but isolated objects and inconsistent reference points made their scale difficult to judge. The online team worried that a more literal presentation would make the range less appealing.
I authorised a ninety-six-line pilot costing RMB640,000 for new photography, revised measurement diagrams and page changes. The team placed critical dimensions close to the purchasing choice and used consistent everyday context in the photographs. We did not claim to have redesigned the products or corrected every possible misunderstanding. Merchandising colleagues retained editorial responsibility for the range, while service colleagues helped identify which questions customers repeatedly asked. Finance required the review to include conversion, returns and implementation cost rather than selecting a single favourable measure.
In the subsequent review, 252 of 8,400 orders were returned for recorded fit-related reasons. Conversion across the pilot pages moved from 2.9% to 2.6%. The two periods differed in promotions and traffic mix, so neither movement could be attributed entirely to the new presentation. We retained the clearer dimensions and continued testing the photographic treatment. The decision gave up some immediate confidence in the sales presentation in return for a more explicit account of what a customer was choosing.
I would design the comparison before commissioning the images if repeating the project. Our initial enthusiasm for fixing a visible problem made the evaluation less informative than it could have been. The enterprise change was a requirement to discuss consequential product facts during the creative brief, rather than asking service colleagues to explain them after purchase. A beautiful image still has a role, but it should not require the purchaser to supply the missing scale from imagination.
Recognising the shop colleague behind an online order
By 2025 several shop managers were reluctant to devote time to customers who intended to order online. Their sales measures gave no credit for that assistance, while they still carried the staffing cost. The digital director argued that adding shop rewards would burden a channel already paying for acquisition and fulfilment. Both accounts were reasonable within their own budgets; neither described the customer’s whole journey.
I asked ten shops to test a simple assistance reference that could be linked to an eventual online order. It was voluntary for the customer and did not require collecting a detailed account of their visit. We trained colleagues to offer help before discussing the reference, capped rewards and reviewed repeated or disputed claims. The finance lead kept transaction revenue in its original channel. Store-assistance credit was a separate operating measure and the reward an additional selling expense. We rejected a proposal to count the full order as store revenue as well, because that would make internal comparisons easier at the expense of a coherent company total.
Over six months, 610 online orders carried validated assistance references from the ten shops. Additional rewards cost RMB186,000. We could show that colleagues were being recognised for recorded help; we could not show that every one of those orders was incremental or would otherwise have been lost. The digital director remained concerned about paying for existing demand, and we retained that objection in the decision to continue a narrower version of the pilot.
I would give managers a clearer explanation of attribution before launch. Some interpreted the first briefing as a contest to recover sales from another channel. The useful change was to make a service contribution discussable without treating channel ownership as the customer’s problem. The next review will include observations of the actual assistance conversation, because a rising reference count alone could indicate better recording, more pressure or more useful help.