Sixty days ago, a Shopee and TikTok Shop seller in Ho Chi Minh City was running 750 orders a month and still watching her bank balance flatline. That is the real shape of this case study: a 1688 import shop that lifted margin from 8% to 22% not by raising prices or spending more on ads, but by fixing how it calculated cost of goods.
What this shop looked like before it changed how it calculated cost
The shop sells kitchen organizers and home storage products, sourced from factories around Yiwu and Guangzhou, sold through Shopee and TikTok Shop Vietnam. Volume was steady at 700 to 800 orders a month, average order value around $9.20. On paper, margin sat at 8%, which is thin but not unusual for a first-year import operation still learning the ropes.
The problem was the gap between what the books said and what the owner felt. Revenue kept climbing month over month, but the cash sitting in the business account did not grow at anything close to the same rate. Ad spend was flat. Return rates were normal for the category. Nothing on the surface explained why an 8% margin business wasn't accumulating 8% worth of cash.
That mismatch is what pushed her to stop guessing and pull every single order from the last sourcing cycle to check the numbers line by line.
Where the mistake was: what the old cost calculation missed
The cost model she had been using was simple: take the listed price on 1688, convert it to VND at the day's exchange rate, call that the landed cost. It is the most common way new importers price product, and it is also the most common way they quietly lose money.
Four things were missing. Domestic freight inside China, from the factory to the consolidation warehouse in Guangzhou, was never added to unit cost. Repacking fees and warehouse storage charges at the consolidation point were treated as a rounding error, not a real line item. Damage and shrinkage, the small percentage of units that arrive cracked, mismatched, or simply missing, was never tracked against actual shipments. And platform costs, commission, payment processing, and the ad spend allocated per unit sold, were subtracted from revenue at the top level but never folded back into a per-SKU cost figure.
Each of these looks small in isolation. Stacked together across 750 orders a month, they were eating the entire margin.
The review process: reconciling every shipped order
The fix started with one sourcing batch: 40 SKUs, all shipped within the same three-week window. Instead of estimating, she pulled the actual supplier invoices, the actual freight receipts from the forwarder, and the actual fee statements from Shopee and TikTok Shop for that period.
Line by line, she matched what she thought each SKU cost against what it actually cost to land in her warehouse and get sold. The gap was not evenly distributed. Some SKUs were priced almost correctly. Others were off by 20% or more, usually the ones with awkward packaging that triggered extra handling fees, or ones sourced from a factory further from the main consolidation hub.
This kind of reconciliation is exactly what the detailed guide to calculating true landed cost from 1688 walks through step by step. She used that framework as the checklist, then applied it retroactively to shipments she thought she already understood.
Before and after numbers: the real comparison
Two SKUs made the pattern obvious. A stackable drawer organizer set was priced at $2.10 landed cost in the old model. The real number, once freight, repacking, storage, and a measured 4% shrinkage rate were added in, came to $2.85, a 36% understatement. A three-pack silicone kitchen mat set showed the same story: $3.40 assumed versus $4.65 actual.
Across the full 40-SKU batch, undercounted cost averaged 18 to 24% of true landed cost. That is not a rounding error, that is the entire profit margin, sitting uncounted on the wrong side of the ledger.
Overall margin moved from 8% to 22%, but not overnight. It took roughly three sourcing cycles, about ten weeks, because pricing corrections and supplier changes only show up in the numbers once new inventory replaces old inventory bought under the wrong assumptions.
What the shop did once it knew the real numbers
With real cost data in hand, three changes followed. First, supplier conversations changed tone. Instead of asking for a vague discount, she brought factory-level cost breakdowns to the table and negotiated based on actual volume and actual landed cost, a very different conversation than the one covered in most generic supplier negotiation advice.
Second, she restructured her shipping approach, consolidating more units per freight run and switching forwarders for the SKUs that were absorbing the highest per-unit domestic freight cost. Third, she killed or repriced the SKUs where true margin was under 5%, products that looked fine on the old spreadsheet and were actually barely breaking even.
Going forward, every new shipment gets costed using the same line-item process before it's ever listed for sale, not a rough conversion applied after the fact.
Lessons for other shop owners
The core lesson: get cost of goods right before you touch pricing, ad spend, or supplier switching. None of those levers matter if the input number feeding them is wrong.
The warning sign to watch for is the same one that triggered this review: revenue rising steadily while cash in the account does not rise at a matching pace. That gap is a signal, not noise.
Rebuild your cost calculation every time you bring on a new supplier or place a materially larger order, since freight terms, consolidation fees, and shrinkage rates shift with volume and route. And this shop is not an outlier. Undercounted cost is one of the most common margin leaks among 1688 importers, especially anyone still converting listed price straight to local currency and calling it done.
FAQ
How do I know if my cost of goods is wrong without a full audit? Compare revenue growth to cash growth over three months. If revenue is up 20% and cash is flat, cost is likely miscounted somewhere.
What's the single most commonly missed cost line? Domestic freight inside China, from factory to consolidation warehouse. It rarely appears on the supplier invoice, so it gets skipped entirely. See the import fee breakdown for first-time buyers for the full list.
Does fixing cost of goods mean raising prices? Sometimes, but often it means cutting SKUs that were never profitable and reallocating that shelf space to ones that actually work.
How often should I redo this calculation? At minimum, every time you switch suppliers or place a shipment large enough to change your freight rate tier.
Does slow-moving inventory make this worse? Yes. Capital sitting in unsold stock compounds a bad cost calculation, since you're carrying storage cost on top of the original error. Worth checking against the inventory turnover guide for 1688 imports and reviewing your shipping cost options at the same time.
If you're running the same math on spreadsheets across multiple suppliers and shipments, this is exactly the gap Ordinex is built to close. Scout and Orders are both in private beta now, built to keep landed cost accurate per SKU without the manual reconciliation. Get on the list at ordinex.cc.