Importing from 1688 for the first time is not complicated. But the cost of first-time mistakes is very real. Most beginners do not fail because of missing capital or weak sales channels. They fail because they repeat a handful of avoidable errors that every experienced importer made early on. This is a roundup of the most expensive ones and how to sidestep each.
Underestimating landed cost and mistaking it for profit
The most common and most expensive mistake. A beginner sees a low price on 1688, estimates a selling price on the platform, subtracts the two, sees a margin, and orders. The problem is that the 1688 price is only one of many cost layers.
Real landed cost per unit includes: the purchase price on 1688, domestic China shipping from the factory to the consolidation warehouse, freight to Vietnam charged by actual weight or volumetric weight (whichever is higher), customs duties and import taxes, the order-agent service fee if you go through an intermediary (typically a few percent of order value), and shrinkage from defects or breakage. Convert everything to dong at the current rate (around VND 3,600 per yuan, check this when you calculate as rates shift). Then subtract the platform fee and expected ad spend per order. What remains is profit.
Experienced importers call that final figure the only number that matters. It usually runs 30 to 40 percent lower than the first estimate. If the math is wrong from the start, every sale is a potential hidden loss.
How to avoid it: build a complete landed-cost sheet before placing the first order. Include every line item. Add them up, then multiply by quantity to see total capital at risk.
Ordering on instinct instead of demand signals
Seeing a TikTok video with millions of views, or a shop moving fast, and copying the order. The reasoning sounds logical. It misses one critical fact: goods take roughly 18 to 30 days to arrive by sea freight, sometimes faster by road or air but at different cost. Within that window, a short trend can cool and a hot product can pull in dozens of shops driving each other down on price.
A reliable demand signal is not view count. It is actual units sold across several different shops, tracked over at least two weeks. If five to seven independent shops are all selling the same product steadily, demand lives in the product rather than in any single channel. If only one shop is winning, they win on their own audience, and you will not inherit that advantage.
Cross-check Shopee: if a product is hot on TikTok Shop but barely present on Shopee, it is most likely a short video-driven trend with high timing risk. If both platforms show steady sales history, you are looking at demand with a foundation, not a spike.
Skipping the saturation check
Beginners focus on "high demand" and forget to ask "is supply already saturated." High demand in a crowded market means the only way to sell is on price. Competing on price compresses margin. If margin was thin to begin with, a price cut means a loss.
A quick check: count the number of shops selling the same item on TikTok Shop and Shopee combined. Under ten shops, there is usually still room. A few dozen at near-identical prices signals an active price war. Then look at the price band. If the current selling price already sits close to landed cost plus platform fees, that market is broken and latecomers lose.
Check the supply side on 1688 too. If dozens of suppliers sell the exact same item at similar prices, the barrier to entry is near zero, and next month more shops will import it. A small differentiation (a bundled accessory, distinct packaging, a size or color variant not yet in the market) can keep you outside a naked price race.
Placing the first order too large
One of the most common capital-trap mistakes. The reasoning goes: larger quantities mean a lower unit cost, so a bigger batch saves money and avoids running out of stock. That logic holds once you know a product sells. For the first order, you do not know that yet.
The purpose of a first batch is not to optimize unit economics. It is to test whether this product sells on your channel, to your audience, at your target price. If the answer is yes, the next order can be larger. If the answer is no, the damage is contained.
Size a test order small enough that a failure does not hurt badly, but large enough to get a trustworthy sales signal. For most compact goods, 30 to 50 units is enough to read. For bulky or high-value items, even fewer works.
Not ordering a sample before a large run
Photos and descriptions on 1688 are not a reliable preview of what you actually receive, especially for goods made with cheap components or low-cost materials. Color differs from photos. Texture is rougher. Dimensions are off. Internal parts are a different grade. None of these are visible until you hold the sample.
Ordering a sample before committing to a large quantity is not a formality. It is the only way to know what you are getting. Sample prices often run three to four times the bulk unit cost, but that is insignificant compared to the loss when an entire shipment arrives out of spec.
When the sample arrives, inspect the specific points that generate complaints: seams, closures, electrical connections if any, smell, and actual measurements against the listing. Write down what you checked. That list becomes your reference when the real shipment lands.
Handing everything to an agent without understanding the process
Many first-time importers use an order agent because they do not read Chinese and are not yet familiar with the process. That is a practical choice. The problem arises when you hand over everything without understanding how it works, because then you cannot detect problems when they occur, and you have no visibility into what you are being charged.
Agents typically charge a service fee as a percentage of order value, plus separate charges for packaging, inspection, and storage. Ask upfront: what is the service fee percentage, which items are charged separately, and are there weight or size limits that apply. Agents are not the problem. Not knowing what you are paying for is.
For straightforward items from a known supplier, some beginners can order directly through 1688 cross-border with translation tools and then use an agent only for the freight and customs leg. This cuts cost but places the order-confirmation responsibility on you.
Leaving return costs out of the margin calculation
TikTok Shop and Shopee both allow buyers to return goods fairly easily. For imported 1688 goods, return rates depend on quality, listing accuracy, and product category. But even a low rate means real cost per returned unit that belongs in your average cost calculation.
When a unit comes back, you absorb at minimum: return shipping, re-inspection time, and often a condition downgrade that prevents selling at full price. If your return rate runs at 5 percent and each return costs an extra few thousand dong in handling, that total comes out of monthly profit.
A simple approach: measure your actual return rate from the first test batch, then load a return-cost allowance into the per-unit cost for every successful sale. It does not need to be precise, but it needs to be present.
Picking the cheapest supplier without checking their history
The same product on 1688 may have dozens of suppliers at different prices. Beginners default to the lowest. That is sometimes correct, but often not.
The cheapest listing might be a genuine factory with low overhead, or it might be a new seller with no track record, or it might be lower-quality goods that do not match the photos. To tell them apart, look at three data points: total transaction count, buyer ratings, and repeat-purchase rate. A supplier with several years of sales history, a large number of completed transactions, and consistent repeat buyers is more reliable than a new listing with a remarkably low price.
The few yuan difference per unit is not worth the loss when an entire shipment is unusable.
Bottom line
Most of these mistakes are not about lacking experience in any deep sense. They happen because of rushing, because a step looks optional until it causes a loss, or because the consequences have not appeared yet. Building a checklist for the first import covering full landed cost, demand and saturation checks, a sample order, and a small test batch is the most practical way to avoid most of the damage. Doing it carefully the first time is almost always faster in total.