Holding a dying SKU too long is one of the quietest ways to destroy capital, because there is no single loss line to stare at, just money sitting in a warehouse not returning anything. This is about the signals that tell you when a SKU has stopped deserving your capital, and how to exit cleanly enough to put that money somewhere better.
Why a bad SKU costs more than it looks
A slow-selling product does not stand still. It drains resources in several directions at once.
Capital gets locked. The money sitting in that stock cannot fund a new product line or top up a winning SKU. If you are running a 20-million-VND operation and 8 million is parked in a SKU selling a trickle of orders, that 8 million is earning close to nothing for you.
Storage cost accumulates. Whether you manage your own space or rent a 3PL, every extra day a unit sits there has a cost. With 1688 imports, by the time stock arrives it has already spent 20 to 30 days in transit. Each additional week of slow sales chips away at the margin you calculated when you ordered.
Listing rank falls. TikTok Shop and Shopee both favor products with consistent sales velocity and conversion. A low-volume SKU slides down search results over time, and once it slides, it sells even less. Breaking that loop almost always requires cutting the price, which kills the margin you were hoping to recover.
Attention gets wasted. The time you spend on purchase orders, inbound checking, and processing the occasional unit for a weak SKU is time you are not spending on the product that has momentum.
Five concrete signals to watch
No single number tells the whole story, but when two or three of these appear together, a decision is overdue.
Sell-through rate falling for four or more consecutive weeks. One bad week can be a seasonal dip or a competitor campaign. Four weeks of continuous decline is a trend. If you are running three SKUs and only one is showing this pattern, that is enough to start the review.
Real margin falling below your floor. Real margin means after platform fees, after ad spend, after order-handling cost. Not selling price minus 1688 price. A lot of imported goods look profitable on gross cost but break even or lose once every layer is counted. If the real margin has fallen below roughly 10 to 15 percent and there is no path to recover it without cutting purchase cost or raising the price, the SKU is eating into your capital rather than growing it.
The market has saturated at your price point. The average selling price for that item on the platform is drifting down and the gap between competitors is narrowing. You cannot price above the cluster without losing orders, and pricing at or below it leaves no room. This is the market telling you a price floor has been set, and anyone entering after this point cannot win.
Return rate running above normal for the category. Imported 1688 goods with a return rate consistently above 10 to 15 percent usually point to a quality issue you cannot fix at the source, or a listing that does not match the real product. Each return carries a reversal fee, a handling cost, and often a unit you cannot resell. The numbers add up fast.
You have adjusted repeatedly without improvement. You tried lowering the price, swapping the cover photo, rewriting the description, running a voucher, testing a micro-influencer, and nothing moved. That is the product or the market telling you the problem is not in the execution. Running more experiments when you have run out of new hypotheses is just spending more money on the same dead end.
Telling "still adjusting" from "actually dead"
Not every weak SKU deserves to be killed. Some are worth holding.
A SKU under eight weeks old may still be building rank. Platform algorithms need time. If weekly numbers are climbing slowly from a low base, that is a different situation from a product that peaked and is now falling.
A predictable seasonal dip is not saturation. If you sell something with a clear off-season and you know which two months are slow, that is not a death signal. The problem is that many sellers cannot distinguish the two because they never mapped their seasonality.
One untested variant remains. Sometimes the fix is a color swap or a size option, not an exit.
The clearest way to tell them apart: if the SKU had a peak and is now consistently declining below its own baseline, that is over. If it is oscillating around a stable floor, that is normal variance.
How to exit without a panic sale
Deciding to kill a SKU does not mean dumping everything at a loss the next day. There are more controlled ways out.
Calculate your break-even clearance price first. What price recovers your purchase cost, inbound freight, and platform fee without adding to the loss? That is your floor. Anything below it means you are paying to get rid of it.
Run a time-limited promotion, not a permanent price cut. A seven-to-ten day sale to push through existing stock usually works better than permanently lowering the listed price. Permanent price cuts anchor a new reference price and are very hard to reverse. Use the platform's own coupon mechanics to pass the discount to buyers without changing the sticker price.
Bundle with a stronger SKU. Pair the slow mover with something that converts well. The strong product pulls the conversion, the weak one clears the stock. Bundle pricing works on perceived value, not the sum of individual list prices.
Sell in bulk to another operator. If you have significant stock left, trader groups and wholesale buyer communities exist in most categories. Bulk prices run 20 to 30 percent below retail, but you free up the capital in days rather than weeks of drip orders.
Set a decision deadline. Instead of letting units sit indefinitely, set a checkpoint: if the clearance plan has not moved X units in three weeks, move to the next option. Without a deadline, the decision never gets made.
Where to put the capital after you exit
Exiting a SKU is step one. The more important step is making sure the freed capital goes somewhere worth it, or you are just rotating from one bad bet to another.
Before you liquidate, have at least one of these ready:
Top up a winning SKU. Which product in your current lineup has rising velocity, stable margin, and stock running thin faster than you want? This is the highest-confidence place to redeploy cash immediately.
A new SKU that has passed your research filter. If you already have a candidate waiting on capital, one you have already checked for demand, saturation, and real landed-cost margin, the freed money goes there directly. This is the ideal outcome: replacing a proven loser with a product that has a better case but has not been tested yet.
Hold cash if you have no clear answer. This sounds passive, but it is reasonable defense. Sitting on the freed capital for one to two weeks beats importing a new SKU in a rush because idle money feels wrong.
Bottom line
Exiting a SKU at the right time is capital discipline, not failure. Every product has a cycle, and recognizing where that cycle ends, so you can move capital to a better opportunity, is what separates operators who manage their catalog from operators whose catalog manages them. The signals to stop are usually visible well before the stock hits zero. Read them early, clear out cleanly, and let the capital work somewhere it can actually return.