Polish shopping app whenUbuy reported on September 15, 2026 that observed Temu orders contained roughly half as many products in August as before the EU’s customs changes. Monthly spending among those users nevertheless remained broadly stable.

That is an interesting signal for sellers planning European inventory. It also needs careful qualification: the release describes activity captured by one shopping app, without enough published methodology to establish a national trend or prove that customs duties caused the change. This article is analysis of that evidence.

Fewer items does not establish a collapse in demand

In the whenUbuy release distributed through ArchNEWS, Temu’s average item count fell to 2.8 per order while average product value increased by about 50%.

The other platforms behaved differently. AliExpress’s item count stayed around 1.1, with monthly spending up 20%. Shein’s count remained around 1.7 and monthly spending stayed stable. These findings do not support a blanket claim that Polish consumers halved their purchases across Chinese marketplaces.

Average product value also is not a price index. A customer switching from inexpensive accessories to a larger household purchase can raise the average without any identical product becoming more expensive. Product mix, promotions and order frequency all need to be separated before drawing conclusions.

The customs change is real, but causation is unproven

The Council of the EU’s final announcement confirms an interim €3 customs duty from July 1, 2026 for qualifying small consignments valued below €150 and sent directly to EU consumers. The charge applies by distinct tariff category within a parcel, rather than simply adding €3 for every physical unit.

That distinction matters when comparing low-priced items and mixed baskets. Several units in the same tariff category do not necessarily create the same duty calculation as several different categories. Sellers need the actual classification and shipping arrangement to calculate exposure.

The timing makes the whenUbuy findings worth examining. Timing alone cannot show how much of the observed change came from duties, platform subsidies, seasonal demand or a different mix of app users. The public release provides no controlled comparison isolating those factors.

Stable spending can still change your inventory needs

Consider a deliberately simplified example. A customer spends €40 a month buying eight €5 products, then switches to four €10 products. Monthly revenue stays unchanged, while unit demand falls by half. This is an illustration, not a reconstruction of whenUbuy’s data.

For a seller, those scenarios produce different purchase orders, packaging requirements and fulfillment costs. The higher-value products might also require more working capital or generate more expensive returns. Revenue alone does not reveal the contribution left after those costs.

Track units per order, order value, purchase frequency, refund costs and contribution together. If you change the assortment, compare like-for-like products as well as the store total. Otherwise, a shift toward expensive SKUs can hide weakening demand for the inventory you already hold.

Separate local fulfillment from direct imports

Do not assume every order placed on a China-linked marketplace travels from China to the customer. EcomCrew’s reporting on AliExpress’s shift toward local warehouses and lower fulfillment costs provides relevant background on how the platform’s operating model is changing.

Local stock and direct imports need separate comparisons. They can differ in delivery time, checkout charges, inventory ownership and return arrangements. Goods imported in bulk still have import costs; moving inventory locally changes when those costs arise and how they are allocated.

For your own competitive review, use a consistent delivery address and an identical product or a clearly matched specification. Record the final checkout amount, discounts, delivery promise and shipping origin where disclosed. Repeat the observation instead of relying on one promotional screenshot.

That gives you evidence relevant to your category. A platform-wide average from Poland cannot establish the price pressure on one SKU in Germany, France or another market.

What would make the data more actionable

The public release does not disclose the sample size, exact comparison window, customer weighting or whether it follows the same users across periods. It also does not fully explain how product values, basket totals and monthly spending are defined and reconciled.

Until those details are available, avoid using the reported percentages to set a purchasing forecast. A shopping app’s users may differ from the broader population, and changes in its coverage can alter averages without equivalent changes in the market.

Use the findings to choose questions for your own order data: are buyers ordering fewer units, choosing different price points or changing how often they purchase? The next meaningful follow-up would be a documented comparison of the same customer group, with category and fulfillment breakdowns. That would help distinguish a lasting change in demand from a temporary change in the measured sample.