The great destocking is about to meet its first peak season

In January of last year, the typical ecommerce operator on our network was carrying 111 days of inventory. Eighteen months later, at the close of this year’s second quarter, that figure had fallen to 89 days, among the leanest readings of the entire period and 5.9 days below the same week last year.

The eighteen months in between were anything but quiet. Tariffs reshaped import costs across the categories most operators sell, with Federal Reserve researchers finding retail prices on Chinese-origin goods up 8.5% year over year by December and pass-through to consumers running at least 30%. Diesel rose roughly 50% year over year. Consumer sentiment fell to among its weakest readings on record.

Any one of those developments could have justified a defensive inventory build. None of them interrupted the drawdown, and that is what makes the trend worth examining.

The starting point may itself be a tariff artifact. Federal Reserve researchers note that many retailers built inventory ahead of the tariffs taking effect, which suggests the 111-day peak reflected a defensive stockpile as much as a normal operating posture. What followed was not only a drawdown but an unwinding. 

What the data shows, across the more than 4,900 supply chain operators moving 485 million orders a year through the Deposco platform, is a decline that has been steady, broad-based and by every appearance deliberate. Thousands of planners made the same call, quarter after quarter: run thinner and turn faster. Elevated borrowing costs supplied a standing motive, since every day of inventory carries a financing cost, and for a year and a half the network has consistently preferred to hold the cash instead. Annualized inventory turns now sit at 4.09x as a result. By almost any measure it has been a disciplined retreat.

The question this fall will answer is whether it went one season too far.

While inventories thinned, demand moved in the opposite direction. Unit order volume across the network grew 4.0% year over year in April and was growing 8.8% by the end of June, a steady acceleration that comes against negative order growth at the same point last year. Every week, more units are moving through less inventory coverage.

The consumer behind that demand is not making the planning any easier. The University of Michigan’s consumer sentiment index fell 16% in two months, landing at 44.8 in May, among its weakest readings on record. Advance retail sales, as reported by the Federal Reserve Bank of St. Louis, rose anyway, from $653.8 billion in March to $662.8 billion in May. Shoppers are expressing economic pessimism while continuing to spend. A demand planner can model pessimism and can model spending, but modeling both at the same time is a harder exercise, and that is the exercise this summer requires.

What makes the situation genuinely new is how the cost of guessing wrong has shifted. Every buy plan is a decision between two mistakes. Overbuying means carrying charges and January markdowns, which is painful but budgetable. Underbuying means unfulfilled demand, loss of customer loyalty and expedited freight purchased at whatever rates the fourth quarter is charging. This year those rates start from parcel costs already up 12.8% year over year, before any seasonal surcharge schedules take effect. Eighteen months of destocking has moved the network’s risk from the cheaper mistake toward the more expensive one, and not many peak plans have been rewritten to account for it.

None of this argues for a return to 111 days. Holding that much coverage was its own kind of expensive, and the discipline that unwound it is real. But the operators positioned for a good peak season are not the ones holding the most inventory. They are the ones who know precisely where their coverage is thin. In practice that means testing the peak plan against unit demand rather than last year’s dollars, since dollars have decelerated and units have not. It means setting replenishment trigger points before supplier lead times exceed the runway to peak, after which reordering becomes a spot-market exercise. And it means remembering that lean does not have to be uniform. Ensure you have full visibility into where your goods sit across your network, and concentrate coverage behind the products that will actually carry the season rather than spreading a shortage evenly across the catalog.

The destocking era has been a long exercise in proving that discipline beats cushion. Peak season is where that proposition meets its first real test, with less buffer and more demand than the network has seen since the drawdown began. Roughly ninety days from now, the operators who spent eighteen months choosing to run thinner will find out how that choice performs under load.

This analysis draws on live platform data from Deposco, a cloud-native supply chain platform processing $84 billion in annual GMV across 4,900+ brands, retailers, and 3PLs. The full Q2 2026 findings, including four on-record directional calls for Q3, are published in Commerce Signal, Deposco’s quarterly intelligence brief on US ecommerce fulfillment. Subscribe here to get each issue as it publishes.