More than half of Ulta Beauty’s e-commerce orders are now fulfilled through its stores. At the same time, digital sales are growing in the high teens and inventory per store has fallen 4.1%.
Each number looks positive on its own. Together, they reveal a harder operational test: Ulta is asking less inventory in each location to serve more demand from beyond the store.
One Pool of Stock, Two Customers
Ulta’s network of more than 1,500 stores now handles the majority of its online orders. This gives the retailer inventory close to customers without requiring every digital purchase to travel through a central fulfilment facility.
But store stock is not spare stock.
The same product may be available to a customer browsing online and another standing in the aisle. Once it is allocated to one order, it is unavailable to the other. As stores take on more digital volume, accurate inventory records and careful allocation become more important than the headline fulfilment percentage.
Ulta ended the second quarter with $2.4 billion of inventory, broadly flat year on year despite opening stores and introducing new brands. Management attributed this to better inventory management, supported by an AI-powered sourcing capability that optimises stock across the network.
The challenge is no longer simply having enough inventory. It is placing each unit where it has the best chance of being sold profitably, while recognising that local stock may be serving demand well beyond the local catchment.
A Cheaper Parcel Can Create Costs Elsewhere
Ulta says greater use of stores helps manage the transportation-cost difference between digital and store sales. That makes sense: fulfilling closer to the customer can reduce distance and make better use of inventory already in the network.
Yet transport is only one part of the calculation.
A store order also requires someone to locate, pick and pack the product. It may compete with other tasks during busy periods. Routing from the wrong locations can produce split shipments or move scarce stock away from customers likely to buy it in-store.
None of this means store fulfilment is the wrong choice. It means the nearest available unit is not automatically the lowest-cost one.
The routing decision needs to consider delivery cost, store workload, local demand and the risk of losing another sale. Otherwise, savings in one part of the network can quietly reappear somewhere else. Cost has a habit of doing that.
The Percentage Is Not the Prize
It would be easy to treat a rising share of store-fulfilled orders as evidence that the model is progressing. But more volume through stores is not necessarily the goal.
The better measure is whether Ulta can serve growing digital demand at a lower total cost while maintaining availability and store productivity.
That becomes harder as inventory per location falls. Leaner stock leaves less room for inaccurate records, poor allocation or an unexpected increase in local demand. A system can route an order perfectly according to the data and still disappoint the customer if the product is not actually on the shelf.
Ulta says supply-chain productivity contributed to modest gross-margin improvement during the quarter, while channel mix remained a pressure. Its store network is clearly part of the response, but the long-term value of the model will depend on more than transportation savings.
The important questions are now:
- Can stores fulfil more digital orders without reducing availability for local customers?
- Does order routing account for store labour and potential lost sales, as well as delivery distance?
- How accurate must location-level inventory become as stock per store declines?
- At what point does additional digital volume require more capacity rather than better allocation?
Ulta’s stores give it a powerful fulfilment network that already exists, sits close to customers and carries a broad product range. But they are still stores, not small distribution centres with mirrors and lipstick.
The opportunity is not to push the maximum possible number of online orders through them. It is to make each unit of inventory work harder without asking the store operation to absorb costs the fulfilment model has chosen not to see.
