Four results in three days, and one shared pattern
Home Depot reported on 18 August 2026, Target and Lowe's on 19 August, and Walmart on 20 August. The U.S. Census Bureau published its second-quarter e-commerce figures on 18 August. Four large retailers and one official macro series landed inside three days, which makes them unusually useful to read together rather than one at a time.
The macro number is the one most likely to be quoted out of context. U.S. retail e-commerce sales rose 12.2% year over year in Q2 2026, against 6.7% growth in total retail sales, and e-commerce reached 17.1% of retail sales. The Census Bureau states plainly that these estimates are adjusted for seasonal variation but not for price changes. They measure nominal spending and channel share. They do not measure units, and they are not a demand index.
The company results carry the same warning in a sharper form. At all four retailers the digital channel grew several times faster than the underlying comparable business:
- Walmart U.S.: comparable sales +2.6%, e-commerce +24% — a gap of 21.4 percentage points.
- Target: comparable sales +3.8%, digital comparable sales +8.7% — a gap of 4.9 percentage points.
- Home Depot: comparable sales +1.7%, online comparable sales +11% — a gap of 9.3 percentage points, and a fifth consecutive quarter of double-digit online growth.
- Lowe's: comparable sales +0.2%, online sales +15.7% — a gap of 15.5 percentage points.
Those spreads are a calculation, not a disclosed metric, and the four companies do not define digital, comparable, transaction and traffic identically. The figures should not be added together or treated as interchangeable. But the direction is consistent enough to be worth a seller's attention: online outgrew the underlying business everywhere, and at two of the four the underlying business was serving fewer customers than a year ago.
The arithmetic that explains two of the four
Comparable sales growth can be approximated by multiplying the change in transactions by the change in average ticket: implied comp = (1 + transaction growth) × (1 + ticket growth) − 1. Applying that to the reported figures reproduces each company's headline almost exactly.
- Walmart U.S.: transactions +1.5%, ticket +1.1% → 2.62% implied against 2.6% reported.
- Target: traffic +3.6%, average transaction amount +0.2% → 3.81% implied against 3.8% reported.
- Home Depot: transactions −1.0%, ticket +2.8% → 1.77% implied against 1.7% reported.
- Lowe's: transactions −2.1%, ticket +2.3% → 0.15% implied against 0.2% reported.
The last two rows are the important ones. Home Depot and Lowe's both reported positive comparable sales while comparable transactions fell. Ticket rose by enough to cover the decline. That is a real result and both companies were open about it — Home Depot described customers engaging in smaller repair and maintenance projects while larger discretionary projects stayed under pressure, and Lowe's described discretionary DIY demand as still pressured, with its transaction decline centred on DIY.
Walmart's Sam's Club shows the same arithmetic running the other way: transactions +7.0% against average ticket −2.5%, producing roughly 4.3% against 4.4% reported. More customers, smaller baskets. Same growth rate, completely different business condition.
This is the trap in miniature, and it is the same one a store owner falls into with a single product.
A rising average order value can mean the offer got better, or it can mean the same number of people paid more for fewer, or that a bundle raised the basket while quietly costing more to pack and ship. The number on its own does not say which. We wrote the store-level version of this problem separately in when higher AOV hides weaker cart economics. The retailer results are a large, well-audited demonstration of exactly that failure mode: two of the four grew sales while serving fewer customers.
Why the online number is a channel metric first
The reasonable question is what produced 24% at Walmart, 15.7% at Lowe's and 11% at Home Depot while the underlying businesses grew 2.6%, 0.2% and 1.7%. Management gave partial answers. Walmart pointed to store-fulfilled delivery, advertising and marketplace. Target said its digital growth was led by same-day delivery growing more than 25%. Lowe's cited its digital experience, loyalty, free delivery, same-day fulfilment and its marketplace, and said it saw higher traffic and increased conversion online across both Pro and DIY. Home Depot said faster delivery speeds were resonating with customers and driving greater engagement.
Those are management interpretations. They identify the mechanisms each company is investing behind; they are not randomised estimates of cause. But they all describe the same kind of mechanism, and none of them describes a broad expansion in category demand.
A retailer's online sales can grow because it is capturing channel share rather than because more product is being consumed. The available routes include faster local fulfilment, a broader marketplace assortment, better search and app experience, more reliable in-stock rates, loyalty and membership, easier returns, retail-media traffic, store inventory becoming visible online, acquisitions and business mix, and price investment. Every one of those can lift a digital number without any underlying category getting bigger.
This matters because of how the number usually travels. The tempting chain of reasoning is: retailer online sales up, therefore category hot, therefore source a similar item, therefore scale advertising. That chain skips the only step that decides whether the money works — whether the growth came from traffic, units, price, mix, assortment, local delivery or channel share, and whether any of it can transfer to an independent store shipping a generic product from China.
A defensible version keeps that step. Establish what produced the growth. Identify a specific product attribute or customer mission that could transfer. Test whether your own conversion and unit economics confirm it. Scale only if the evidence survives your fulfilment costs.
What the retailers actually built
The strongest shared signal in the four packets is not that these companies sold online. It is how hard they are working to compress the time between the order and the doorstep.
- Target said its stores serve as fulfilment hubs for more than 95% of its sales, that same-day delivery grew more than 25%, and that it fulfilled nearly 30% more same-day and next-day units than the previous year.
- Home Depot said more than 65% of deliveries on in-stock parcel products are same day or next day, and that it launched Express Delivery nationwide with delivery on tens of thousands of products in three hours or less.
- Home Depot also said it reduced delivery lead times in the U.S. by approximately 45% over the previous 18 months, and that approximately 55% of big and bulky deliveries on stocked products now arrive within two days.
- Lowe's said customers responded to the free delivery and same-day delivery options it launched earlier in the year, and described same-day fulfilment as an increasingly important part of its omnichannel experience.
- Walmart said its e-commerce growth showed strength in store-fulfilled delivery, and that e-commerce contributed about 510 basis points to Walmart U.S. comparable sales growth.
The Home Depot figures are the most striking, because big and bulky is the category that was supposed to stay slow. A company that can move a stocked bulky item in two days, and tens of thousands of parcel products in three hours, has changed what a customer thinks a normal wait looks like.
That is the part that reaches a dropshipping store. Not the category. The reference point.
A customer is rarely comparing your product against nothing. They are comparing your product, price, proof, delivery and return terms against the closest credible local alternative. As that alternative gets faster, the difference becomes more visible, and a generic item needs a clearer reason to justify the wait. We have set out the full method for measuring and pricing that difference in the delivery gap — the local-substitute benchmark, the route measured as a distribution rather than an average, and the four conditions that should be true before stock moves anywhere.
You are not competing with a product margin
The third pattern explains why local price competition can look irrational from the outside while being perfectly rational for the retailer.
Walmart reported global advertising growth of 38%, Walmart U.S. advertising growth of 38%, Walmart Connect growth of 43% excluding Vizio, and global membership fee revenue growth of 17%. Target reported Roundel gross billings up nearly 20%, Target Plus marketplace GMV up more than 40% and Target Circle 360 membership revenue up more than 40%. Target also said it had lowered prices on more than 10,000 items over the previous year, and that 95% of its school supplies assortment was priced at or below the prior year.
Those figures are not directly comparable with each other. Advertising gross billings is not revenue. Marketplace GMV is not revenue. Membership fees carry different economics again. But they establish one structural fact: the customer relationship at these companies is monetised through more than the margin on the item in the box.
A large retailer can earn from product gross margin, supplier and seller advertising, marketplace commissions, membership fees, payment and financial services, fulfilment services, first-party data and retention, high-frequency repeat categories, and scale procurement. A single-product store running paid social usually depends on one product margin, one acquisition channel, one international parcel, one customer purchase and one refund decision.
That asymmetry is why comparing a factory unit price against a retailer's shelf price is not a competitive analysis. The retailer may be pricing the item as an entry point into a relationship that pays elsewhere. The seller has no elsewhere unless they build one.
The practical response is not to match the price. It is to know which of your own four margins is actually under pressure — product, acquisition, fulfilment-adjusted, or lifetime. A strong return on ad spend can still hide poor cash contribution once refunds, reships and support are counted. We covered the repeat-purchase half of that question in LTV is not permission to lose money on the first order.
The tariff refunds that flattered the headline
Three of the four results contained large IEEPA tariff refunds, and they are a clean example of a headline number moving for a reason that has nothing to do with customer demand.
- Target recognised $994 million pretax, which added 3.7 percentage points to both its gross margin and operating margin rates and $1.65 to earnings per share.
- Home Depot said it received $730 million, of which $685 million reduced cost of goods sold in the quarter.
- Lowe's reported approximately $80 million, or 30 basis points of gross margin, and said it was largely offset by elevated fuel and transportation costs in the quarter.
- Walmart said tariff refunds benefited its gross profit rate and that it intended to prioritise remaining refunds into price investments, but did not disclose the amount in its release.
Target's reported operating income rose from $1,317 million to $2,560 million, or 94.4%. Mechanically subtracting the disclosed $994 million refund leaves about $1,566 million — roughly 18.9% above the prior year, and about 5.9% of sales. That is still healthy growth. It is a very different statement from "operating income nearly doubled because consumers got much stronger." The subtraction is our arithmetic, not a measure Target designates or endorses.
Lowe's made the competitive consequence explicit: it said competitors used tariff refunds to lower prices later in the quarter, that it viewed this as transitory rather than a new normal, and that some of that pricing action was aimed at driving units and clearing seasonal inventory. A seller who reads a competitor's aggressive pricing as evidence of a strong category may be reading a one-off refund and an inventory problem instead.
Two signals worth watching, not acting on yet
Home Depot's smaller-project strength is a customer mission, not a product list. Thirteen of its sixteen merchandising departments posted positive comparable sales, Pro outperformed DIY, and the pressure sat in larger discretionary projects. The transferable idea is the mission — fix rather than replace, maintain an existing asset, complete a smaller affordable job, avoid a larger expense.
That mission does not transfer evenly to a parcel from China. A strong need can make a product less suitable for slow fulfilment, not more. An urgent replacement part, an exact-fit component, or anything needed to finish today's repair is a poor direct-parcel candidate precisely because the need is real. Planned, compact, low-compatibility-risk, easily demonstrated items travel better.
The AI-commerce statistics are the second watch item, and they are the ones most likely to be over-read. Lowe's said its MyLow assistant has supported more than 25 million questions since inception and that the conversion rate of customers who use MyLow is triple that of customers who do not. Target said its digital traffic sourced from external AI platforms is growing more than 3.5 times the industry compared with a year ago, that total wish-list creations were up more than 50% with items added more than doubling, and that conversion across its key back-to-school pages was up nearly 20%.
None of that establishes that installing an assistant will multiply conversion. Customers who open a product assistant may already carry high intent. More complex missions can produce both higher assistance use and larger baskets. Target changed assortment, prices, inventory placement, store execution, app content and seasonal merchandising in the same quarter. Neither company reported a randomised holdout, and neither disclosed the absolute base. These are observational figures from companies with an interest in the result — interesting enough to test, not strong enough to copy.
Three read-throughs to reject outright
First, tariff-refund-driven earnings growth is not consumer strength. The refunds changed reported margins and earnings. They say nothing about units bought.
Second, a named branded product growing is not a sourcing instruction. Target reported Lego sales up more than 30%, plush up more than 20% and its style-forward $10 headphones running more than 35% ahead of last year. The transferable observation is that accessible price points, novelty, visual identity, collectability and gifting worked. The non-transferable part is the brand, the licence and the exclusive collaboration. Sourcing a lookalike of a branded or licensed product is a legal and payment-processor problem, not a shortcut.
Third, retailers getting faster is not an instruction to put every product into a local warehouse. Their network scale does not remove your forecasting, cash and obsolescence risk. Local stock splits demand across countries, locks cash into unproven variants and strands slow-moving tail variants. It is a targeted tool, not a default.
What to measure in your own store this week
The useful response to all of this is not a product list. It is a short set of numbers that tell you whether your own growth is the healthy kind or the Home Depot kind.
- Sessions and qualified landing-page views, so you have a denominator that is not orders.
- Order conversion and unit conversion separately — paid orders per session, and units per session.
- Units per order, so you can see whether bundle growth is masking fewer customers.
- Net price per unit after discounts, and refunded units rather than only refunded orders.
- Contribution per order after fulfilment-failure costs, then contribution per session after acquisition cost.
- Repeat purchase by acquisition cohort, and every one of the above split by destination country and route.
Then apply one decision rule. Do not call a product "validated by retailer results" unless you have at least one of: category-level unit or order growth rather than sales growth; a specific transferable customer mission; a specific transferable product attribute; your own conversion and contribution evidence; or repeatable demand data from beyond one retailer's channel gains.
The bottom line from these four results is not that e-commerce is weak. Consumers are still moving online, and the strongest retailers are converting stores and inventory into digital fulfilment assets faster than most people expected. The bottom line is narrower and more useful: online growth is a channel metric before it is a product-research signal, ticket can rise while the customer base shrinks, and convenience has become part of the product. Do not only ask whether people buy the item online. Ask whether your offer survives the local alternative, the delivery gap and your true post-fulfilment economics.
Evidence boundary
Every company figure above comes from the official Q2 release or the official-hosted earnings-call transcript for Walmart (FY2027 Q2, reported 20 August 2026), Target, Home Depot and Lowe's, and the macro figures come from the U. S. Census Bureau's Q2 2026 quarterly retail e-commerce release. The digital-growth spreads, the ticket-and-transaction decomposition and the Target refund subtraction are our arithmetic from reported inputs, not disclosed measures.
Census retail figures are nominal and not adjusted for price changes. The four companies define digital, comparable, transaction and traffic differently, so the figures are not strictly comparable with each other.
Management explanations of why a metric moved are labelled as such and are not causal evidence; no company disclosed a randomised conversion lift from faster delivery, product-level unit demand for unbranded goods, or a delivery-time elasticity that transfers to an independent store. The seller actions are RyanFulfil's operating interpretation, not claims made by any of these companies and not guaranteed outcomes.