Grocery delivery has spent more than a decade trying to prove it can be a real business and not a perpetual marketing experiment. The companies racing to drop bananas at your door have raised billions, burned billions more, and reshaped how Americans buy food. Yet beneath the slick apps and 30-minute promises lies a brutal economic question: who actually keeps any of the money? The answer, after years of data from US grocers, third-party platforms, and quiet quarterly disclosures, is more nuanced than either the bull case or the doom-loop narrative suggests.
In short
- Margins are razor thin. A typical US grocery basket carries a 1 to 3 percent net margin in stores; delivery adds 8 to 15 percent of order value in extra cost.
- Platforms (Instacart, DoorDash, Uber) extract roughly 15 to 30 percent of the order in fees, often more once advertising and surge pricing kick in.
- Retailers profit on data, ads and private label long before they profit on the delivery itself.
- Shoppers (gig workers) earn $14 to $22 per hour gross, but net pay after vehicle, fuel and tax can fall to $7 to $12.
- The winners are not always the brand on the bag. Advertising, white-label tech, and refrigerated micro-fulfillment are quietly the highest-margin layers of the stack.
Why grocery delivery economics matters in 2026
US grocery is a $1.6 trillion industry, and online penetration has climbed past 14 percent according to US Census Bureau retail e-commerce data. That share is small in percentage terms and enormous in absolute dollars. Even a single point of margin movement translates into billions in profit or loss across the system. The bigger picture matters too, because grocery is the most frequent retail trip the average household makes, and whoever owns the digital touchpoint owns the most valuable data set in consumer goods.
If you are building, investing in, or competing with a grocery delivery operation, the unit economics decide everything. They decide whether you can afford a second warehouse, whether your private-label margins can subsidize your delivery losses, and whether your investors will allow a third capital raise. They also decide whether your shoppers earn a wage they will return for next month. Understanding the math is no longer optional; it is the only way to read the rest of the retail story, which we cover in our pillar on the state of retail.
Key terms and definitions
Before we dig into who keeps the money, it helps to settle the vocabulary. Grocery delivery uses words that overlap with restaurant delivery and e-commerce, but the economics behind each term are different.
- Marketplace model: A platform (Instacart, DoorDash, Uber Eats) lists a retailer’s inventory, takes the order, dispatches a contractor, and shares revenue with the store.
- First-party (1P) delivery: A grocer (Kroger, Walmart, Albertsons) owns the app, picks orders in store or in a dedicated facility, and either employs drivers or hires a delivery contractor.
- Micro-fulfillment center (MFC): A small automated warehouse, often bolted onto a store, that picks ambient and chilled SKUs faster than humans can.
- Dark store: A retail-shaped facility with no shoppers inside, used purely to fulfill delivery orders.
- Click-and-collect: Online order, in-person pickup. Cheaper than delivery for the retailer, popular with US households since 2020.
- Take rate: The percentage of order value the platform keeps after paying the retailer and shopper.
- Contribution margin: Revenue minus the variable cost of a single order (picking, packing, delivery, payment fees). The number that decides whether scale helps or hurts.
These definitions matter because they shape who carries which risk. A marketplace platform carries marketing and tech costs; a first-party retailer carries inventory and labor. When a grocer says “delivery is profitable,” ask which version of delivery they mean.
How grocery delivery economics actually work
A useful way to read the economics is to follow a single $100 grocery order from checkout to doorstep. The order looks identical to the shopper. The income statement underneath does not.
The $100 order, broken down
Across published data from Instacart’s S-1, DoorDash investor decks, and Kroger’s 10-K filings, a representative $100 US grocery delivery order distributes roughly as follows.
| Component | Marketplace order | Retailer 1P order |
|---|---|---|
| Cost of goods sold (groceries) | $72 | $72 |
| Store labor for picking | $3.50 | $4.50 |
| Last-mile delivery (driver pay + fuel) | $7 | $6 |
| Platform / tech overhead | $4.50 | $2.50 |
| Payment processing | $2 | $2 |
| Customer service, refunds, shrink | $2.50 | $2.50 |
| Marketing & promo allowance | $4 | $3 |
| Operating profit at order level | $4.50 | $7.50 |
That $4 to $7.50 of operating profit is gross of corporate overhead, real estate, and capital expenditure, so the true net is much thinner. It is also before the side revenue, which is where the real story lives.
Where the money actually shows up
If you only look at the basket math, grocery delivery is a near-zero business. The line items that move the picture into the black are usually three: retail media, membership fees, and private label.
- Retail media networks. Instacart Ads, Walmart Connect, Kroger Precision Marketing and Albertsons Media Collective sell ad placements to consumer brands. Margins on retail media run 70 to 90 percent. Instacart reported advertising revenue alone covered most of its operating costs in 2024.
- Membership. Walmart+, Instacart+, Amazon Prime and Costco’s online tier collect $99 to $139 per household per year. Even a 30 percent attach rate among delivery users transforms the unit economics.
- Private label. Store brands like Kroger’s Simple Truth or Aldi’s Specially Selected carry 25 to 40 percent higher margin than national brands. When private-label penetration on delivery rises, the basket math swings.
The headline is simple: the delivery itself loses or barely breaks even; the surrounding ecosystem makes the money. If a platform cannot build a profitable ad business, or a retailer cannot push private label, the model does not work no matter how many baskets they ship.
Who keeps what: a layer-by-layer view
Five parties touch a grocery delivery order: the shopper (consumer), the retailer, the platform, the gig worker, and the brand on the shelf. Each layer has its own margin profile.
| Layer | Revenue source | Typical margin | Risk profile |
|---|---|---|---|
| Consumer brand (e.g., General Mills) | Wholesale price + trade promo | 15 to 25 percent | Loses shelf control; pays for visibility |
| Retailer (e.g., Kroger) | Markup + media + membership | 2 to 4 percent net | Carries inventory, labor, real estate |
| Marketplace platform (e.g., Instacart) | Take rate + ads + subscriptions | Negative to 8 percent | Marketing-heavy; data-rich |
| Gig worker | Per-order pay + tips | $7 to $14 net per hour | Carries vehicle, fuel, tax |
| Tech vendor (e.g., Symbotic, Takeoff) | SaaS + hardware | 20 to 40 percent | Customer concentration |
Notice the inversion: the layers furthest from the shopper (tech vendors and brand advertisers) earn the highest margins, while the layer closest to the shopper (the gig worker) earns the thinnest. That structural inversion explains why every player tries to climb into the high-margin layers, often by acquiring an ad business or launching a private-label SKU.
Common mistakes operators make in grocery delivery
Operators who run grocery delivery programs (whether at a national chain or a regional independent) tend to repeat the same expensive errors. Most are not strategy errors; they are math errors hidden in plain sight.
- Treating delivery fees as revenue, not a cost offset. A $4.99 delivery fee usually does not cover the $7 to $9 of last-mile cost. Counting it as a revenue line distorts every dashboard downstream.
- Subsidizing every basket equally. A $40 single-bag order and a $180 full-cart order cost almost the same to deliver. Flat-fee promos drain margin disproportionately on small baskets.
- Pricing parity with stores. If online prices match in-store prices, the retailer eats the entire delivery cost. Light price markups (3 to 6 percent online) are now standard at major US chains; ignoring this leaves money on the table.
- Underinvesting in pick accuracy. A 4 percent substitution or out-of-stock rate is normal; an 8 percent rate kills repeat orders. Fresh items are especially sensitive, which connects to the cluster topic of fresh food supply chains.
- Confusing GMV with profit. Many operators chase order volume because it improves marketplace ranking. Without margin discipline, every additional order can deepen losses rather than amortize fixed costs.
- Ignoring tip structure on shopper retention. Tip-pooled or low-tip routes burn out contractors. High turnover raises the average pick error rate, which raises refund rates, which kills contribution margin in a quiet loop.
These mistakes share one trait: they all stem from looking at the top line instead of the contribution margin. Any operator who cannot recite their per-order contribution margin from memory is, statistically, losing money on delivery.
Examples from US retail and e-commerce
The math becomes more concrete when you map it to specific companies. The big US players have made very different bets, and the early 2026 picture shows which bets are paying off.
Walmart: scale and click-and-collect as ballast
Walmart’s e-commerce business reportedly turned profitable on a contribution basis in fiscal 2025, primarily because more than 60 percent of online grocery orders in the US are picked up in store rather than delivered. Click-and-collect carries roughly half the variable cost of home delivery, which pulls the blended margin up. Walmart Connect, the company’s ad platform, then pours high-margin revenue on top of the basket. Walmart+ adds another $98 per year per member.
Kroger: the data play
Kroger’s joint venture with Ocado for automated micro-fulfillment was, by mid-2025, a partial retreat: the company paused new facility builds while expanding store-picked delivery. The economic reason was simple. The Ocado-style sheds cost hundreds of millions and required higher order density than most US suburbs deliver. Meanwhile, Kroger Precision Marketing has grown into a multi-billion-dollar advertising business with margins that dwarf the grocery floor. The lesson is that retail media can fund delivery losses for years; the lesson is also that capital-heavy automation does not always pay off in lower-density US markets.
Instacart: a platform that became a media company
Instacart’s 2023 IPO disclosed the structural truth out loud. Roughly 30 percent of revenue, and a large share of profit, came from advertising rather than delivery. The platform fee on retailers, plus the customer-side service fee, barely covered shopper pay and overhead. The ad business turned the company profitable. The risk is concentration: if Walmart and Kroger sell their own ads directly to consumer brands at lower take rates, Instacart’s economic engine is the part most exposed.
Amazon Fresh and Whole Foods: bundled with Prime
Amazon does not break out grocery delivery profitability, and the structural reason is that it does not need to. Grocery is a frequency play that feeds Prime renewals. When 200 million Prime households shop more often, the entire flywheel benefits. The economics that matter are membership retention and ad attach, not the bag of strawberries at the door.
Albertsons: third-party plus first-party
Albertsons runs a hybrid model: it sells through Instacart, DoorDash, and Uber Eats while also offering first-party delivery via its own app. The economic trade is visibility versus margin. The platforms supply demand that Albertsons could not generate alone; the first-party app carries higher contribution margin and richer data. Most regional grocers will end up in this hybrid posture, and our 2026 outlook covers this in our 2026 retail industry outlook.
Aldi and Lidl: the private-label advantage
The hard-discount grocers approach delivery from a structurally different starting point. Because more than 70 percent of their SKUs are private label, every order sold through delivery already carries a richer margin than a comparable basket from a traditional chain. Aldi has used Instacart aggressively as a customer-acquisition channel rather than a profit center, treating the marketplace fee as the cost of reaching shoppers who would never enter a small-format store. Lidl has been slower to scale online but follows a similar logic. The deeper lesson is that the product mix matters as much as the delivery model: a grocer with strong private label can absorb platform fees that would crush a national-brand-heavy competitor.
Whole Foods and the premium segment
Premium grocers carry their own economic quirks. Average basket sizes at Whole Foods and Erewhon run 30 to 60 percent higher than the US mainstream, which spreads picking and delivery costs across more revenue per order. The trade-off is a more demanding shopper, where substitution tolerance is lower and customer-service costs are higher. The premium tier also leans more heavily on specialty subscriptions, prepared meals, and ad-supported placements from emerging better-for-you brands willing to pay above-average rates for shelf visibility.
Forecasting the order-level P&L for your own operation
If you are inside a grocer or a platform and need to build a credible model, the trick is to separate variable costs that scale with each order from fixed costs that scale with the program as a whole. The former drives contribution margin; the latter drives break-even volume.
| Cost type | Examples | Behavior | How to manage it |
|---|---|---|---|
| Variable, per order | Picking labor, driver pay, fuel, payment fees | Scales linearly | Optimize basket size, batch orders, route density |
| Semi-variable | Customer service, refunds, shrink | Scales with errors | Tighten pick accuracy, training, exception flows |
| Fixed program-level | App development, MFC depreciation, regional ops | Flat regardless of volume | Amortize across higher GMV and ad revenue |
| Strategic investments | Retail media, membership tooling, data platform | Front-loaded, long payback | Track separately so they do not pollute order P&L |
A common analytical mistake is to push fixed and strategic costs into the per-order line, which makes every order look unprofitable and tempts leadership to cancel the entire program. Separating the layers reveals that the order itself is often a sound (if thin) business, while the loss sits in the strategic investments that are supposed to pay off over years.
A practical forecasting playbook for an operator launching a delivery program looks like this. First, lock the variable cost per order using time-and-motion data from existing pickers and a realistic estimate of last-mile cost in your geography. Second, model the basket-size distribution your shoppers actually generate, because mean and median differ sharply in grocery. Third, layer in expected ad revenue per order based on category mix; CPG brands pay more for visibility in pet, baby and personal care than in produce. Fourth, set membership penetration assumptions on a curve, not a flat number, because Walmart+ and Instacart+ data shows steady multi-year ramps rather than instant adoption.
Tools, partners and vendors worth knowing
If you are evaluating or building grocery delivery infrastructure in the US, the vendor landscape matters because the right partner can flip your unit economics. The wrong one can sink them.
- Marketplace platforms: Instacart, DoorDash, Uber Eats. Best for fast reach, weakest for margin control.
- White-label e-commerce: Mercatus, Rosie, Self Point, Swiftly. Power independent grocer websites and apps; meaningful fixed costs but stronger first-party data.
- Micro-fulfillment automation: Symbotic, AutoStore, Takeoff Technologies, Fabric. Compelling at high density; capex-heavy.
- Last-mile delivery: Roadie, AxleHire, Veho, Shipt. Often used as overflow capacity for spikes and rural ZIP codes.
- Retail media platforms: Criteo Retail Media, Citrus Ad (Epsilon), Pacvue. Power the ad layer that funds the rest.
- Forecasting and shrink reduction: Afresh, Shelf Engine. Specifically built for perishables; high ROI on fresh categories.
Picking the stack is a strategic decision, not a procurement one. A grocer that outsources its app, its delivery, and its ads will struggle to keep any of the marginal dollar. A grocer that owns all three (Walmart-style) faces a much larger fixed cost base and needs scale to justify it. For deeper coverage of the vendor landscape, see our overview of tools and vendors for supermarkets and grocers in 2026.
What this means for the next two years
The trajectory through 2027 looks less like the rapid-growth, capital-flooded story of 2020 to 2022 and more like a slow consolidation. Several forces are visible in the data already.
- Retail media becomes the central business. Within three years, ads, data, and on-site retail media will likely be the largest profit pool in US grocery, larger than the delivery service itself.
- Click-and-collect outgrows delivery in suburbia. The economics simply work better outside the dense urban core, and US households have shown a strong preference for free pickup over paid delivery.
- Gig pay normalizes upward. Several state-level minimum-wage frameworks for app workers (notably in New York and Washington) are pushing platforms toward employed or hybrid models. Expect $2 to $4 of additional delivery cost per order in regulated markets.
- Automation goes selective. Instead of full dark-store buildouts, retailers will add automation to specific high-density nodes and use store picking everywhere else.
- Private label penetrates online faster than in store. Algorithmic recommendations favor store brands; this is the cleanest margin lever any grocer has.
Pulled together, these forces describe a market that is still growing in absolute terms but maturing fast in margin terms. The era of growth at any cost is over. Operators who understand the layered economics will outlast those who keep treating grocery delivery as a single line item, and the broader context lives inside our state of retail pillar, which connects this to department stores, experiential retail, and the rest of the modern grocer playbook.
FAQ
Is grocery delivery actually profitable for the retailer?
On the basket itself, usually not. Contribution margin on a US grocery delivery order ranges from negative 2 percent to positive 5 percent depending on basket size, picking method, and delivery model. The order becomes profitable only when retail media, private label, and membership revenue are added on top.
How much does Instacart actually keep on a typical order?
Across platform fees, service fees, and tip-related adjustments, Instacart’s take rate runs roughly 15 to 25 percent of order value. Advertising adds several more percentage points of effective revenue per order, but only on orders involving paid-placement SKUs. Net contribution per order is positive only because of the ad business.
Why do groceries cost more on the app than in the store?
Most major US chains apply a 3 to 10 percent online markup to offset picking, packing, and platform costs. Some categories, like fresh produce and meat, often carry larger markups because shrink and substitution risk are higher. The price you see is the retailer’s way of recovering part of the delivery cost without breaking the headline delivery fee.
Do gig shoppers actually make a living wage?
Gross pay on US grocery platforms typically runs $14 to $22 per hour, including tips. After vehicle, fuel, maintenance, and self-employment tax, net earnings often land between $7 and $12. Earnings vary widely by market, hours worked, and tip culture. Regulatory pressure in major cities is starting to lift the floor.
Are dark stores and micro-fulfillment centers the future?
For dense urban areas and high-volume corridors, yes. For most of US suburbia, the math is harder because order density per square mile rarely justifies fixed automation costs. Expect a hybrid future where stores remain the primary fulfillment node and automation is bolted on where density allows.
How does click-and-collect change the economics?
Click-and-collect removes the last-mile delivery cost and shifts the labor of transport back to the consumer. That single change typically lifts contribution margin by 4 to 7 percentage points compared with home delivery. It explains why most major US grocers heavily promote pickup and many keep delivery fees high enough to nudge shoppers toward the cheaper option.
What is the single biggest lever a grocer can pull on delivery profitability?
Retail media. Building or partnering with a high-quality on-site advertising platform can add multiple points of high-margin revenue across the entire delivery business. After media, the next strongest levers are private-label penetration and basket-size thresholds for free delivery.
Will grocery delivery prices keep rising?
Probably yes in the short term, especially in markets with new gig-pay regulations. Higher service fees, dynamic delivery pricing, and tighter promo discipline are all visible across major platforms heading into 2027. The era of below-cost delivery, subsidized by venture capital, is largely over.