Amazon spent 12 years learning that same-day delivery is a geography problem before it’s a speed problem, and paid about $4 billion for the lesson. Prime Now launched in Manhattan in 2014, a warehouse full of the things people buy when they can’t wait, promising any of it at the door inside an hour.
In dense metros it worked beautifully. Everywhere the customers sat farther apart, it didn’t. After seven years, Amazon folded the service into its app and stopped talking about it.
Then, the same idea came back in May at half the time. Amazon Now promises 30 minutes in Atlanta, Dallas-Fort Worth, Philadelphia, and Seattle, picked by three or four people working a room about the size of a corner store. The concept didn’t improve between 2021 and 2026, the network did.
Your customers followed none of that and wouldn’t care if they had. They only ever saw the result, and the result reset what counts as a reasonable wait. A free delivery now buys you 2.7 days of patience, down from 3.5 a few years ago, and for groceries it’s under a single day.
Whether or not you have Amazon-like capabilities, that average is the standard you’re held to now.
What Amazon Trained Your Customers to Expect
The standard was built by repetition rather than advertising, one uneventful delivery at a time. U.S. Prime members took delivery of more than 8 billion items on the same day or next day last year, a 30% jump in 12 months. Walmart worked the same ground out of its stores, running millions of sub-30-minute drops across better than 19,000 ZIP codes in a single quarter.
Order that way for a few years, and delivery stops being something you evaluate. The shopper adding dish soap and bananas to a cart at 9 p.m. has already settled three things. Fast doesn’t cost extra, the window on the screen is real, and their street gets served like everyone else’s. The first two you can fix with pricing and process. The third is a promise about geography, and nobody checks it until the day it fails on them.
The New Baseline: 2.7 Days and Under a Day for Food
AlixPartners has asked Americans the same home delivery questions every year since 2012, which makes it the longest continuous read this industry has. This year, 94% of respondents said free shipping affects what they buy. Marc Iampieri, who co-leads logistics at the firm, described Prime-era speed as “the floor, not the ceiling” for retail, which is a polite way of saying the expectation compounds and never resets.
Grocery reached that floor before anyone else. Roughly 80% of online grocery delivery orders arrived same day last quarter, and e-grocery has grown by more than 20% for six consecutive quarters.
Buried under all of it is a number that should change how you pitch this internally. Walmart charges around $10 to deliver, and a few dollars more for a three-hour window. Better than 30% of its customers pay for the upgrade.
Nationwide Coverage Is Harder to Buy Than Speed
Speed is something you can buy in one place. Amazon proved it in 2014, when Prime Now hit an hour in Manhattan from one warehouse. Add drivers, tighten pick times, cut dwell, and one metro gets faster inside a quarter.
Coverage doesn’t work that way. Each new market costs about what the first one did and arrives with a fraction of the orders, so cost per stop climbs exactly where order density drops. That’s what killed Prime Now in the second city and the fourth, and it’s why $4 billion bought Amazon four cities. Money doesn’t create density. Volume per square mile does, and you can’t manufacture that alone in a market where you have a handful of stores.
Walmart’s answer was to already own it. Turning 10,900 stores into warehouses is the only reason it covers 95% of American households inside three hours. But you’re not Walmart.
Somebody else’s density is for rent, though. The provider working that ZIP code pools orders across dozens of merchants, which makes the stop economics work where yours never would. Kroger and Albertsons both buy into that instead of buying vans for every market they’d like to serve.
What a Missed Same-Day Delivery Order Costs
The failures a thin market produces look cheap on the day they happen. A refund or a redelivery, a support ticket, an hour of your dispatcher’s day. Call it a few dollars and some patience.
The customer is the expensive part. In that same AlixPartners work, 52% said one or two botched deliveries would end the relationship, and close to 90% said a late order does damage even after the apology. Veho also found in August that 40% have walked over a package that never arrived, without filing a complaint or leaving a review. The refund closes the ticket while the order history quietly stops.
None of that lands anywhere you’d look for it. Delivery costs rose year over year for 83% of retailers, and 64% say the channel still doesn’t pay next to a sale rung up in the store. Which makes the rate on your provider contract a lousy number to manage by. It counts the trip and nothing else.
Meeting the Bar Without Handing Over Your Brand
Procurement figured this out before marketing did. For the first time in 14 years of the AlixPartners survey, retailers ranked reliability ahead of price when picking a lead provider. And 55% now work with somebody outside UPS, FedEx, and USPS. One provider is one point of failure, and a bad week in one region costs a day no one gives back.
Buying coverage the easy way means a marketplace, and marketplaces charge in a currency that never shows up on the invoice. Their name goes on the tracking page. Their app owns the notification, the delay alert, the apology. You get the order, and they get the customer, plus a cut of whatever the customer spent.
Keeping both takes more infrastructure. Route every order to whatever best fits it, like a van for the bulky item or your own driver for the run two miles from the store. Then let all of it arrive under your name. You’re renting the density, not the relationship.
The Bar Moves Again Next Year. Build for That.
The bar will move again, because it always does: 2.7 days becomes 2.2, three hours becomes one across the top 40 metros, and shoppers will absorb it inside a week without saying a word. Build around this year’s number, and you’ll be rebuilding by spring.
That’s what Burq is for. One system runs hundreds of delivery providers alongside your own drivers, and it works the same for a florist heading into Mother’s Day, a restaurant group, or a retailer bringing its own contract rates. The engine picks the provider, watches the route while it’s live, and reroutes in seconds when something starts to slip. That prevents 60-80% of potential delivery failures, cuts “where is my order” inquiries by 40-60%, and holds a 99%+ success rate across more than $500 million delivered a year.
It took Amazon a decade, two attempts, and $4 billion to put 30-minute delivery into four cities, so nobody expects you to win that race overnight. It doesn’t get you off the hook, though. The person with dish soap in their cart at 9 p.m. still has expectations.
Book a demo with Burq and walk a live hybrid dispatch, with your drivers and the network in one view and the reroute logic running on a real exception.
