Hard to believe, but August is here. It’s the last full month in which a delivery hiccup teaches you something instead of costing you someone, and most of what happens to your operation in December is getting decided right now.
Whether you’re deciding it on purpose or not.
Some background on why 2026 feels different. UPS, FedEx, and USPS used to move 85% of this country’s parcels between them. In 2025, they moved 61% of 23.9 billion, and Amazon’s own network now out-delivers each of them. The pool you buy capacity from got reshuffled quickly, and mostly without asking you.
The Q3 job is to make that bigger carrier pool behave like one network. You choose the provider per order, your brand stays on tracking, customer data returns to your systems, and your team catches the failure before the customer. That’s what controlled delivery network growth should look like.
Before adding another market or handing a provider more peak volume, we’d run five checks, in this order.
Figure Out What a Delivery Actually Costs You
Start with what it took to get the package to the door. The delivery provider’s quote is nice and clean because none of your cleanup is on it. If the first attempt fails, your dispatcher’s chasing the order while support’s answering the customer, and you may still pay to send it out again. A refund can be waiting at the end of all that.
Add it up and divide by successful deliveries. That’s the number that tells you whether the new provider helped and how much volume it should get next.
August is boring, which is exactly why it works. The network’s running normally. By October, peak’s already bending the numbers, so you’ll be measuring strain and calling it a baseline.
We’ve never met an ops lead who regretted having this number. We’ve met a few who first saw it in a November postmortem. When finance wants the AI-versus-manual math, you’ll start from the same sheet.
Test the New Capacity While Mistakes Are Still Cheap
Now put every backup lane against that cost-per-successful-delivery number before peak. Most of the capacity behind this year’s delivery network growth has barely been live for six months, so there’s still plenty you haven’t seen from it.
The oddest piece is USPS. After spending years steering small shippers away from its local network, it opened more than 18,000 neighborhood delivery units to bids. The winning lanes came online from July into August, which means they’re carrying real boxes right now.
Target now ships from more than 100 stores at roughly $2.50 less per package than the major national parcel networks. Amazon keeps bringing 30-minute delivery to new cities. None of that is your problem directly. But it sets the grading curve.
So the August move is simple: every zone gets a primary and a backup, and the backup carries live orders before Labor Day. If it only exists as a login you created in March, it doesn’t count. And a marketplace account is not the same purchase as infrastructure, which you find out at the worst possible price.
Put Your Drivers and the Network on One Screen
Once each backup has carried live orders, put it on the same screen as your in-house fleet. We see the same thing on discovery calls all the time: company vans running at 60% while an outside provider gets paid for orders those vans could have taken. The fleet sits in one system and the provider network in another, so the dispatcher never sees the choice.
Keep the vans. Dispatch them from the same place as the rest of your network. An order goes to the company van when it can handle the job for less. The next one goes outside because that provider has the right vehicle or a better record on the lane. Burq customers running this kind of hybrid setup have cut dispatch-planning time by up to 90%.
When finance asks what Tuesday cost, there’s one answer instead of four exports.
Fix Recovery Before You Add Another Handoff
Every provider you add adds a handoff, and a handoff is where an order can sit for hours with nobody’s name on the problem. The pickup that never happened, or a driver parked two miles out with no update, or a delivery photo of the wrong porch. Each one costs more than the redelivery, mostly in ways no invoice shows.
Retail has already said what it wants from you here. A Macy’s supply chain director put it plainly at Home Delivery World: “When you’re not reliable is when we’re going to lose our customers.” Ryder’s ecommerce lead sees shoppers gladly trading speed for a window they can trust. And McKinsey has 6-in-10 ranking a kept promise over a fast one.
Manual recovery has a ceiling: however many dispatchers you can hire. So Burq’s Pulse AI watches every order in flight, reroutes the 60-80% of failures it can catch early, and answers the where’s-my-order question before it gets asked, which cuts those tickets 40-60%. Live support sits behind the rest.
Check Whose Name Is on the Tracking Page
Somewhere in your current mix, odds are, a customer is getting a text with somebody else’s logo on it. That’s the quiet cost of a wider network, and it compounds, because the doorstep is the one brand moment your customer never skips.
It’s also worth more this quarter than it’s ever been. Online grocery has grown more than 20% a year for six straight quarters and now takes about 19 cents of every grocery dollar. Back-to-school is tracking to a record $43.3 billion, and 62% of those shoppers had started buying by early July. That traffic is on your trucks already.
Run the audit before the calendar flips. Look for whose name is on the tracking page, whose number sends the update, who keeps the customer record across every location. Same test whether the box holds groceries, Friday’s dinner rush, a patio set, or a flower delivery.
What You Want to Be True by September
The season itself should be ordinary. Holiday ecommerce is forecast to grow about 6.6% this year, close to last year’s pace. What changed is the ground under it. The provider pool, the pricing, and the customer’s patience for a broken promise all moved inside 12 months, and none of that can be hired for in November.
By September you want five things to be true: the real cost on paper, a proven backup in every zone, your drivers and your providers on one screen, something watching for failures around the clock, and your own name on the tracking page.
That last stretch is what Burq is for. One integration puts hundreds of vetted delivery providers alongside your own fleet on one set of controls, with Pulse AI riding along on every order. It moves more than $500 million in deliveries a year at a 99%-plus success rate across 14,000-plus businesses, and the enterprise setup takes your own contracted rates, with the SOC 2 Type II, ISO, and GDPR paperwork your security review will ask for.
Book a demo and bring the zone that scares you. Watching Pulse AI dispatch and recover your own orders will tell you more than another article can, and there’s still a month to act on what you see.
