Merchant Center Shipping Settings in 2026: Handling, Transit and Free Delivery

A shopper clicks a $49 product, reaches the cart, and finds $8.90 shipping with a 7–14 day estimate. The click is paid for; the order never happens. In your reports this looks like a landing page problem. It is not. You bought traffic that your delivery terms cannot serve, and Google never had the data to filter for it.

Merchant Center shipping settings are not accounting paperwork. They are product data: they decide which items are eligible to serve in several countries, what annotation the listing carries, and what total price the shopper compares before clicking. This guide covers how the shipping block is structured, how to set handling and transit times honestly, when free shipping beats a discount, which feed attributes override account settings, and what to check before peak season.

Why delivery data moves impressions, CTR and order economics

Shipping cost and speed appear in the listing itself — sometimes as a “Free delivery” or “Get it by” annotation. That plays out on three levels:

• Eligibility. In many countries valid shipping settings are a serving requirement. No settings, no impressions, regardless of budget.
• The click. Shoppers compare the landed price — item plus delivery — and the promised date. A listing at $49 with free two-day delivery beats the same product at $46 with $9 shipping in ten days.
• The conversion. If your feed is more optimistic than your warehouse, you are paying for clicks that are guaranteed to hit an unpleasant surprise at checkout.

Shipping cost is part of the price. Optimizing bids while ignoring delivery means bidding on a product that is more expensive than you think it is — in the only opinion that matters, the shopper’s.

It is the same mechanic as the price competitiveness report: buyers compare totals, not line items.

How Merchant Center shipping settings are structured

Four layers, and most confusion lives between the second and the third.

Layer	What it defines	Common mistake
Shipping service	Country, currency, which products it applies to	One service for markets with very different logistics
Zones	Regions, postal codes, region groups	No separate zone for remote areas
Rate	What delivery costs in that zone	An average rate instead of the real spread
Delivery time	Handling plus transit	Transit entered, handling forgotten

A shipping service can be limited to a subset of products: tag them with shipping_label in the feed and select that label in the service’s advanced settings. That is the only clean way to separate, say, oversized furniture from small hardware with different rates.

Handling time and transit time, told honestly

The delivery estimate on a listing is the sum of two things:

• Handling time — from payment to handover: picking, packing, carrier pickup. Entered as a min–max range in business days.
• Transit time — from carrier pickup to delivery. Also a range, and worth setting per zone.

Four things break the math:

1. Cutoff time. If orders after 3 p.m. ship the next morning, say so in the cutoff setting, or the system promises same-day handover you do not perform.
2. Business-day calendars. Weekends and holidays are excluded by default, but your warehouse may work Saturdays — or close for a regional holiday Google does not know about.
3. Ranges beat averages. “3–7 days” is honest; “5 days” means half your orders arrive late by definition.
4. Different zones, different transit. One national estimate either over-promises to distant regions or under-sells your speed in nearby ones.

Some categories and markets support much tighter settings — down to same-day windows measured in minutes to hours, plus a “Get it by [date]” badge. Those formats reward real logistics coverage and expose any gap between promise and reality almost immediately.

Rate types and when each one fits

Rate type	Fits when	Risk
Flat fee	Homogeneous catalog, one market	Loses money on bulky items
Always free	High margin, shipping priced in	Small baskets erode margin
Free over a threshold	You want a higher average order value	A threshold far above AOV motivates nobody
Table by zone / weight / price	Wide catalog, large geography	Hard to maintain by hand
Carrier rates	Direct carrier integration	Listing price fluctuates

A working starting point: set the free-shipping threshold roughly 15–25% above your current average order value. Below that you are giving delivery away to people who were already buying; far above it, the threshold stops being an incentive and becomes an irritant. Treat the range as a benchmark to test, not a rule.

When free shipping beats a discount

Do not compare “$8 off the price” with “$8 off shipping” — compare profit per order. Free shipping:

• leaves the product price untouched, so your position in price comparisons holds;
• applies once per order rather than per item, which makes it cheaper than a discount on multi-item baskets;
• earns an annotation on the listing, so it buys attention as well as economics.

A discount wins where a price threshold decides the comparison and where baskets are almost always single-item. Either way, judge it on margin rather than revenue — the method is in POAS vs ROAS and in media buying unit economics.

Feed attributes vs account settings

Account settings are the default; feed attributes are the exceptions. Knowing which wins saves hours of confusion.

Attribute	What it does	Use it for
shipping	Sets delivery price for a specific product and region	Items with non-standard logistics
shipping_label	Tags a group so a rate applies only to it	Oversized, fragile, frozen goods
max_handling_time	Overrides maximum handling days	Made-to-order, pre-orders
min_handling_time	Overrides minimum handling days	Build-to-order products
transit_time_label	Points a product at a specific transit table	Multiple shipping origins
shipping_weight, dimensions	Feeds weight- and size-based rate tables	Carrier and table rates

The rule: if the attribute is present in the feed, it wins for that product. Hence the classic “I changed the rate in the interface and nothing happened” — the price was hard-coded in shipping at the item level. Bulk changes to those fields are best handled with feed rules and supplemental feeds rather than a platform ticket.

Shipping and local inventory

If you have physical stores, delivery is no longer the only way to get the product. Pickup and local offers compete with courier shipping for the same click, so they have to be configured together: a “ready today” pickup option makes paid five-day delivery in the same city pointless. What changed in that area is covered in local inventory ads being on by default.

What delivery actually costs you

Most shipping decisions are made on instinct because very few teams calculate cost per order. The minimum workable formula:

Delivery cost per order = carrier rate + packaging + pick-and-pack labour (minutes × hourly rate) + return rate × (outbound + inbound shipping) + non-delivery rate × full logistics cost.

The last two terms are where margin disappears in any category with fit or size uncertainty. At a 20% non-delivery rate, every delivered parcel carries a quarter of the logistics cost of the undelivered ones — so “free shipping” is not $4.50, it is closer to $5.60. Put that number into order economics before you announce a threshold.

Then the question becomes answerable. Not “can we afford free shipping,” but “how much conversion lift does free shipping need to pay for itself.” One line: required lift = delivery cost ÷ (order margin − delivery cost). If the answer is a 40% lift, drop the idea. If it is 6–8%, you have a testable hypothesis.

Three workable shipping models

Model	Fits	What to configure
Shipping priced into the product	High margin, uniform size, one market	Zero-rate service, product price with headroom
Free over a threshold	Wide catalog, AOV needs lifting	Rate table plus a threshold 15–25% above AOV
Transparent paid rate	Bulky goods, thin margin, long distances	Zones plus weight or dimension rates, honest estimates

Mixing models inside one account is fine — as long as it is done with labels. What is not fine is running one model in the feed and a different one at checkout. The gap between the listing promise and the final total is not only lost orders; it is a structural conversion problem that teams then spend months hunting for in cart design.

Delivery, returns and repeat purchase

Shipping terms do not stop working after the first order. Two effects that rarely reach a campaign report:

• Returns. The faster the promise, the less tolerance a late delivery gets. Two days late against “tomorrow” generates complaints; two days late against “3–7 days” does not.
• Repeat purchase. Predictable delivery lifts repeat rate more reliably than a one-off discount. Over an LTV horizon that changes your allowable acquisition cost — which is exactly why shipping belongs in the same model as your bids.

Once you bid on profit rather than revenue, delivery stops being a logistics line item and becomes a variable inside the bid itself. How that plays out at product level is covered in Product Value Optimization in PMax and Shopping.

Mistakes you can see in the P&L

1. Optimistic delivery dates. CTR rises for a week, cancellations and refunds rise for a quarter. Bidding learns on conversions, not clicks, so the system eventually serves you less.
2. No zone for remote regions. Either you subsidize expensive logistics or the shopper discovers the real cost at checkout.
3. Free-shipping threshold below AOV. Shipping is effectively always free, margin drops and the incentive never fires.
4. Stale holiday calendar. In December your promises slip by two or three days — the worst possible month for that error.
5. Feed and site disagree. The listing promises one thing, checkout charges another. Beyond lost orders, that is also a destination-quality problem: see Google Ads destination requirements.
6. One average rate for everything. Light items overcharge the customer; heavy items undercharge you.

Building a rate table that survives contact with reality

Rate tables are where shipping settings usually mutate into an unmaintainable forty-row monster. The discipline that keeps them usable:

1. Zones first, weight second. Group geography into three to five zones by actual logistics cost rather than administrative boundaries, then add three or four weight bands inside each. Fifteen combinations cover most retail catalogs.
2. Set bands from data, not round numbers. Plot a histogram of order weights over a quarter and cut at the gaps in the distribution, not at “under 1 lb / under 10 lb” invented at a whiteboard.
3. One label, one logic. Do not try to encode oversized, fragile and frozen in a single shipping_label. Every exception with its own economics earns its own label.
4. Leave headroom. Carrier rates move and table edits take time. A small buffer in the upper bands keeps you out of loss-making territory between revisions.
5. Document it. One page describing what each label means and who owns it saves weeks the first time somebody else has to edit the table.

One more: when a carrier raises rates mid-season, resist the urge to push the increase straight into every listing. Check which weight bands actually lose money first — the damage is often concentrated in the top 10–15% of orders, and repricing those beats repricing the catalog.

Pre-peak checklist

1. Compare actual delivery times over the last 30 days against what the settings promise. A gap over one day is a settings problem, not a reporting one.
2. Update holiday calendars and cutoff times for the season.
3. Confirm zones cover your entire selling geography, including the awkward edges.
4. Recalculate the free-shipping threshold from this year’s AOV, not last year’s.
5. Export items carrying shipping and max_handling_time overrides and verify each override is still intentional.
6. Check that bulky items have weight and dimensions populated.
7. Compare landed price — product plus shipping — against competitors on your top 50 SKUs.
8. Align promotions so they do not contradict delivery terms: see Merchant Center promotions and annotations.

Measuring the effect

Shipping changes are almost never evaluated correctly by eye, because they collide with seasonality. The minimum honest setup:

• record the change date and segment affected products via shipping_label or custom labels so you keep a control group;
• look past CTR to click-to-order conversion and cancellation rate;
• measure profit per order after actual delivery cost, not revenue;
• for segmentation, custom labels in the product feed do the job.

If impressions drop after a shipping change, check neighbouring feed issues before rolling back: product identifiers — GTIN, MPN and brand affect matching and visibility independently of delivery terms. And if you are opening a new market, shipping has to be configured together with the language side of the campaigns — see AI campaign localization for Search and international campaign launch.

When the settings look right and products still do not serve, the constraint has usually moved to the account: history, structure, category eligibility. That diagnosis is part of what PPC Rebels does inside its Google Ads services, and new markets are typically opened on agency Google Ads accounts that already have serving history.

FAQ: Merchant Center shipping settings

Are shipping settings mandatory?

In most countries where Shopping ads run, yes — products are not eligible to serve without valid delivery data. There are exceptions, but they are a poor foundation for a strategy.

What matters more, cost or speed?

It depends on the category. Impulse purchases are usually decided by speed, considered purchases by landed price. Test it on a product segment instead of arguing about it.

Can I offer free shipping on only part of the catalog?

Yes. Tag the group with shipping_label and attach a separate shipping service with a zero rate to that label.

How do I set different delivery times per region?

Through zones inside the shipping service — each zone gets its own transit range. If you ship from several warehouses, add transit_time_label to point products at the right table.

What happens if I promise faster than I deliver?

CTR improves first, then cancellations, refunds and complaints follow. Because bidding optimizes on conversions and order value, the click advantage is consumed by falling traffic quality within weeks.

Do weekends count in handling time?

Estimates run in business days, and the non-working-day calendar is configurable. If your warehouse works Saturdays, say so, or you will under-promise and lose the speed advantage you actually have.

Does the shipping attribute override account settings?

Yes, for the product that carries it. If interface changes seem to have no effect, look for hard-coded shipping values in the feed first.

How does shipping relate to price competitiveness?

The report compares product prices, but shoppers compare totals. A price advantage can be entirely erased by expensive delivery, so read the two together.

What should I do with oversized products?

Isolate them with their own label and a service priced by weight or dimensions, and populate shipping_weight and size attributes. An averaged rate on a mixed catalog loses money on exactly those items.

How often should shipping settings be revisited?

At least twice a year before peak periods, plus whenever carriers, rates or warehouses change — and any time your average order value moves, if you run a free-shipping threshold.

Does shipping affect Performance Max?

Indirectly but materially. Delivery data shapes conversion rate, and the algorithm allocates budget by predicted value, so weak delivery terms translate into products that quietly never scale.

Can I A/B test shipping terms?

Not as a native experiment. Test by product segment instead: change terms for one labelled group, hold another as control, and compare comparable periods rather than before-and-after.

Shipping is the second price tag on your product. The shopper sees it before the click and the algorithm sees it before the budget decision — the closer it is to the truth, the cheaper every order becomes.