If you're here because you saw the new Flow action and want to know whether you can stop clicking "Buy shipping label" 200 times a day: yes. Shopify added a buy-shipping-label action to Flow in mid-October, so a workflow can now pick a carrier service, buy the label and attach tracking without anyone touching the Orders page. You need Shopify Shipping switched on, billing attached, and a fulfilment location that actually holds the stock. Then the question stops being can I and becomes which orders should I let it touch. That second question is the whole job, and for Australian merchants it comes down to three variables: destination zone, parcel weight, and whether the order is one Flow should be allowed near at all.
The timing is not an accident. Click Frenzy lands in November, Black Friday a fortnight later, and then a summer Christmas with Australia Post cut-off dates that get earlier every year. Label buying is the step that quietly becomes the bottleneck at 300 orders a day, because it is manual, repetitive and nobody can do it from their phone.
What the Flow action actually does, and what it doesn't
The action operates on a fulfilment order, not an order. That distinction matters more in Australia than people expect, because if you run a Sydney 3PL and a Melbourne overflow location, one customer order can produce two fulfilment orders and your workflow will fire twice. That's usually correct behaviour. It is not always what the finance person expects to see on the shipping invoice.
You choose the carrier service, the package, and the extras. What Flow can't do is reason about your parcel. It will not notice that three candles in a 40 × 30 × 20 cm box are going to be rated on cubic weight, not the 2.1 kg on the scale. If your rate card applies a cubic conversion factor of 250, that box works out at 0.40 × 0.30 × 0.20 = 0.024 m³, times 250 = 6 kg chargeable. You just bought a 6 kg label on a 2.1 kg order and your margin on that SKU went with it. Configure the package dimensions in Shopify properly before you automate anything, or the automation will industrialise the mistake.
Australia Post vs Sendle: where each one earns its place
Both sit inside Shopify Shipping in Australia, so you can route between them in a single workflow. The split we use most often:
- Sendle for metro-to-metro parcels in the 500 g to 5 kg band where the customer is a residential address and there's no deadline attached. Door-to-door, no post office trip, carbon-neutral by default, which plays well if your brand makes sustainability claims. Keep those claims accurate, because the ACCC has been actively pursuing vague environmental marketing and "carbon neutral delivery" written loosely on a shipping page is the kind of thing that gets read closely.
- Australia Post for anything going to a PO Box or Parcel Locker, anything remote, anything where the customer paid for Express, and anything under 500 g where a prepaid satchel beats a parcel rate. Coverage is the argument. Sendle's network does not reach every remote postcode, and a failed pickup in Broken Hill costs you more in support time than the rate difference ever saved.
On rates, resist the urge to pick one carrier and never revisit it. Compare your own last 500 shipments, not a published table. Export orders with weight and postcode, bucket them by zone, and work out the blended cost per parcel each way. We've seen stores save 12 to 18 per cent on the metro half of their volume by moving it to Sendle while leaving everything west of Ceduna on Australia Post. We've also seen stores where the saving was close to nothing because 70 per cent of their orders were under 500 g satchels and Australia Post already won that tier.
The rules table: zone, weight, service
Here is the logic we start from for an Australian store with one warehouse on the east coast. Treat it as a draft to argue with, not a template to paste.
Metro east coast (2000–2249, 3000–3207, 4000–4179, plus Adelaide and Canberra metro)
- Under 500 g: Australia Post prepaid satchel tier. Flow buys, auto-fulfils, done.
- 500 g to 5 kg: Sendle standard. This is the bulk of most D2C volume and the clearest win from automation.
- Over 5 kg: Australia Post parcel post, and route the fulfilment order to a human if it's over 15 kg because someone needs to check the box dimensions.
Perth and WA metro
- Everything goes to a slower, cheaper service by default and the customer-facing promise says 6 to 10 business days. Do not let Flow buy Express to WA on a free-shipping order. The freight cost on a 3 kg parcel Melbourne to Perth will eat the entire contribution margin on a $90 order.
- Over 3 kg to WA: hold for review. These are the parcels where consolidating two items into one box changes the economics.
Regional, NT and remote postcodes
- Australia Post only. One carrier, no cleverness.
- Add a condition that checks the shipping address against a metafield list of postcodes you've had delivery failures to. If it matches, Flow tags the order and stops. Shipping automation for regional Australia is mostly about knowing which postcodes lie to you about serviceability.
Tasmania
- Treat as regional for transit time even where the rate looks metro. Customers in Hobart are used to it. Customers who were promised three days are not.
Writing the weight conditions without getting bitten
Flow exposes order weight in grams. Every time we've seen a weight rule misfire, it's because someone typed 5 instead of 5000. Build the tiers as explicit ranges with no gaps: 0 to 499, 500 to 4999, 5000 to 14999, 15000 and above. If a product has no weight set in the variant, Shopify treats it as zero and your order drops into the lightest tier, which is how a 4 kg hamper ends up with a 500 g satchel label. Before you turn the workflow on, run a quick check of variants with weight = 0 and fix them. That's an hour of work that prevents a week of refunds.
One more condition worth adding at the top of every label workflow: total line item quantity. If an order has more than, say, six units, send it to a human. Multi-unit orders are where packing decisions change the parcel, and no weight rule can see that.
Auto-fulfil and auto-label are two different switches
People use "auto fulfil orders" and "automate shipping labels" interchangeably and then get confused when tracking emails go out before the parcel is packed. Buying a label creates a fulfilment with tracking, which triggers the shipping confirmation email. If your pick-and-pack happens the next morning, the customer has a tracking number that shows nothing for 18 hours and your support inbox fills up.
Two ways to handle it. Either run the label workflow on a schedule just before the courier pickup, so labels and parcels exist at roughly the same time. Or buy labels early and suppress the notification, then send a single "it's on its way" email once the carrier scans it. We prefer the first because it's simpler and nothing silently breaks. The second is better if your warehouse wants labels printed in a batch at 7am.
Orders Flow should never auto-label
This list is the real value of the whole exercise. Every one of these came from something going wrong.
- Pre-orders and backorders. The single most common failure. If you sell pre-orders via a tag, a product tag, or a line item property, add a first condition: if any line item product has the pre-order tag, exit the workflow. Buying a label for stock that arrives in six weeks means a dead tracking number, an angry customer, and a label refund request you'll forget to lodge. Selling-plan and scheduled-fulfilment orders need the same guard.
- Anything with a risk flag. High-risk orders, and in practice any order where the Afterpay or Zip transaction is still pending capture. Let the payment settle first.
- Orders with a gift note or delivery instruction. Someone needs to read it.
- Orders above a value threshold. Pick a number that reflects your AOV. If a $600 order is going out, the signature-on-delivery decision is worth 30 seconds of human attention.
- International. Customs declarations, HS codes and the low-value imported goods rules mean the paperwork matters more than the label. Automate the New Zealand lane if you ship there constantly. Leave the rest alone.
- Dangerous goods. Aerosols, lithium batteries, anything flammable. Different service, different declaration, not a Flow job.
- Split-location orders until you've tested them properly. One order, two fulfilment orders, two labels, one confused customer.
Three workflows worth building first
- Metro light parcels. Trigger on order paid. Check: not tagged pre-order, risk not high, weight under 5000 g, postcode in your metro list, single location. Buy a Sendle label, tag the order auto-labelled. This alone usually covers half your volume.
- WA and regional guard. Trigger on order paid. If postcode is in the WA or remote list and weight is over 3000 g, add tag freight-review and send a Slack message. No label bought. This one saves money rather than time.
- Peak-season cutoff. From mid-December, any order placed after the last standard cut-off gets tagged, routed to Express, and the label bought immediately so it makes the same-day pickup. Turn it off in January or you'll be paying Express rates in February.
Rough maths on the first one: if you ship 220 orders a day in peak and 55 seconds goes into finding the order, choosing a service and buying the label, that's 220 × 55 = 12,100 seconds, a bit over three and a third hours. Automate 60 per cent of them and you've handed two hours a day back to the warehouse, every day, for six weeks.
What still goes wrong
Rate lookups fail. The carrier returns nothing for an address, Flow can't buy, and the workflow ends without telling anyone unless you added an error branch. Always add the error branch: tag the order, post to a channel, move on. Address quality is the other one. Shopify won't stop a customer typing "Unit 4" into the city field, and a label bought against a broken address is a label you pay to reprint. If a meaningful share of your orders have messy addresses, an address validation step at checkout pays for itself faster than the label automation does.
And if your packing logic genuinely can't be expressed in Flow conditions, say because carton selection depends on SKU combinations, that's a case for a small custom app rather than fifteen branching workflows nobody can maintain. We build those; it is usually cheaper than people assume. Custom Shopify app development is the right tool when the rules engine runs out of room.
Where to start this week
Export your last 500 orders with weight, postcode and shipping charged. Bucket them into the four zones above. You'll find one tier that is both high volume and boringly predictable, and that's the workflow to build first. Leave the rest manual until January.
If you'd rather have someone map the rules against your actual order data before peak, our Australian team does this as part of a store audit, and you can see how we work with local brands on the Australia page.