Switching shipping software without losing orders comes down to one rule: nothing gets canceled until the new tool has shipped real orders for a full week alongside the old one. Export your shipment history, address book, and open-order state first; run both tools in parallel while the new one earns trust; resolve unshipped labels and pending refunds on the old platform before the final cutover. Done in that sequence, a switch is a quiet week of overlap — not a weekend of missing tracking numbers.

What should you export before you cancel anything?

The day you decide to switch is an export day, not a cancel day. Once a subscription lapses, some platforms curtail access to your own history — so pull everything while your login still works:

Shipment history, as far back as the platform allows. You want tracking numbers, dates, services, and costs — this is your evidence base for any future buyer dispute, tax question, or carrier claim about an old shipment.
Your address book: saved recipients, sender addresses, and any warehouse or return addresses configured in settings.
Package presets — the box sizes and default weights you have refined over hundreds of shipments. They look trivial until you re-create them from memory and get three wrong.
Open state: a list of every purchased-but-unshipped label and every refund/void still pending. This list becomes your mid-switch worksheet in the section below.
Integration inventory: which stores are connected, which automations or rules exist, and any API keys in use by your own scripts.

Save it all as CSVs somewhere boring and durable. Most of it imports cleanly into the next tool — GoatLabels takes address and order CSVs through an import flow with column mapping and per-row validation, and saved mapping presets mean the second import is a two-minute job. History that will not import still earns its keep as an archive.

How do you run two tools in parallel for a week?

Parallel running is the step people skip, and it is the step that makes the switch boring — the goal of this entire checklist. The pattern:

Days 1–2, shadow mode. Set up the new platform and quote the same orders in both tools without buying from the new one yet. You are checking that addresses validate, rates make sense, and your package presets translate. On GoatLabels this is free to do exhaustively — quoting costs nothing and there is no subscription meter running while you evaluate.
Days 3–5, split live traffic. Buy real labels for a portion of orders on the new tool — start with your simplest domestic parcels, then add an international order to exercise the customs flow. Everything else stays on the old tool. You now have production evidence instead of a demo impression.
Days 6–7, invert the split. New tool takes everything; the old tool sits idle but alive, as your fallback. If nothing forces you back for two full days of normal volume, the migration has passed its trial.

Day	Task	What proves it worked
1	Export history, addresses, presets, open-order list from the old tool	CSVs open cleanly; row counts match what the old dashboard shows
2	Import addresses into the new tool; recreate package presets	Spot-check ten imported addresses against the source; presets quote sensibly
3	Buy first live labels on the new tool (simple domestic orders)	Labels print, scan at drop-off, and tracking events flow
4	Route one international and one edge-case order through	Customs details accepted; the awkward address validates
5	Connect store syncs on the new platform	New orders appear in the new tool without manual entry
6–7	All volume on the new tool; old tool idle	Two days of normal shipping with zero fallbacks to the old tool

A note on cost during overlap: a week of parallel running means one extra week of the old subscription, which stings on principle. Pay it anyway — it is the cheapest insurance in this whole process, and if the old bill is the motivation for leaving, the specific exit guides for ShipStation, Shippo, and Stamps.com cover the timing quirks of each platform's billing cycle.

What happens to unshipped labels and pending refunds mid-switch?

That open-state list from your export day is now your worksheet. Every purchased-but-unshipped label on the old platform gets one of two fates:

Ship it from the old platform. If the parcel is genuinely going out, use the label you already paid for. Do not re-buy it on the new tool; you would then have two live labels for one box and a void to chase.
Void it on the old platform. If the order was canceled or you would rather re-buy on the new tool, void the label now, while you still have full account access. Carriers set an outside window for refunding unused labels, and a void requested through a platform you have already left is a support ticket you do not want to write.

Pending refunds deserve the same urgency in reverse: refunds already in flight on the old platform should land before you close the account, because they typically return to the payment method or account balance on that platform. Keep the old account open — even if idle — until the last pending refund resolves.

This is also where platform payment models diverge in a way worth knowing before you pick a destination: on GoatLabels, a void returns to your prepaid wallet as instantly spendable credit, so mid-operation voids never strand money in a refund pipeline. If refund friction is part of why you are leaving, the Pirate Ship switch guide walks through that difference in detail.

How do you re-point store connections without missing orders?

Store syncs are the one place a gap can genuinely lose an order, so sequence it deliberately:

Connect the store to the new platform first, before disconnecting the old one. GoatLabels syncs Shopify, Etsy, WooCommerce, and TikTok Shop; for channels without a connector — DMs, invoices, marketplaces like Mercari or Whatnot — orders come in through Billy by paste or screenshot instead, so inventory that intake path too.
Let both platforms see orders for a day. Duplicate visibility is harmless; you simply fulfill from one. A missed order is not harmless.
Mark the new platform as the fulfillment source of truth, and confirm that tracking numbers it generates flow back to the store and out to buyers.
Disconnect the old platform's sync only after a full day of orders has demonstrably arrived in the new one and tracking has round-tripped.

One TikTok Shop caveat to carry into any migration: TikTok requires USPS labels for TikTok orders to go through TikTok Shipping — that is platform policy, and no third-party tool replaces it. A synced tool covers your other carriers where allowed and every other sales channel.

When is it actually safe to cancel the old subscription?

Cancel when every item on this list is true — and not before:

Two consecutive days of full volume shipped on the new platform, no fallbacks.
Zero purchased-but-unshipped labels remaining on the old platform.
Every pending void and refund on the old platform has resolved and the money has landed.
Store connections re-pointed, with tracking confirmed flowing back to buyers.
Exports archived — history, addresses, presets — somewhere that outlives the account.

Then cancel, calendar a reminder to confirm the final bill is the final bill, and keep the export archive forever. The whole process is seven to ten days of mild overlap, and the drama it prevents is precisely the drama you never get to see.

FAQ

Will I lose my shipment history when I cancel? You lose access on the old platform's terms, which is why the export comes first. Archive tracking numbers, dates, services, and costs as CSVs before canceling — disputes and tax questions about old shipments outlive subscriptions.

Do I have to re-enter my address book by hand? No. Export it as a CSV and import it into the new tool. GoatLabels maps your columns with per-row validation on import and saves the mapping as a preset, so subsequent imports take minutes.

What about labels I bought but have not shipped yet? Ship them from the old platform or void them there before you leave — never re-buy the same shipment on the new tool while an old label is live. Carriers set an outside window on unused-label refunds, so sweep these early in switch week.

How long should the parallel period last? A week of real volume is the sweet spot: long enough to hit an international order, an edge-case address, and a normal Monday spike; short enough that double-running costs stay trivial.

My sales channel has no connector on the new platform — is that a dealbreaker? Not necessarily. On GoatLabels, channels without a sync (DM sales, Whatnot, Mercari, invoices) run through Billy — paste the order text or a screenshot and review the drafted label — so the checklist above still applies, with Billy standing in for the connector.