Every WooCommerce shop that ships through a third-party warehouse has, somewhere in a drawer (or in a PDF nobody opened), a data processing agreement with that warehouse. Or it should. And almost none of those shops have looked at what their fulfilment connector actually sends, or can read, compared to what that agreement says.
I did look. That is more or less why Wareferry exists.
Disclaimer
I am not a lawyer. I am a developer who has read the GDPR more times than is healthy, and this is my reading of it as applied to a very specific problem: getting orders from WooCommerce to a warehouse. If your shop has real legal exposure, talk to someone who does this for a living. What follows is engineering, not legal advice.
Two articles that matter here
Article 5.1.c: data minimisation
The principle is short: personal data shall be “adequate, relevant and limited to what is necessary in relation to the purposes for which they are processed”. Not “what might be useful someday”. Not “what is easier to send in one go”. What is necessary for the purpose.
And the purpose, in fulfilment, is about as concrete as a purpose can get: put these items in a box and get the box to this person.
Article 28: the processor
When a logistics operator receives your customers’ names and addresses to ship your orders, it is processing personal data on your behalf. In GDPR language, you are the controller and the warehouse is a processor. Article 28 says you should only use processors that provide sufficient guarantees, that the relationship needs a contract, and that the processor acts on your documented instructions.
Read that again with a fulfilment connector in mind. “On your documented instructions.” Your instruction is “ship this order”. Not “have a look around the store and take whatever you need”.
Where the pull connector gets awkward
The typical WooCommerce fulfilment integration works by pull. You create a REST API key (often read/write, often on orders and customers), paste it in the operator’s portal, and their system polls your shop for new orders.
The problem is not what they do with that key. I’m happy to assume good faith. The problem is what the key allows. A WooCommerce API key is scoped by resource, not by order or by date. A key that can read today’s orders can read the orders from 2019, the cancelled ones, the ones from customers who asked you to delete their account and whose orders you kept for tax reasons. All of it is one request away.
So the processor ends up with standing access to far more personal data than the purpose needs. Maybe it never uses it. But minimisation is about what you make available, not about what you hope nobody will touch. And if that operator’s systems are breached, “they had read access to the whole store” is not a sentence you want to write in a breach notification…
What a warehouse really needs
Writing one plugin per operator means reading a lot of order APIs and checking which fields are mandatory. The answer is reassuringly boring. To ship one order, a warehouse needs:
- The recipient’s name.
- The shipping address.
- A phone number, when the carrier requires one (some do for certain countries or services; many don’t).
- An email, only if the carrier sends delivery notifications and you want them.
- The order lines: SKU and quantity.
- Your order reference, so the tracking can come back to the right place.
That’s it. No billing address, no order history, no customer notes, no list of other customers, no payment method. In other words, the minimal set is small enough to fit on the label, because, well, most of it ends up on the label.
How I approached it in Wareferry
Wareferry is a family of WooCommerce plugins, one per operator (Huboo, Bigblue, Quivo and the rest). The design choice that matters here is push instead of pull. Your shop sends each order to the warehouse, with the fields above and nothing else. The operator never receives your WooCommerce keys and never gets admin access. The connection uses your own account with the warehouse, so the credentials flow the other way: you authenticate to them.
The nice consequence is that minimisation stops being a promise in a contract and becomes a property of the system. The warehouse can’t read what was never sent. There is a longer explanation in how it works.
The request log, masked
There is one more place where personal data likes to hide: logs. Any integration worth using keeps a record of what it sent and what came back, because when an order fails at three in the morning you need to know why. But a log that stores full names, addresses and phone numbers in plain text is just a second customer database, with worse access control.
So the Wareferry request log masks personal fields. You can see that the request went out, which order it was, which SKUs, what the operator answered and what error it returned. You don’t see the customer’s phone number sitting there for anyone with access to the logs. Enough to debug, not enough to leak.
Honest status
Wareferry is in development. Nothing is on sale and there are no prices yet. I’m not going to claim it has been running in production for years, and I’m not going to claim it makes you “GDPR compliant”, because no plugin does that on its own. What it does is remove one big, unnecessary chunk of exposure from a process that most shops never looked at twice.
If your data processing agreement says one thing and your connector does another, you may want to join the waitlist. I’ll write when it’s ready.
