The same message looks credible on a phone and obviously wrong on a computer. That is not a matter of the reader's judgement but of the two clients showing different amounts of the same header — and the shorter display leaves out precisely the part that matters.
The sender line has two halves
The From: header consists of a freely chosen name and the actual address. The name is plain text with no rules attached: it can say whatever it likes and need have nothing to do with the address behind it.
Why the authentication checks say nothing here
This is where many explanations fall short. SPF, DKIM and DMARC check the domain of the address. They do not check the display name — it is text in the message header, nothing more.
A message from k8x2q@sender-4471.example carrying a bank's display name can therefore pass all three checks, and entirely legitimately: the domain sender-4471.example belongs to the sender, their server is authorised, their signature is valid. The result is a green tick on a message pretending to be something else.
It is the same blind spot as with domains that look indistinguishable from a real one: authentication says "this domain really did send it". It says nothing about whether it is the domain you have in mind.
The sharpened version
Because the display name is arbitrary text, an address can sit there too:
From: "service@northbank.example" <k8x2q@sender-4471.example>
On the phone the sender line then shows an address that looks right. It is only the name, though. Anyone who has learned to "check the address" is caught at this point by exactly the habit that was supposed to protect them.
Two amplifiers come on top:
- The contact picture. If the display name matches a saved contact, some clients show that contact's photo. The message then does not merely look right, it looks familiar.
Reply-To. A header of its own that decides where a reply goes. It can point at a third address, and it is usually displayed only once the reply has already been written.
What helps, in order of effort
Tap the name. In every mail client on a phone, tapping the sender opens the full address. That is one second and the only action that makes the missing half visible.
Read from the right. An address belongs to whatever sits immediately before the last dot. In northbank.example.customer-service.test the domain is customer-service.test. Everything in front is freely chosen and is there for exactly that reason.
Reverse the order when money is involved. Do not check the message and then act; open the matter independently: launch the bank's app, look at the order history in your account, ring the number printed on the card. If the thing is real, it is visible there too. If it is not, the check was unnecessary.
Open it on a computer when something feels off. The larger screen shows more — the address, the destination of links, often a note about an external sender.
For organisations
Two measures work better here than any training:
- Mark external senders. A note the mail server adds to every message from outside makes the contradiction immediately visible: a display name from your own management next to "from outside the organisation" needs no further knowledge to spot.
- Protect internal names. Some systems can quarantine messages whose display name matches a person in the directory while the address does not. That is the only check that looks at the display name at all.
Everything else is the familiar set of markers listed elsewhere. The display name is simply the part of it that small screens systematically favour.