You send an invoice, close the window and move on. Later the customer says it never arrived. Your Sent folder shows the message. Their inbox does not. Both of you may be describing exactly what you can see; the missing part is what happened between those two screens.

Email does not normally travel straight from your computer to the other person’s computer. Services handle it along the way, rather like postal sorting offices handing over a letter. Understanding those handovers makes a delivery problem much easier to discuss. It also explains why “sent”, “accepted” and “read” are different things. 1

The journey between the screens

First, your email program or invoicing application gives the message to an outgoing mail service. A mail server is simply a computer running software that sends or receives email. Your business can use such a service without owning the computer, just as it can use a delivery company without owning its vehicles. 1

A program is the software you use to do a task: here, write or generate an email. If you use webmail, the browser shows the mail service’s page and passes your actions to that service. If your invoicing application sends notices automatically, it performs the sending step instead of you typing a message. The screen is the place where you make the request; the service is where the next part of the work happens. The precise submission arrangement varies, but it is distinct from the later transfer between mail servers. 1

Consider an invoice notice addressed to [email protected]. The part before @, accounts, identifies a mailbox at that domain. The part after it identifies the domain responsible for that address. The sending service first needs to reach a system handling that domain, rather than find the accounts employee’s laptop. This is why the recipient does not normally have to leave their computer running for the sending service to hand over the email. 1

The outgoing service needs to find the receiving service. In an address such as [email protected], the part after the @ is the domain: example.com. The sending service consults the Internet’s domain directory, called DNS. In the usual route, it looks for the domain’s mail-routing entries, called MX records. These tell it which servers handle incoming mail for that domain. 1

The sending and receiving systems then exchange the message using agreed rules called SMTP. There may be intermediate services, too. If the destination temporarily cannot accept the message, the sending system can keep it in a queue and try again. A permanent failure can lead to a non-delivery notice. Sending therefore involves more than one attempt or one instantaneous movement across the Internet. 1

This exchange has an order. The sending system introduces itself, supplies the delivery return address and names the intended recipient. The receiving system replies at each stage. It can accept the recipient address before the sender has transferred the message itself.

Only after that transfer can the receiver give the final response accepting the complete message. Support therefore needs to distinguish “the recipient address was accepted” from “the whole message was accepted after transfer”. The first response alone cannot confirm the second event. 1

If the receiving service answers that it cannot currently accept the message, the sender may retain it for another attempt. Think of the queue as a list of work waiting to be completed, not a second inbox for the customer. If the recipient address is permanently unacceptable, repeating the same attempt will not repair the address. The returned explanation can tell the administrator whether to wait, correct the address or investigate another problem. 1

Suppose the customer’s mail service accepts our invoice email. That confirms a handover to that service, not that the customer has read the attachment. The message still has to be handled on the receiving side. A record of acceptance can help locate the problem, but it cannot tell us that someone approved the invoice. SMTP transfers mail; it does not record a purchasing decision. 1

Why one message has several addresses

The postal comparison helps with another source of confusion. A paper letter has delivery information on its envelope and a name written inside. Email also separates the information used to transport the message from the information shown as part of the message. 1

The electronic envelope includes the recipient addresses and a return address used for delivery problems. Inside the message, the From field says who the author is. A separate Reply-To field can suggest a different address for replies. Your email program usually shows only a small part of this information at first. The rest is in the message’s headers: the information lines attached to the email. 1 2

For example, a business might use an invoicing service to send mail displaying [email protected]. That service could handle failed deliveries at a different address, while customer replies go to the business’s accounts team. Different addresses are not automatically evidence of fraud. They have different jobs. The important point is that the name you see in the inbox is not the whole delivery record. 1 2

There can also be a display name, such as “Supplier Accounts”, beside the From address. That name is text in the message, not a separate delivery destination. If your program shows only the name, opening the sender details can reveal the address behind it. Reading that address is a useful first step, but reading a claim and verifying the claim are different operations. The domain checks in the next lessons provide additional information. 2

This also explains why checking an email’s origin is not as simple as reading the displayed name. Later checks need to establish which domain a sending service is entitled to use, not just which name somebody wrote in the message. We will examine those checks separately rather than trying to learn every abbreviation at once.

Finding the missing handover

For our missing invoice, “Did you send it?” is only the opening question. The useful next question is “Did the customer’s mail service accept it?” If not, the sender’s provider can look at the delivery attempt and any returned error. If it did, the investigation can move to the receiving side. This follows the route instead of changing unrelated settings and hoping the message appears. 1

Return to our invoice. The invoicing application may have recorded that it handed the notice to its outgoing provider. That provider may then have recorded a later handover to the customer’s provider. These are two records of two different events. Ask support which record it has, and whether there is a later failure notice. This follows responsibility from one service to the next instead of expecting the Sent folder to explain the whole journey. 1

Give support the destination address, approximate sending time and a harmless identifying reference. Ask what stage their record actually confirms. “Our application created the email” and “the destination accepted the email” are not interchangeable answers. Avoid circulating a real invoice more widely than needed just to diagnose its notification.

If the invoice is already available in a familiar business portal, you can check it there through your usual access route while the email problem is investigated. The document and the notice announcing it need not be the same thing. A delayed notice does not, on its own, mean the underlying document disappeared.

An encrypted journey is a separate question

Mail services can encrypt a connection between them: they transform the traffic so someone merely observing its path cannot read the message in its ordinary form. The mechanism commonly used to start this protection is called STARTTLS. But one protected connection does not prove that every stage of a multi-service journey is protected, and it does not protect every stored copy. A padlock on the supplier’s website describes yet another connection. 10

So when a provider says “the email was delivered securely”, there are at least two things to understand: where the message was handed over and how that connection was protected. Neither tells the accounts team whether a new bank account in the message should be trusted. Knowing the email’s route helps us work out what happened to it. Deciding whether to act on its contents is a separate task.

Sources

[1] IETF RFC 5321 — SMTP

[2] IETF RFC 5322 — Internet Message Format

[10] IETF RFC 3207 — STARTTLS