“Sent” should refer to evidence, not just a green message
If a contact form confirms submission but no email arrives, do not reinstall the form or change suppliers blindly. Trace one test from the page to the mailbox and identify the last stage that can prove it received the enquiry: browser, application, mail service or recipient system. Commission the repair at that break rather than replacing the whole chain.
To begin, provide the form URL, exact time and time zone, intended destination, confirmation wording and a test reference that contains no sensitive information. Do not send passwords, API keys or unrestricted mailbox access. A specialist can request properly limited access through an appropriate channel if the investigation genuinely requires it.
Worked example: a quotation request at 11:42
Consider a fictional window and door company. Someone completes its quotation form at 11:42 on 31 October, includes “TEST-1142” in the message and sees “Thank you, we have received your request”. Nothing reaches the sales address. Rather than submit ten more forms, the manager records the page, time, browser, expected destination and a screenshot of the result.
The web supplier checks the application and finds a submission with that time and reference. The content was stored, but the notification event was rejected before reaching the mail service because the recipient setting was empty in the live environment. The green message proved that the form had been stored, not that a notification had left the application.
The fix assigns the intended recipient, retains a record of each enquiry and makes the confirmation issue a reference code. A second test appears in the record, is accepted by the mail provider and reaches the agreed mailbox. The failure path is checked too: the visitor is not promised success when the application already knows that storage or notification failed.
This is not a real client or a universal cause. With no application record, the break would be earlier. If the mail provider showed acceptance by the recipient server, the investigation would continue in the mailbox, its rules or quarantine.
Use one test record instead of diagnosing by instinct
Create a test with recognisable but non-confidential data. Use a message such as “TEST-1142”, an authorised reply address and a clearly fictional telephone number if that field is required. Record the exact URL, date, time and time zone, device, browser, displayed response and intended mailbox.
MDN describes an HTML form as a way to configure an HTTP request that sends data from the browser to a server address; the server must then receive, validate and process it. That separation matters. Pressing the button, passing browser validation and receiving a response do not by themselves prove that an email was created and delivered.
Ask the supplier to find this test by time and reference, not to expose unrestricted logs. If the application stores enquiries, confirm that the record exists and who can recover it. If it stores none, request at least a limited interaction identifier and outcome. OWASP explains that application logs can add context unavailable in infrastructure logs and support business-process monitoring and investigation of unusual conditions.
Read the last available evidence, then continue from there
At the browser stage, confirm that the test received a real response rather than displaying a local animation. In the application, look for receipt, validation, storage and the notification attempt. At the mail provider, find the sending identifier and status. At the destination, use the exact time and sender to review receipt, spam, quarantine, rules and aliases.
Acceptance is not the same as final delivery. WordPress documents the distinction explicitly: a successful `wp_mail()` result means that the method processed the request without errors, not that the recipient received the email. Even when the site is not built on WordPress, the business question remains useful: what exactly does the status in front of us confirm?
The evidence should form a chain. If the application received nothing, the spam folder is not the place to start. If the mail service rejected the message, changing form colours or fields will not repair it. If the recipient server accepted it, reinstalling the form consumes time and may introduce another fault without explaining the first.
Compare a repair with two more resilient options
A minimal repair restores the broken point: recipient, service credential, configuration or error handling. It fits when the site already keeps a reliable submission record and the failure is isolated. Request before-and-after evidence and a description of the change without secrets appearing in the report.
An operational improvement stores each enquiry in an authorised panel or system and treats email as an alert rather than the only copy. This helps where a missing request has commercial impact or several recipients are involved. Decide who can view the data, how long it remains and how it is deleted; indefinite storage is not a neutral improvement.
Periodic monitoring can run controlled tests or alert on accumulated failures. It may suit a critical form, but it does not replace diagnosis or guarantee that every future message will arrive. It should distinguish a page failure, provider rejection and recipient filtering so the alert leads to a specific action.
What the completed repair should hand back
The handover should state the identified cause, applied change, reference for a successful follow-up test, verified destination and visible behaviour when an error occurs. Record who administers the sending service and mailbox, where events can be checked and which data is retained. A screenshot of the green confirmation alone is not sufficient.
You can check spam, rules, addresses and a simple test internally when you control both site and mailbox. Professional help is worthwhile when there is no record, site and email belong to different suppliers, rejections or credentials are involved, attachments or personal data are handled, or the failure is intermittent. Limit access to the system under investigation and do not exchange credentials through ordinary email or chat.
A useful next step is a precise supplier note: “On this URL, test TEST-1142 at 11:42 Europe/Madrid on 31 October showed confirmation, should have reached [email protected], and is absent from inbox, spam and quarantine.” That gives the supplier evidence to find. “The form does not work” makes everyone start again.
Sources and documentation
Content created with AI assistance. Check the sources, date and scope of each guide.