When sending an email campaign or a simple transactional message to an SFR address, the email may end up in the spam folder or even disappear without notification. The problem does not always stem from the content of the message. At SFR, spam filtering combines sender reputation, technical authentication, and content analysis, with thresholds that have tightened in recent years.
SPF, DKIM, and DMARC Authentication: The Technical Prerequisite That SFR Checks First
Even before analyzing the body of the email, receiving servers check if the sender’s domain is properly authenticated. Three protocols come into play: SPF (which declares which servers are allowed to send on behalf of the domain), DKIM (which cryptographically signs each message), and DMARC (which tells the receiving server what to do if SPF or DKIM fails).
Since February 2024, Google and Yahoo have imposed these three protocols on bulk email senders. French internet service providers, including SFR, are gradually aligning their filtering criteria with this new standard. A domain without a published DMARC record in its DNS zone will see its messages classified as spam or silently rejected.
You can understand how SFR’s spam filter works by consulting specialized resources, but the starting point always remains the DNS configuration. A poorly formed SPF (syntax with too many lookups, “all” mechanism instead of “~all” or “-all”) is enough to trigger a rejection.
Checking Your Configuration in Practice
Use a tool like MXToolbox or your email service provider’s diagnostics to ensure that the SPF record includes the servers being used. For DKIM, the public key must be accessible in a dedicated DNS subdomain. DMARC, at a minimum, should be in “none” mode with a reporting address to start collecting data on authentication failures.
Without these three records in place, any other optimization is useless. Technical filtering precedes content analysis.
Sender Reputation and Complaint Rates: The Signals SFR Monitors
A technically authenticated email can still end up in spam. SFR’s servers take into account the reputation of the sending IP address and the sender’s domain. This reputation builds over several weeks and depends on cumulative factors.
- The spam complaint rate: each time an SFR recipient marks a message as unwanted, it degrades the sender’s domain reputation. The threshold not to be exceeded is generally below a few tenths of a percent of the messages sent.
- The bounce rate: regularly sending to non-existent or disabled SFR addresses signals poor maintenance of the contact list, which filters interpret as spammer behavior.
- The sudden sending volume: going from a few hundred emails per day to several thousand without a gradual ramp-up triggers alerts on the server side.
Returns on this point vary depending on periods and IPs, but it is observed that SFR applies stricter restrictions to new sending IPs. Starting with small volumes and gradually increasing remains the most reliable strategy for building a positive reputation.
Cleaning Your SFR Contact List
Systematically remove addresses that generate hard bounces after the first send. Inactive addresses for several months deserve a re-engagement campaign: if they do not respond, remove them. This list hygiene has a direct impact on deliverability to SFR inboxes.
Email Content and SFR Filtering Triggers
Once authentication is validated and the reputation is correct, the content of the message is scrutinized. SFR’s spam filter analyzes several elements of the email, and some triggers are less obvious than one might think.
The text-to-image ratio matters. An email composed solely of a large image with little or no HTML text will be penalized. Filters cannot read text embedded in an image, which makes the message suspicious by default.
Links in the body of the email are examined. A blacklisted redirect domain, a URL shortener (like bit.ly), or a link to a site without an SSL certificate can be enough to trigger spam classification. Each URL should be checked with a reputation tool before integration.
Keywords Are No Longer the Only Criterion
The approach of avoiding terms like “free” or “promotion” remains relevant, but it has become secondary. Modern filters analyze the overall structure of the message: poorly formed headers, broken HTML code, absence of a plain text version, hidden tags. A technically clean email performs better than an email without “forbidden” words but poorly coded.
- Always include a plain text version (multipart/alternative) in addition to the HTML version.
- Avoid HTML code generated by a word processor (Word, Google Docs) that inserts unnecessary tags.
- Place a visible and functional unsubscribe link at the top or bottom of the email, in accordance with requirements that have been generalizing since 2024.
- Test each campaign with a spam scoring tool (Mail Tester, GlockApps) before sending to SFR addresses.
Configuring SFR Webmail on the Recipient Side to Retrieve Legitimate Emails
The problem sometimes lies on the recipient’s side. An SFR contact who does not receive our emails can take action from their webmail interface to correct the situation.
In SFR webmail, the “Spam” folder stores filtered messages for a limited time. Moving an email from the spam folder to the inbox sends a positive signal to the filter: the sender is reclassified as legitimate for future sends.
They can also add the sending address to their contacts or SFR address book. This action creates an implicit whitelist that bypasses automatic filtering for that specific sender.
Asking Your Recipients to Act
In a welcome email or confirmation of registration, add a line asking the recipient to add the sending address to their contacts. This simple practice significantly reduces the spam classification rate on subsequent sends.
Deliverability to SFR inboxes relies on a two-level effort: technical configuration on the sender’s side (DNS, reputation, content) and actions on the recipient’s side (whitelisting, reporting false positives). Neglecting either level is enough to compromise the arrival of emails in the inbox.
