What is SpamAssassin Score and How to Fix It

The digital age inundated inboxes with emails.

Of the 333 billion emails sent daily, 162 billion are spam emails, making up 49% of the total volume.

Now, email filtering software like SpamAssassin plays a pivotal role in sifting through this deluge, ensuring that only legitimate emails land in our inboxes. For email users, this filtration system is a godsend, as it eliminates potential threats and unwanted clutter from their inboxes. 

However, for us email marketers, this poses a significant challenge. Because even with these handy filters, mistakes can happen. Legit emails might get wrongly flagged as spam, causing them to vanish into the void. 

• A deliverability rate lower than 95% is concerning, often leading to a higher bounce rate. 
• The bounce rate, categorized into hard and soft bounces, shouldn't exceed 3%. 

To avoid this and improve email deliverability rates, you need to keep your SpamAssassin score as low as possible. 

• Do you want to have a grip on the SpamAssassin scoring system? 
• How can you improve your SpamAssassin score? 
• Are you aware of how SpamAssassin works? 
• Have you set up your SpamAssassin score correctly? 

That's precisely what we're here to explore today. Let’s find out:

What is SpamAssassin?

SpamAssassin is a powerhouse when it comes to email deliverability testing.

 It is an impactful email filtering system designed to be your inbox's first line of defense against spam. Developed by the Apache Software Foundation, SpamAssassin has been keeping unwanted emails at bay since 2001.

Integrated with many popular email platforms, it's a go-to tool for ensuring your emails reach their intended recipients. Whether you're using GlockApps, Email Acid, Litmus, or any other deliverability tool, chances are they all lead back to SpamAssassin.

SpamAssassin has become a staple in email deliverability. It's often included in email services like Fastmail, Gmail, and private email servers, giving users the power to customize and create their own spam filters.

This open-source project is all about customization and adaptability. It uses a variety of tests to assign scores to emails, analyzing everything from the content of the email to the sender's IP address. You can even fine-tune SpamAssassin by adding your own rules or training it with your email data. 

How SpamAssassin Works: Decoding SpamAssassin Score

By analyzing emails and assigning a spam score,SpamAssassin helps differentiate between legitimate emails and spam.Understanding how SpamAssassin scores emails is crucial for optimizing your email deliverability.

SpamAssassin employs over 700 tests, including Bayesian filtering, blocklists, and DNS checks, to assess the likelihood that an email is spam. Each test contributes to the overall SpamAssassin score, which helps determine whether an email is delivered to the inbox or flagged as spam.

Here's a breakdown of how SpamAssassin scores emails:

1. Negative numbers: These indicate that the email is unlikely to be spam, keeping your overall score down.
2. Neutral: A score of "0" means that this factor has little impact on the overall spam score.
3. Positive numbers: These suggest a higher probability that the email is spam.

Multiple Rules Triggered
‍

‍When an email passes through SpamAssassin, it undergoes various tests, or "rules," to determine its spam score. These rules can include checking for suspicious keywords, evaluating the email's formatting, and examining the sender's IP address.

Scoring Each Rule

‍Each rule that is triggered by the email contributes a certain score to the overall SpamAssassin score. For example, if an email contains a suspicious keyword, it might trigger a rule that adds 2 points to the score. If the email's formatting is irregular, another rule might add 3 points.

Adding Up Scores

‍SpamAssassin adds up the scores from all triggered rules to calculate the email's overall score. For instance, if the suspicious keyword and formatting rules were triggered, the total score would be 5 (2 + 3).

Threshold Determination

‍The default threshold for classifying an email as spam is 5.0. If the overall score exceeds this threshold, the email is considered spam. However, this threshold can be adjusted based on user preferences or the email service provider's settings.

It's important to find a balance between minimizing false positives (legitimate emails classified as spam) and false negatives (spam emails that go undetected) when setting the threshold.

For instance, if your email scores 3.0 on SpamAssassin, it's less likely to be classified as spam compared to an email with a score of 5.0.

However, different email providers may have varying anti-spam settings, so it's essential to aim for the lowest possible score to ensure your emails reach the inbox.

With every email it processes, SpamAssassin generates a header that includes a list of rules along with their corresponding points. It also indicates whether they contribute positively, negatively, or have no impact on the email's overall score.

Here's an example of a SpamAssassin header for an email:

_{X-Spam-Flag: YES}

_{X-Spam-Score: 7.2}

_{X-Spam-Level: ******}

_{X-Spam-Status: Yes, score=7.2 required=5.0 tests=HTML_MESSAGE,}

 _{RCVD_IN_DNSWL_NONE,SPF_HELO_NONE,SPF_NONE,URIBL_BLOCKED}

    _{autolearn=unavailable autolearn_force=no version=3.4.4}

_‍

In this example, the email has been flagged as spam (X-Spam-Flag: YES) with a SpamAssassin score of 7.2 (X-Spam-Score: 7.2). The X-Spam-Level shows the "spamminess" level using asterisks, with more asterisks indicating a higher score.

The X-Spam-Status provides more details, showing the specific tests that contributed to the score, such as HTML_MESSAGE, RCVD_IN_DNSWL_NONE, SPF_HELO_NONE, SPF_NONE, and URIBL_BLOCKED.

Let's try to understand it further in detail:‍

Rule-based Spam Filters

Rule-based spam filters like SpamAssassin rely on a set of rules to detect spam emails. These rules are essentially patterns or conditions that, when met by an email, indicate that it is likely to be spam. For example, a rule from the SpamAssassin rule set:

_{# Rule for detecting phishing emails}

_{header __PHISHING_URL          eval:check_uridnsbl('PH_SURBL', 'blacklist.example.com')}

_{meta __PHISHING_DETECT         (__PHISHING_URL && !__LOCAL_CONTACT)}

_{score __PHISHING_DETECT        5.0}

_{describe __PHISHING_DETECT     Possible phishing attempt}

‍

In this rule, the header condition checks if the email contains a URL that is listed in a phishing blacklist (blacklist.example.com in this example).

If the URL is found and the email does not contain a local contact (indicating it's not a legitimate communication), the meta rule triggers. This meta rule increases the email's spam score by 5.0 points, indicating a high likelihood of a phishing attempt.

These rules are just a couple of examples of how SpamAssassin uses patterns and conditions to identify spam emails. By combining multiple rules and assigning scores to them, SpamAssassin can effectively categorize emails based on their likelihood of being spam.

Probabilistic Spam Filter

Probabilistic spam filters, like SpamAssassin's Bayesian classifier, enhance the effectiveness of rule-based filters. It uses machine learning to improve spam detection. 

Unlike hard-coded rules, which rely on specific patterns or conditions, the Bayesian classifier learns from labeled examples of spam and non-spam emails to make probabilistic predictions.

• The Bayesian classifier assigns a probability to each email, indicating the likelihood that it is spam. This probability is based on the presence or absence of certain features in the email, such as specific words or patterns. 
• As the classifier is exposed to more examples, it refines its predictions, becoming more accurate over time.

SpamAssassin users can customize the Bayesian classifier by providing it with labeled examples of spam and non-spam emails.

This allows the classifier to learn from the user's specific email patterns and improve its accuracy.

Additionally, users can enable the 'bayes_auto_learn' parameter, which automatically feeds emails that the filter has confidently classified as spam or non-spam back into the classifier for further training.

Custom Rule Set of SpamAssassin 

In addition to the built-in rules, SpamAssassin also allows for the installation of custom rulesets. These third-party rulesets can be easily added by placing them in the sitewide configuration directory. 

While these custom rulesets can be highly effective, it's important to note that they have not undergone the same level of quality assurance as the main ruleset. Therefore, using third-party rulesets is done at your own risk.

Available custom rulesets for SpamAssassin:

• SARE: The SpamAssassin Rules Emporium (SARE) offers a variety of rulesets for different purposes, such as blocking specific types of spam or targeting certain languages or regions.
• KAM Rules: Known as the KAM ruleset, this collection focuses on blocking known spam sources and patterns.
• SpamCop: The SpamCop ruleset integrates with the SpamCop service to block emails from known spam sources.
• URIBL: The URIBL ruleset checks email links against URIBL's list of known spam domains to block emails containing such links.
• SURBL: The SURBL ruleset checks email body and links against SURBL's list of spam domains found in URLs to block emails containing such links.
• RHSBL: The RHSBL ruleset checks the right-hand side of email addresses against RHSBL's list of known spam domains to block emails from these domains.
• DCC: The Distributed Checksum Clearinghouse (DCC) ruleset checks email checksums against the DCC database to detect bulk email.
• Pyzor: The Pyzor ruleset integrates with the Pyzor service to detect and block spam based on checksums.

What are the Common SpamAssassin Tests?

SpamAssassin tests emails for various attributes and patterns associated with spam.

These tests include checking for spammy content, presence on blocklists, proper DKIM and SPF record configuration, suspicious links and attachments, spam-related terms, and disallowed scripts.

‍

Here are some common tests and what they mean:

‍

URIBL_GREY (1.1): Indicates that the URL in the content is present on the URIBL greylist, suggesting potential spammy behavior.

URIBL_BLACK (2.0): Indicates that the URL in the content is present on the URIBL blocklist, indicating a higher likelihood of spam.

URIBL_SBL (0.6): Indicates that the URL is present on the Spamhaus SBL blocklist, which is used to identify spam sources.

SPF_PASS (-0.0): Indicates that the SPF (Sender Policy Framework) check was successful, which is a good sign for email authenticity.

SPF_FAIL (0.9): Indicates that the sender does not match the SPF record, suggesting a potential spoofing attempt.

DKIM_VALID (-0.1): Indicates that the message contains at least one valid DKIM (DomainKeys Identified Mail) signature, which is a positive indicator of email authenticity.

LONG_HEX_URI (2.4): Indicates that at least one file included in the body has a very long, solely hexadecimal URI, which can be a spam characteristic.

HTML_MESSAGE (0.0): Indicates that HTML is included in the message, with negligible impact on the spam score.

AWL (-0.9): Indicates that the "From" address is included on the recipient's white-list, which is a positive indicator for the email's legitimacy.

DRUGS_ERECTILE (2.2): Indicates promotion of drugs related to erectile dysfunction, a common spam topic.

FREEMAIL_FROM (0.0): Indicates sending from a free email account (e.g., Gmail), with negligible impact on the spam score.

MPART_ALT_DIFF_COUNT (1.5): Indicates that the HTML and plain text parts of an email are different, which can be a spam characteristic.

HTML_IMAGE_RATIO_02 (0.8): Indicates a low text-to-image ratio in the body, which can be indicative of spammy content.

These tests are just a few examples of the many factors that SpamAssassin considers when determining the spam score of an email. 

The system also considers the historical engagement of recipients who have received emails from you.

If recipients consistently don't open your emails, it can negatively affect your SpamAssassin score. Regularly cleaning your email list is essential to maintain a positive score. 

Smartlead's Master Inbox feature ensures that email responses are displayed in real-time. 

This means that as soon as a response is received, it instantly appears in the central inbox.This real-time tracking and notification system help you quickly identify and troubleshoot any problems.

5 Proven Tips for Improving Your SpamAssassin Score

We've drilled down a few key tips which can help you improve your SpamAssassin scoring. Let's check them out:

1. Build a Good Sending History

Email filters consider your sender reputation, which is influenced by your sending history. Sending quality emails that recipients engage with positively impacts your reputation.

For example, consistently sending relevant and valuable content to your subscribers can improve your sender reputation.

To enhance your sender's reputation, Smartlead automatically sends warm-up emails from your mailbox to other warm-up users.

• Smartlead's advanced system emulates genuine user behavior, including sending, opening, saving from spam, and replying to warm-up emails. This authenticates the emails, helping to establish a positive reputation for your email account.
• When you activate warm-up, you will receive warm-up emails from unfamiliar users as part of the email warm-up process. These emails will contain a unique "Custom warm-up identifier Tag," identifying them as part of the warm-up process.

2. Use a Reputable IP

Sending emails from a reputable IP address can improve your email deliverability. For successful cold email campaigns, it’s recommended to maintain a domain reputation of 85 or higher.

However, if your domain reputation falls close to or below 70, it’s considered poor and can significantly impact your email deliverability.

For instance, using a dedicated IP address from a reputable ESP can enhance your email deliverability compared to using a free email account. Smartlead.ai provides unique and exclusive IP allocation for every campaign. Also the tool offers:

• Deliverability Test approved IPs only.
• Dynamic IP allocation ensures pre-warmed up IPs or fresh IPs for every campaign.

Moreover,  sending emails to spam traps can severely damage your sender reputation and deliverability. Here's how Smartlead handles spam traps:

• Verification Tools: Smartlead uses two separate verification tools that check the entire pool of email addresses, which currently exceeds 800,000. These tools run multiple times a day to ensure that the email addresses are valid and not associated with spam traps.
  

• Detection and Prevention: If any suspicious activity is detected, such as sending emails to known spam traps or honeypots, Smartlead takes immediate action. It blocks the user account associated with the offending email address to prevent any further issues. 

This proactive approach helps maintain a high sender reputation and ensures that emails are delivered to the intended recipients' inboxes.

3. Authenticating Your Emails

Email authentication, such as SPF and DKIM records, helps verify the legitimacy of your emails. Authenticating your emails reduces the chances of them being flagged as spam.configuring SPF and DKIM records for your domain can improve your email deliverability.

4. Avoid Spammy Words

Be mindful of using words and phrases commonly associated with spam. We have enlisted a few common spammy words and phrases to avoid in your emails:

• Free: While offering something for free can be attractive, this word is often associated with spam emails. Instead, use terms like "complimentary" or "at no cost."
• Urgent: Using "urgent" can create a sense of pressure, which is a tactic often used in spam. Consider using "important" or "time-sensitive" instead.
• Guarantee: Promising a guarantee can trigger spam filters. Use terms like "assure" or "ensure" to convey confidence without using the word "guarantee."
• Act now: Similar to "urgent," this phrase can create a sense of urgency that is common in spam. Instead, use phrases like "limited time offer" or "exclusive deal."
• Buy now: Direct calls to action like "buy now" can be seen as spammy. Try softer CTAs like "explore our products" or "learn more."
• Click here: Phrases that explicitly ask for clicks, like "click here," can trigger spam filters. Use descriptive CTAs that explain the action, such as "view our latest collection."
• Congratulations: While genuine congratulations are fine, this word is often used in spam emails. Use it sparingly and only when appropriate.
• Cash: Phrases related to money, like "cash," can trigger spam filters. Use terms like "funds" or "payment" instead.

Pay attention to text:image ratio

While visual elements can enhance your emails, avoid using too many images. A good practice is to keep the text:image ratio balanced, with text comprising at least 60% of the total message body. This helps maintain a good score and improves deliverability.

5. Technical Fixes

‍Ensure that your HTML code is well-structured and compliant with email coding standards. Avoid using excessive code or complex formatting that could trigger spam filters.

Instead of using <font color="red">Buy Now!</font>, which might trigger spam filters, use <span style="color:red;">Buy Now!</span> for better compliance.

‍ Pay attention to your email headers, including the From, To, and Subject lines. Make sure they are clear, relevant, and not misleading. Avoid using spammy or deceptive headers. Use a concise and relevant subject line like "Exclusive Offer Inside" instead of "You're a Winner!" to avoid misleading headers.

The Bottomline

Mission accomplished! Understanding SpamAssassin scoring can be complex, but once you grasp the system and its measurements, interpreting your results becomes easier.

To avoid an undesirable SpamAssassin score, it's crucial to comprehend the scoring system. If you have any questions about SpamAssassin, feel free to ask in the Smartlead Slack Channel.

Frequently Asked Questions

‍

1. How does SpamAssassin work?

‍SpamAssassin works by using a variety of tests to analyze emails and assign a spam score. These tests include checking for spammy content, presence on blocklists, proper DKIM and SPF record configuration, and more. The system then uses these scores to determine whether an email is delivered to the inbox or flagged as spam.

2. What is a SpamAssassin score?

‍A SpamAssassin score is a numerical value assigned to an email by the SpamAssassin system. This score indicates the likelihood that an email is spam. The higher the score, the more likely it is that the email will be flagged as spam.

3. How can I improve my SpamAssassin score?

‍There are several ways to improve your SpamAssassin score, including building a good sending history, using a reputable IP address, authenticating your emails, avoiding spammy words, and paying attention to the text:image ratio in your emails.

4. What are some common SpamAssassin tests?

‍Some common SpamAssassin tests include checking for spammy content, presence on blocklists, proper DKIM and SPF record configuration, suspicious links and attachments, and more. These tests help determine the likelihood that an email is spam.

5. What is the range of SpamAssassin scores?

‍SpamAssassin scores typically range from -10 to 10, with lower scores indicating a lower likelihood of spam and higher scores indicating a higher likelihood of spam.

6. How can I check my SpamAssassin score?

‍You can check your SpamAssassin score by looking at the email headers of an email. The headers will include a line that indicates the SpamAssassin score, along with other information about the email.

Want to engage in real-time discussions, receive product updates, and gain industry insights related to SpamAssassin Score? Join our community of over 4000 cold email experts.

7. Can I change my SpamAssassin scoring?

‍You cannot directly change your SpamAssassin score, but you can take steps to improve it by following best practices for email deliverability and avoiding behaviors that could trigger spam filters.

8. What is the SpamAssassin threshold score?

‍The default threshold for classifying an email as spam is 5.0. If an email's SpamAssassin score exceeds this threshold, it is considered spam. However, this threshold can be adjusted based on user preferences or email service provider settings.

‍