Email Data-Science Posts

The old Mailchimp blog had several classic articles about email. These articles were data driven, and many were written by the fantastic John Foreman (whose book and talk I recommend).

These articles were scrubbed from the Internet. So here they are again, for private reference:

John Foreman
A Purchased List is a Dead List
Jul 24, 2014 MailChimp prohibits using purchased, rented, or scraped email lists, which are now illegal in Canada under CASL and violate their acceptable use policy. Such lists perform poorly, leading to low engagement and high complaint rates.

AOL and Hotmail Users Spend More than Gmail Users, and Other Research Finds
Jan 8, 2013 MailChimp’s 2012 research highlights reveal that Hotmail users tend to spend more than Gmail users, likely due to an older demographic with more disposable income. Additionally, daily deals emails have low engagement but also low unsubscribe and abuse rates, while purchased email lists result in significantly higher complaint rates, leading to potential ISP blocks.

Comacast and Gmai: all your typo email are belong to us
Dec 20, 2011 Typos in email addresses, especially in domains, can lead to successful delivery to typosquatting sites, which may result in security risks or misuse of your content. It’s crucial to watch for these errors and consider using double opt-in to ensure list quality.

Competing with the World Cup for Subscribers’ Attention
Jul 9, 2014 The 2014 World Cup significantly impacted email marketing engagement, with noticeable drops in activity during games and peaks before and after matches. Marketers should consider their subscribers’ time zones and avoid sending important campaigns during major events to maintain engagement.

Detecting Click Clusters in Your Data Using Spreadsheet Math
May 15, 2013 Most people segment their email lists based on predefined criteria, but there’s potential in allowing the computer to analyze data and suggest segments based on subscriber clicks. By employing data mining and unsupervised learning, we can uncover new insights and optimize our email marketing strategies.

Digging Deeper into Wavelength and EGP Data: Finding Interest Clusters in MailChimp’s Network
Oct 23, 2012 MailChimp’s Wavelength tool analyzes subscriber data to identify interest clusters among users, enhancing targeted marketing efforts. By employing cosine similarity calculations, it reveals how closely related different subscribers are, uncovering unique audience segments and insights for improved engagement strategies.

Effect of Social Networks on Email Engagement
Feb 22, 2011 The article discusses how social networks can enhance email engagement, emphasizing the importance of diverse content sharing on platforms like Twitter and Facebook. Research indicates that incorporating social elements in emails can lead to increased engagement, suggesting that marketers can effectively blend email and social strategies for better results.

Fun with Data Science
Dec 7, 2011 The article introduces John, a data scientist at MailChimp, who analyzes email campaign performance through geographic distribution using advanced algorithms like clustering and dendrograms. His findings reveal significant insights on list performance, such as the correlation between subscriber locations and bounce rates, as well as the benefits of utilizing newer email domains for better engagement and deliverability.

How Gmail Tabs Affect Your Email Marketing: An Update
Dec 12, 2013 The article discusses the impact of Gmail’s tabbed inbox on email marketing, revealing that while open rates have decreased significantly for Gmail users, the system appears to help users manage their inboxes effectively, as unsubscribe rates remain stable. Additionally, it highlights variations across industries, suggesting that marketers may need to adjust strategies to engage subscribers better.

Inactive Subscribers are Still Valuable Customers
Feb 9, 2016 The article emphasizes that inactive subscribers, while not actively engaging with emails, still hold significant value for businesses. They contribute to revenue, purchase more frequently, and are less likely to churn compared to non-subscribers, suggesting that businesses should focus on re-engagement strategies rather than removing them from email lists.

Insights from MailChimp’s Send Time Optimization System
Jul 14, 2014 The article discusses MailChimp’s Send Time Optimization (STO) feature, which recommends optimal sending times for email campaigns based on individual engagement data. It highlights key insights such as weekdays being preferable for sending, a peak time of 10 AM, and the importance of understanding the audience’s characteristics, like age and occupation, for effective email timing.

Introducing Hairball v2: Hairball with a Vengeance
Apr 9, 2012 The article introduces Hairball v2, an updated version of the Adobe Air application designed for complex list segmentation in MailChimp, allowing users to perform advanced analysis of campaign performance and recipient engagement. Key features include improved segmentation based on historical data, performance comparison, and more precise filtering of click data to enhance targeted marketing efforts.

Introducing Wavelength
Jan 18, 2012 Introducing “Wavelength,” a MailChimp service designed to help users find similar publishers by analyzing their email lists. It emphasizes the importance of building organic, permission-based email lists instead of purchasing them and outlines how Wavelength can connect users with relevant newsletters based on subscriber overlap.

MailChimp’s Email Genome Project
Feb 9, 2011 The MailChimp Email Genome Project aims to analyze user data to improve email marketing effectiveness and deliverability. By examining patterns from millions of subscribers and emails sent daily, MailChimp seeks to address common dilemmas in email marketing and prevent abuse of the email ecosystem.

Multivariate Testing for MailChimp Pro
Sep 17, 2015 The article discusses MailChimp’s new Multivariate Testing feature, which allows users to test up to eight different elements of an email simultaneously to enhance customer engagement and marketing strategies. It emphasizes the importance of data-driven decisions and provides insights into effective testing practices and reporting metrics to optimize email campaigns.

Project Omnivore: Declassified
Jan 27, 2010 MailChimp’s Project Omnivore, initiated in late 2008, aimed to create a scalable tool to detect spam, particularly from well-meaning marketers unaware they’re spamming. By utilizing advanced genetic optimization techniques and massive data analysis, Omnivore assesses email campaigns and user behavior, sending warnings or suspending accounts for suspicious activities, while still relying on human review for final decisions.

Project Omnivore: Three Years of Gorging on Data
Nov 13, 2012 This article discusses Project Omnivore, which MailChimp launched to monitor and analyze email list quality and sender behavior. By utilizing an AI model that evaluates email addresses based on extensive historical data, Omnivore helps ensure a healthier email ecosystem while allowing legitimate users to send emails without unnecessary barriers.

Research: Spammy Email Design Mistakes
Oct 13, 2010 The article discusses an experiment by MailChimp that analyzed email campaigns flagged as spam by reviewers from Amazon’s Mechanical Turk. It identifies common design mistakes that lead to false positives, such as overly sales-oriented text, poor layout, and reliance on stock templates, ultimately providing insights into effective email marketing design.

Send Time Optimization in MailChimp
Dec 3, 2013 This HTML document outlines the “Send Time Optimization” feature introduced by MailChimp, which helps users identify the best times to send emails based on subscriber engagement data. The optimization takes into account global email engagement trends and specific subscriber history to improve click-through rates, particularly focusing on the most engaged audience segments.

Sending Frequency: More Is Not Always Better!
Apr 23, 2013 This article discusses the misconception that sending more emails leads to greater engagement and revenue. It emphasizes that increased email frequency negatively correlates with engagement per campaign, suggesting an optimal sending frequency to maximize total clicks rather than simply increasing volume.

Smarter Bounce Management Rules with Engagement
Sep 1, 2010 The article discusses the challenges of managing email bounces for Email Service Providers (ESPs), emphasizing the distinction between hard and soft bounces and the complications that arise from server errors and deceptive bounce messages. It highlights the importance of effective bounce management for maintaining domain reputation and outlines MailChimp’s strategy of incorporating engagement metrics into bounce cleaning decisions to improve deliverability and list management.

Subject Line Data: Choose Your Words Wisely
Nov 13, 2013 This article analyzes the impact of subject line wording on email open rates, revealing that even a single word can significantly affect engagement. Key findings include the effectiveness of personalization, urgency, and certain keywords, highlighting the importance of choosing words wisely in email campaigns.

Subscriber Engagement Half-Life
Feb 18, 2011 The article discusses findings from Dan Zarrella’s webinar on email marketing, revealing that the average click rate for subscribers significantly drops below 1% after four months. To maintain engagement, marketers must entertain and deliver value quickly while continually attracting new subscribers to offset potential churn.

Subscriber Engagement Over Time
Nov 25, 2013 The article analyzes subscriber engagement over time using extensive data, finding that click rates start above 5% and decline to around 3.5% after multiple campaigns, with differences observed across industries and email domains. It emphasizes the importance of subscriber context and preferences in shaping engagement, recommending strategies like double opt-in to enhance initial interaction levels and maintain interest.

This Just In: Subject Line Length Means Absolutely Nothing
Sep 17, 2012 The article explores the misconception that the length of email subject lines significantly affects open and click rates. Analyzing data from 12 billion emails, it concludes that there is no meaningful correlation between subject line length and engagement, emphasizing the importance of personalized content over gimmicks or strict length guidelines.

Update on Omnivore, new 3 Strikes Rule
Aug 27, 2010 The article provides an update on MailChimp’s Omnivore program, which is designed to combat email abuse by analyzing user behavior and implementing anti-spam measures. It discusses the new “Three Strikes” rule, stating that if any sender is suspended three times within six months, their account will be permanently shut down, aiming to enhance the overall integrity of the email ecosystem.

What Does Your ISP Say About You?
Nov 26, 2013 This article explores how email addresses can reveal insights about individuals, including their likely age and interests based on the email domain used. It analyzes various domains such as Gmail, Hotmail, Yahoo, AOL, and Comcast, showing how different demographic characteristics and preferences can be inferred from users of these services.

What Good Marketers Can Learn from Spammers
Oct 18, 2012 The article discusses what effective marketers can learn from spam emails, specifically highlighting that while spammers often use a single clear call to action and few distractions, marketers should balance word count, image count, and URL count to optimize click rates. The research indicates that campaigns with concise content and relevant links tend to perform better than those with excessive text or images, ultimately suggesting that quality and clarity are key to successful email marketing.

Where Have All the Government Clicks Gone?
Oct 15, 2013 The article discusses the impact of the 2013 U.S. government shutdown on email engagement among government employees, noting a significant drop in interactions from affected agencies. It highlights how the furloughs led to decreased activity, contrasting it with unaffected agencies like the SEC, which saw increased engagement during the fall marketing season.


If you loved this post you’ll superlove my monthly emails ✉️