How do email analytics platforms classify email client usage, handle web proxies, and differentiate mobile from desktop views?

Key findings

• Web proxy impact: Gmail and Yahoo (along with AOL and Verizon Mail) utilize web proxies to cache images, which largely anonymizes the user agent string, making it difficult for analytics platforms to determine the specific client or device.
• User agent string reliance: Classification of email client usage primarily depends on analyzing the user agent string when an email's tracking pixel is opened. This string contains details about the software and operating system.
• Mobile vs. Desktop ambiguity: Due to image caching, opens through web proxies mix mobile, webmail, and other platform opens, making it challenging to get an accurate percentage of mobile vs. desktop users.
• Limited granularity: Identifying specific desktop client details, such as Outlook 2013 vs. Outlook 2013 120 DPI, is often not possible if the user agent string does not contain that specific level of detail.
• Occasional bypasses: Some platforms or browser versions, like Edge 12 historically, might occasionally override caching, allowing the full user agent string to be captured for Gmail opens, providing temporary visibility.

Key considerations

• Data accuracy: Recognize that reported email client usage, particularly for major providers, will have limitations due to image caching and web proxies. This can affect the accuracy of open rate measurements.
• Design implications: Despite data limitations, a significant portion of opens still occur on mobile devices. Designing responsive emails that adapt well to various screen sizes remains crucial, as highlighted by Mailchimp's research.
• Alternative tracking methods: Consider employing advanced tracking techniques, such as loading different pixels at CSS breakpoints, to infer device type, especially for iOS users behind proxies, although this method is not foolproof.
• Focus on engagement: While client data is useful, focusing on broader engagement metrics like clicks and conversions may provide more actionable insights, as these are less affected by image caching, allowing you to understand recipient behavior more accurately.

Key opinions

• Desire for detail: Marketers frequently seek granular data to understand their audience's preferred viewing environment, such as specific Outlook versions or the exact percentage of mobile opens.
• Proxy data confusion: The "web proxy" classification in analytics tools (like Return Path) is a source of confusion because it blends mobile, webmail, and other desktop opens, making it impossible to separate them.
• Impact on optimization: Without clear mobile/desktop breakdowns, marketers find it challenging to justify or prioritize responsive design efforts and tailor content effectively for different devices, impacting overall email deliverability.
• Workaround exploration: Some marketers are exploring technical workarounds, such as using CSS breakpoints with different tracking pixels, to indirectly infer mobile versus desktop usage behind proxies, particularly for iOS users.
• Mobile traffic dominance: Despite data hurdles, there's a general understanding that mobile traffic continues to grow, pushing the need for mobile-first email strategies regardless of precise analytics.

Key considerations

• Adapting strategies: Marketers should focus on creating highly responsive emails that provide a good user experience across all devices, even when precise client data is unavailable.
• Holistic engagement view: Consider combining pixel-based open data with other metrics, such as click-through rates and website analytics, to build a more comprehensive picture of recipient behavior and how ISPs track engagement.
• Understanding proxy behavior: Accept that web proxies are a permanent part of the email landscape for major ISPs and design strategies around their impact on data visibility.
• Leveraging available data: Even with proxy limitations, the available email client data can still offer valuable insights into non-proxied opens and overall trends.

Key opinions

• Proxy as an umbrella: Web proxies (like those used by Gmail and Verizon Media Group) act as a catch-all, grouping various user environments, including mobile and webmail, under a single category due to their image caching practices.
• User agent string limitations: The precision of client classification is directly tied to the completeness of the user agent string. If a proxy intercepts and strips this information, detailed reporting is impossible.
• Slippage occurs: Occasionally, certain platforms or browsers (e.g., Edge 12 in the past) might not fully adhere to caching, allowing some full user agent strings to 'slip through' and provide more detailed client data for a limited time.
• Granular data challenge: Identifying highly specific client attributes, such as Outlook 2013 vs. Outlook 2013 120 DPI, is generally not supported by standard user agent strings and, therefore, cannot be accurately reported by analytics pixels.
• Imperfect mobile differentiation: While some methods exist to infer mobile vs. desktop use behind proxies (like CSS breakpoints for iOS), these are acknowledged as imperfect solutions that do not provide definitive answers regarding how opens are generated.

Key considerations

• Proxy as a default: Assume that a significant portion of opens from major providers will fall under a generic proxy category, and adjust expectations for granular client data accordingly.
• Focus on overall trends: Instead of obsessing over precise numbers for specific clients, focus on broader trends in mobile vs. desktop adoption based on unproxied data and industry reports.
• User agent string analysis: For analytics platforms that provide raw subscriber-level data, understanding how to parse and interpret user agent strings can still yield valuable insights for non-proxied opens.
• Industry best practices: Adhere to webmail forensic guidelines and robust email design principles that ensure optimal rendering across a wide range of email clients, rather than relying on perfect data.

Key findings

• Client-server architecture: Email systems operate on a client-server model, where various email clients (desktop, webmail, mobile apps) communicate with mail servers to retrieve and display content, as detailed by technical documentation.
• User agent string purpose: User agent strings are designed to provide server-side applications with information about the user's software environment, including the operating system, browser, or email client.
• Proxy function: Web proxies serve as intermediaries, fetching content (like images) on behalf of the user. In the context of email, this often involves caching images, which can replace the original client's user agent string with the proxy's own.
• Privacy and performance: Image caching through proxies is implemented by ISPs (Internet Service Providers) for both performance benefits (faster loading, reduced server load) and privacy enhancements (masking user IP addresses and specific client details from senders).

Key considerations

• Standardized protocols: While email relies on protocols like SMTP (Simple Mail Transfer Protocol), client-side interactions for rendering (like image requests) fall under HTTP/HTTPS, where user agent strings are typically included.
• Email client diversity: Different email clients, including various mobile apps (e.g., Gmail app on iOS vs. Android), may have distinct ways of handling image requests, sometimes bypassing or modifying proxy behavior.
• Evolving proxies: The behavior of web proxies and image caching is not static; ISPs can update their systems, which might lead to temporary shifts in data visibility, as seen with some browser versions occasionally providing full user agent strings for Gmail opens.
• Data limitations for privacy: Recognize that the reduced granularity in client data is often a deliberate trade-off by ISPs to enhance user privacy, aligning with broader internet trends. This impacts how email deliverability rates are calculated.