The landscape change is under way

Google’s decision to reduce the User-Agent string under the client hints initiative is having a profound impact on the web ecosystem. Being able to understand web traffic will become more complex and difficult to manage in-house, with impacts in multiple areas as described below, yet many companies are completely unaware that this massive change is happening.

Click here to view all of our technical resources on User-Agent Client Hints.

Contact us to learn more

What are User-Agent Client Hints?

Google’s original Client Hints initiative was a positive contribution to the web landscape: providing non-personal identifiable information about the client which can be easily parsed from headers improves the web experience for everyone. However, Google subsequently introduced User-Agent Client Hints as a justification to remove key information from the User-Agent string. As a reminder, the User-Agent string is the header defined in the HTTP specification to provide information about the client capabilities and limitations on the very first request to the web server.

Ultimately, reducing the User-Agent string compromises the web experience for many, since the ability to understand the nature of a device connecting to a resource on the first request is a vital aspect of keeping the web accessible to all.

What does User-Agent string reduction mean for existing device identification solutions?

Many businesses use open source or homegrown solutions to parse the User-Agent string in order to understand their web traffic and serve their audience appropriately. This enables them to distinguish bot traffic, desktop users & diverse connections such as mobile & IoT devices. Google’s substitution of client hints for the User-Agent string materially increases complexity for implementers of such solutions, just to stay at the same level of functionality. As a consequence, many open-source solutions will no longer be maintained and the cost to resource in-house solutions materially increases. This will result in a poorer understanding of web traffic which in turn leads to compromised end user experiences.

Another issue is the fact that the new client hints will only identify whether a visitor is coming from a mobile or desktop source. However, the reality of the web is that it supports a diverse number of different browsers and device types. It’s a rapidly extending spectrum and one size does not fit all when it comes to serving content.

The Impacts of User-Agent Client Hints

Privacy

The justification for Google to make this change is based on the fact that the User-Agent string compromises privacy. However, the risk is marginal, with no evidence provided of an actual as opposed to theoretical problem. The fingerprinting capabilities are negligible compared to the fingerprinting surface available to Google through the provision of the browser and/or OS for many end users.

The main issue for Google appears to be the fact that the User-Agent string is available to all platforms on an equal basis, ensuring a level playing field for the web.

Performance

The inherent nature of User-Agent client hints means that the capabilities of devices cannot be known on the first request. The likely outcome is a one-size-fits-all approach for serving content.

As connected devices vary considerably in connectivity, screen capabilities and processing power, lower end devices result in a poorer user experience if a ‘one-size-fits-all’ approach is taken.

Accessibility

Users with less capable devices or who are in areas with poor or expensive connectivity, are further disadvantaged if content is not tailored to the capabilities of their device.

Serving heavy images over a narrow pipe to lower performance devices results in elevated bounce rates as a direct consequence. Indirectly, it reduces the access of the less privileged to the web.

Diversity

The diversity of devices accessing the web has exploded in the last ten years. We’re now seeing devices like TVs, kiosk displays, games consoles and in-vehicle systems being used more often.

These devices have different use cases & capabilities, so for the best experience the content served must reflect this. The User-Agent string enables web server awareness of visiting devices, and Google will only identify if a connection is mobile or desktop on the first request. This is, unfortunately, a very simplistic view of web clients. A TV for example has neither a touchscreen or a mouse, so the page design needs to be different to accommodate this.

Monetization

The diversity of the web is enabled by ease of monetization, typically through advertising. However, Google’s changes will impact OpenRTB advertising platforms, increasing complexity and reducing reliability. As a result, Publishers are more likely to use Google Ads rather than use independent advertising platforms.

Reducing monetization diversity sources will also reduce competition, which ultimately harms the ecosystem.

How DeviceAtlas Helps

Despite this confusion caused by Google's changes, DeviceAtlas cleanly handles both the User-Agent client hints headers as well as the traditional standard. We provide a simple interface to process the set of headers available to the web server no matter what is making the connection. DeviceAtlas uses the headers to return details on the connecting device, OS and browser, including: