The advertising industry has undergone tremendous change over the past few decades. In the early 2000s, display advertising dominated the digital advertising landscape, but as technology advanced, so did consumer behavior. Banner ads, once the golden standard, have been traded in for more engaging video ads and native ads.
Mobile devices emerged as the dominant medium for online activity, leading to the rise of mobile advertising and new ad formats like rewarded video. Social media platforms like Facebook and Twitter exploded in popularity, giving advertisers new channels to reach audiences. At every twist and turn, our resilient industry found opportunities and developed technology to meet the moment.
For years, consumers willingly offered up their personal information, trading their privacy, some knowingly, some unknowingly, in return for personalization and free content. Now, the pendulum has swung in a different direction, and consumers are focused on maintaining their privacy as much as possible.
As we've done in the past, the AdTech industry is taking steps to reflect this new reality. In the next few weeks, Google will implement one of its most disruptive changes in recent years – the User-Agent (UA) Reduction.
In this article, we'll look at how Chrome User-Agent and Client Hints work, the UA string reduction timeline, and what measures MGID has taken to keep up with the changes and ensure advertisers will be able to reach target audiences, optimize campaigns, and maximize their ROI.
What is User-Agent?
User-Agent (UA) is a piece of software that acts on behalf of a user, typically a web browser, and sends requests to servers for resources such as web pages. The UA string, which is a piece of text, is included in an HTTP request header and provides critical information about the user agent, including the browser type, version number, operating system, device type, and screen resolution.
The data can be used for different purposes. If the UA identifies a user browsing from a mobile device, the server can return a mobile-optimized version of the webpage which is easier to read and navigate on a smaller screen. Website owners and publishers can use the data to analyze traffic and optimize their sites for their audience's most common types of devices and browsers. User-Agent data can also be used for security purposes, such as identifying and blocking suspicious or malicious traffic.
Advertisers can use it to tailor their ads to the user's device, screen size, or show an advertisement for a product available in the user's country. Google primarily uses UA data for optimization and personalization purposes for its own products and for the websites and content that it serves to users.
What is the User-Agent Reduction?
Before we get into the 'what,' let's start with the 'why' of it all. Over the past few years, consumers have become acutely aware of how trackers and cookies are used to track their online behavior and collect thousands of pieces of data about them - data that is then sold to third parties, including advertisers and data brokers. While that data was being used to deliver on the personalization promise, it was also being used to predict, influence and manipulate users' decision-making processes without the user's consent or knowledge.
Governments around the world heard consumer privacy concerns loud and clear. They responded with rules and regulations like Europe's GDPR, California Consumer Privacy Act (CCPA), and China's Personal Information Protection Law (PIPL) designed to protect their citizens' privacy by obtaining active consent.
Although every business must follow the laws for where they operate and where their consumers reside, some big tech companies have chosen to take additional steps to limit collectible data and further protect their users. Apple made a splash and plenty of headlines with its 2021 rollout of iOS14.5, which included its App Tracking Transparency (ATT) feature. ATT is a user-facing feature that gave iPhone and iPad users the ability to allow (opt-in) or deny each app's ability to track them across other companies' apps and websites.
Google has been trying to address privacy-related concerns from several different angles. In January 2020, Google first announced it would phase out third-party cookies. Although the death of the third-party cookie has been delayed multiple times, Google has now said it would start the deprecation process in the second half of 2024.
Also, in 2020, Google publicized its intention to reduce the amount of default data shared through the User-Agent string. From a privacy standpoint, User-Agent data presents two issues. First, end users have no choice in what data is shared and no option to consent or opt-out. Second, the extensive data transmitted in the UA string can be used to passively fingerprint and identify users. From a technical standpoint, the excessive length of the UA string has been known to cause servers to make errors when parsing the string. To eliminate all three problems with one solution, Google came up with User-Agent Reduction and Client Hints for Chrome.
As the name suggests, User-Agent Reduction reduces the amount of information passed in the User-Agent header to improve user privacy and reduce the tracking of users across the web. User-Agent Reduction is already underway and has been rolled out in phases. The first few phases, which covered Chrome 95 – 100, were dedicated to awareness and trials. When Chrome 101 was released in April 2022 (phase 4), the minor version numbers were replaced with zeros, so Chrome/101.3.1.1 appeared as Chrome/101.0.0. In phase 5, which was tied to the release of Chrome 107 in October 2022, the desktop UA string was reduced, with both the desktop OS version and CPU info replaced by fixed values.
In February 2023, Chrome 110 was released, moving us into phase 6, which reduced the Android mobile and tablet UA string so it contained fixed values for the version and device model. In May, we will reach the final phase. With the release of Chrome 113, the reduced UA string will be applied to all page loads on Windows, macOS, Linux, Chrome OS, and Chrome on Android.
What are Client Hints?
Google isn't going to leave developers, publishers, and advertisers high and dry. Although most businesses should be fine operating with the default values provided by the reduced UA, those that need complete UA data for anti-fraud functionality, to render device-specific content, or businesses like supply-side platforms (SSPs) can access it through Google Chrome's User-Agent Client Hints (UA-CH) API via JavaScript or HTTP headers. The default information sent with UA-CH includes the browser and its major version, the operating system, and whether the browser is being used on a mobile device.
UA-CH provides access to the same data that was available in UA strings before reduction, but the key is that it's done with a more privacy-centric approach. Essentially, this stops the default action of sending all data unless it's explicitly requested.
How do Client Hints work?
Client Hints, or Google Hints, provide a standardized set of HTTP headers that contain information about the user agent. When a web server receives a request that includes Chrome User-Agent Client Hints, it uses the information in the headers to tailor its response to the user agent. For example, the server can determine the device type and screen size and send content optimized for that device, helping to improve performance and the user experience without compromising user privacy or allowing for tracking and profiling based on User-Agent details.
Types of Client Hints
There are two types of client hints: low entropy and high entropy.
Low entropy Client Hints, provided by default, share limited information that still allows web servers to optimize content for specific user agents without revealing identifying details. Data would include the software name and its version, such as Chrome 112, and the operating system, such as Mac OS X.
High entropy Client Hints, shared by explicit request, provide more detailed information about the user agent, such as the exact version of the browser and operating system. Instead of sharing the version as Chrome 112, the data would include the full software version (112.0.5615.138), full operating system (10.14.6), and complete device information (Mac OS X 10_14_6).
How MGID Prepared for the Google Update
Times are tough. Consumers are thinking about the unstable economy and rising inflation. Advertisers are being asked to reach the same or higher KPIs even as their ad budgets are slashed; making every dollar count is more important than ever for performance marketers.
The switch from User-Agent Reduction to User-Agent Client Hints could have dramatically impacted our advertisers. However, as a leading AdTech platform, we prepared for the upcoming changes in advance. Had we not, this update could have affected our ability to define device type (mobile versus desktop), define Android OS version and desktop OS and define the device model. In turn, that would have led to inaccurate and inefficient targeting and wasted budgets.
Being able to precisely target users by device, operating system, and phone price are non-negotiables for many advertisers. These parameters, among others, allow advertisers to optimize their ad spend, reach their target audience more effectively, and increase the likelihood of driving conversions.
We have already started to support Client Hints to minimize these adverse outcomes for direct and programmatic advertising.
How MGID uses Client Hints
Browser and OS support for Client Hints and UA string reduction can vary widely and greatly depends on what is available for each specific platform. Some browsers and operating systems have implemented support, while others may have limited support or none at all.
At MGID, we will follow this order:
- If the browser doesn't support Client Hints, we'll look at UA, even if it's reduced.
If the browser does support Client Hints:
- If we get just low entropy (default) hints and UA is complete, we'll use UA only.
- If we get just low entropy hints and UA is reduced, we'll use Client Hints.
- If we get high entropy (on request) hints, we'll use Client Hints.
- If we get high entropy (on request) hints and have full UA, we'll use Client Hints.
- If we see that UA is not reduced, we won't ask about high entropy hints because we assume they will include the same data as UA.
In the face of this significant change, we have good news. As an advertiser using MGID, you don't need to do or change anything. We've taken care of all the details for you within our platform. However, if you use an ad tracking platform, you will need to confirm it has taken the appropriate steps to prepare for and support the adoption of Client Hints because not every solution has.
In Which Cases Are Data Discrepancies Possible, and Why?
We don't always receive Client Hints from all our traffic sources. Sometimes we get reduced UA strings, which means we won't be able to identify phone price ranges or OS versions. You might see these discrepancies in your tracking solutions, or once a user clicks on your ad.
We understand that this might be frustrating, but it is what every ad platform will face once this update is live.
The New Privacy-Centric Era
Consumers and governments have spoken: it's time to embrace and respect user privacy. As we navigate these new waters, we should be prepared to be flexible and adaptable. Chrome's User-Agent Client Hints is an early solution with great potential. By providing limited information about the user agent, UA-CH can help to protect user privacy and prevent tracking and passive fingerprinting while still delivering optimized web content, ads, and a good user experience.
However, client hints are not without their limitations, particularly for advertisers who will need to be cautious about how they use this data to avoid infringing on users' privacy. It's worth noting that the extra call for UA-CH will result in more latency, including for ad delivery. Not all operating systems and browsers support UA-CH, but given Chrome's dominance and market share, if it's successful, more will adopt it, and it will become an industry standard. In the unlikely event this giant experiment doesn't work, it will certainly give rise to even more innovative solutions.
Ultimately, the success of the advertising industry will depend on its ability to strike a balance between effective targeting and responsible data collection. Advertisers must be mindful of users' privacy concerns and work to build trust through transparency and accountability. By doing so, we can create a more sustainable and ethical advertising ecosystem that benefits everyone involved.
