Connected TV (CTV) has opened up a world of opportunity for advertisers, publishers, and streaming platforms. According to Nielsen, streaming currently represents 43.8% of overall TV consumption in the U.S. Rather than switching on the family TV, consumers are now streaming content through apps on a wide range of devices like Smart TVs, game consoles (e.g. PlayStation), streaming sticks (e.g. Roku) and set-top boxes (e.g. Apple TV). CTV means any of those connected devices that stream TV content via the internet instead of traditional cable or satellite.
Traditional TV manufacturers and set-top box vendors were built for broadcast and cable TV. However, CTVs are connected devices with apps that send HTTP requests in browsers and serve ads programmatically. The problem streaming providers and advertisers face is figuring out what devices their customers are using. This information is crucial for identifying and connecting with valuable audiences.
CTV offers an incredible opportunity for advertisers and streaming platforms to personalize user experiences and target customers effectively. StackAdapt reported that CTV advertising spending is forecasted to reach a massive $46.89 billion USD in 2028. Thus, we pose the question ‘What devices are your visitors using and how do you know?’
The CTV Blindspot
One of the biggest blind spots in digital advertising is CTV identification. As mentioned, a CTV can now be a Smart TV connected to the internet unlike a traditional TV. The difference: when a customer streams a show on a Smart TV, the browser provides a User-Agent string. Within this UA string, it is possible to identify the device type.
TV manufacturers and app developers often come from a broadcast background, where the concept of a UA string simply didn’t exist. So when prompted for one, they might insert random text like the app name with a version number or leave it blank altogether. Some TV brands send empty or unstructured UA strings, making it almost impossible to properly identify the device. This lack of CTV knowledge results in poor monetization, weak analytics, inaccurate targeting, and unreliable reporting.
If streaming providers are only using an open-source UAParser or an in-house solution, it is most likely that they do not know what devices are being used. In fact, most UAParsers openly admit that they don’t detect set-top boxes at all!
Why Traditional Detection Fails
Many streaming platforms haven’t fully absorbed what it means to operate in an HTTP environment. Most data libraries cannot tell the devices customers are using and if they do, the results can be extremely inaccurate. Even if businesses use tools like a UAParser, it can fail to detect many devices, including Apple TV or Roku, recognizing them as a "Smart TV".
Without accurate device detection, businesses face many consequences:
1. Poor User Experiences: Content and ads may not be optimized for specific devices, leading to a frustrating experience (e.g. serving an ad for a large TV on a small Samsung phone).
2. Inaccurate Ad Targeting: It becomes difficult to serve the right ads, tailor CTAs for different devices, or even identify household brands (e.g. Apple vs. Android). This impacts the ability to guide users to the correct app store, ultimately leading to lower revenue and a poor user experience.
3. Weak Reporting: Data on user behavior patterns and ad performance is unreliable and inaccurate.
4. Loss of Potential Revenue: Inaccurate device data hinders advertisers ability to optimize ad campaigns. Ads sent to the wrong screen are ineffective.
As more traffic shifts to CTV devices, these problems will worsen if left unaddressed.
How to Accurately Identify CTV Traffic
In-house solutions and open-source UAParser tools can miss thousands of connected TV devices. These solutions don't provide defined models but instead identify just the manufacturer and device type. They only return models directly from the user agent header, which aren’t standardized across manufacturers and can lead to inconsistencies. Thus, for streaming providers and advertisers, it is near to impossible to know what devices their customers are really using.
In contrast, a device intelligence solution like DeviceAtlas can identify the Primary Hardware Type for a vast range of devices. It has a comprehensive database of 108,000 devices.
Below is a table showing the CTV devices identified by DeviceAtlas. It can detect 8,723 TVs, 3,980 set-top boxes, 232 projectors and 80 game consoles. Combined, that is 13,015 different types of CTV devices that could potentially be missed without a device intelligence solution.
Experts at DeviceAtlas tested a UAParser and discovered that it miscategorized the following devices:
This means that for some clients, a UAParser can miscategorize almost 70% of their set-top box traffic.
Therefore, caution is required whilst using open-source libraries to classify CTV traffic. The risk of inaccuracies is very high. Trusted industry leaders like Netflix and CBS understand the value of accurate device detection and rely on DeviceAtlas's data for their operations. Knowing exactly what device is requesting an ad greatly boosts customer satisfaction and reduces subscription churn.
Successful Case Study: Adoppler
Adoppler, a full stack advertising technology platform offering OTT and video monetization solutions, faced challenges with inaccurate device detection in the fragmented CTV device landscape. Their in-house solution, built on open-source libraries failed to deliver sufficient accuracy and device insights. Adoppler noted that in many cases those libraries determined devices incorrectly or didn’t provide information on CTV devices at all.
To overcome these challenges, Adoppler partnered with DeviceAtlas for accurate device intelligence. After implementing DeviceAtlas, Adoppler immediately saw a 110% increase in CTV identification. It improved ad targeting, monetization, and partner trust while giving Adoppler more time to focus on its core business instead. You can learn more by reading our case study.
Conclusion
CTV is rapidly becoming the dominant way to consume media and the future of advertising. Streaming providers and advertisers need to understand the importance of knowing what devices their customers use. Investing in a device intelligence solution can solve the CTV identification problem by revealing the correct user devices, leading to more successful ad targeting, better user experiences, and improved revenue.