Why Customer Experience is Key to Reducing Churn
Churn is a fact of life for video services. While some services have a lower churn rate than others, all services regularly lose a proportion of subscribers. According to Statista, in August 2024, the net churn rate of SVOD subscribers in the US was around 3.5%, and as reported by Ampere Analysis, 42% of US subscribers regularly subscribe, cancel, and resubscribe. This is a major problem for video services. Faced with the ongoing challenge of reducing churn and increasing retention, video platforms have realized that customer acquisition alone isn’t enough and nor is aggressive pricing. What’s really needed is a better understanding of how users engage with content and what kind of experience they’re having.
Understanding the Customer Journey
There are many reasons why a subscriber may choose to end a video streaming subscription. Some users may feel they don’t get enough value from their subscription because they don’t watch enough content to justify the cost. Other users may have taken out a subscription to watch a particular show, then canceled. And some will choose to walk away from a service because of quality and service issues, which they may or may not have raised with the service provider.
This is a challenge for OTT platforms because it can be difficult to fully understand what an individual customer’s journey looks like. This lack of insight makes it hard to spot the early signs of disengagement, which can result in churn, and without this customer insight, it’s nearly impossible to act before it’s too late.
Churn risk is also highest in the early stages of a subscription, and therefore, early interactions with a service are critical. Yet because most video services still take a largely reactive approach to dealing with issues and potential issues that could result in churn, and many subscribers don’t report problems, a provider may be unaware of problems that users are experiencing.
Increased Focus on Quality of Experience
There’s an increasing realization that a subscriber’s Quality of Experience (QoE) has a direct impact on retention, yet to gain a real insight into that experience, services need more than just aggregate-level data. In the past, most video analytics tools focused on engagement, primarily on what viewers were watching and when, rather than the quality of their viewing experience. While that kind of data is valuable, on its own, it isn’t enough. The focus has gradually changed over time, and there is now more interest in the QoE and also in identifying what actions improve it.
While providers inherently understand that QoE does impact retention, a lot of the data points are separate. They need to understand what causes viewers to leave, at a granular level, and for that, the separate data sets need to be brought closer together. This enables the impact of QoE on retention to be shown and proven. Some of the most important QoE data include:
- Video Startup Time
- Rebuffer Percentage & Buffering Time
- Error Ratio
- Video Bitrate
- Seek Time
- Start Failures and Exits Before Video Start
Although some services do use QoE data, it tends to be at an aggregate level. While understanding rebuffering rates across a global audience can show trends and highlight service-wide issues, it doesn’t offer much clarity when an individual subscriber calls with a specific service or quality issue. Some video services do have more granular session data, which is particularly valuable, but it is often held in silos. Subscriber, video QoE data, and customer care data are often held separately, which is problematic. If service providers want to tackle churn effectively, they need to bring this data together, at which point it also becomes more powerful.
QoE analytics example
A More Proactive Approach to Customer Care is Needed
Video service customer care is mostly reactive in its approach. Typically, customer care teams are either reacting to a ticket where a subscriber has reported an issue or responding to an outreach on social media. This approach is problematic for a number of reasons. Even when experiencing an issue, only a subset of subscribers will actually make contact about said issue in the first place. Others will just feel annoyed and frustrated and then go ahead and cancel their subscription. This type of behavior is commonplace. If providers take a more proactive approach, identifying issues before subscribers get in contact, they can solve problems for more users, which in turn improves the user’s experience and helps reduce the likelihood of churn.
More proactive approaches include upfront testing in real viewer environments and observability to become aware of issues early, before they impact viewers. During the development of a streaming service, often coding environments (IDEs) and device simulators are used to speed up and simplify development. However, viewers are watching on real, physical devices that may behave differently in the real world. It’s important to ensure that the streaming experience is tested on physical devices that are the same as those viewers have in their homes and pockets.
Bitmovin Stream Lab
Given the vast range of devices in the market, this requires automated testing to cover the necessary device range. Further, observability can help streaming services become aware of issues as soon as they occur. When an issue is detected, for example, increased buffering rates on a specific CDN, an alert can be sent in real-time to support teams. This proactive approach enables issues to be solved before they become widespread across the viewer base.
Real-time observability example
Second, even when subscribers do get in touch when they have an issue, support teams often don’t have the right information at their fingertips to understand and visualize the issue. Yet, with the right tools, customer care agents can access session-level data to pinpoint issues for individual users. If a user calls in and the agent can immediately see that their last few sessions had long start-up delays or frequent rebuffering, the conversation instantly becomes more productive. It shows the customer that the provider understands what’s gone wrong, and more importantly, can fix it.
AI can play a key role here to help support teams quickly understand complex viewer sessions. For example, a long viewing session of 1-2 hours may contain hundreds of data points as each change in download speed, quality, buffering events, CDN switches, adverts, and seeks are tracked. Support teams may also be supporting not just video but also app experiences, login/authentication, and payment services alongside video. This means the support agent may not have the specialist skills needed to understand what comprises good and bad video sessions among the large number of data points. AI can help by summarizing the playback session from the viewer’s perspective, alongside comparing it to industry benchmarks. For example, was the startup time good or bad, did the number of buffering events affect the viewer’s enjoyment of the content, and did the viewer’s network contribute to poor performance? The use of AI to interpret viewer sessions helps support agents to respond to a larger number of requests with accurate recommendations for how the session could be improved for the viewer.
Bitmovin AI Session Interpreter
Beyond solving current problems, with the right insights and a proactive approach, video providers can also spot trouble before it starts. If a viewer isn’t watching a lot of content, they’re at higher risk of churn. If a video provider can identify that risk early on, they can take action to address it. Equally, if a subscriber is watching a lot of content, across a lot of different devices, but they’re on a basic subscription plan, maybe there’s an opportunity to offer them an upgrade to a premium tier.
Insights Enable Action
Naturally, video services are interested in any data that helps to predict churn or that gives indicators about potential churn. The most reliable source of data for this purpose is first-party data, which is provided by the viewer, or that is generated directly from viewer behavior. This includes whether or not the customer is active and using the service, and for those that are active, are they experiencing errors that are causing their viewing experience to be interrupted? How long are their streams taking to load, and how often are they reloading the same content? There are a lot of insights that can be gathered from how the viewer is engaging with a service, when they are engaging, and the experience they’re having. It’s common sense that the more of this kind of information that’s made available to customer care representatives, the better they’re able to help the subscriber.
Other kinds of data can also be used to enrich QoE data. For example, for ad-supported streaming subscription services, the data around how users experience and engage with ads is extremely useful. If a viewer is on an ad-supported plan but has a high abandonment rate during ads, that viewer may well be better served by an ad-free tier. Alternatively, it may be that a specific ad campaign is seeing higher abandonment rates or viewers who skip the ad early. Alongside ad engagement insights, it’s also crucial for ad-supported services to monitor ad QoE. If ads fail to load because of errors, this directly reduces revenues for the service. It’s also important for this data to be available in real-time without delay. If ad errors are increasing on a certain ad server, the streaming service may fall back to an alternative ad server to ensure ads continue to be delivered and revenue isn’t lost. This decision must be made in real-time to limit impact.
Ad error percentage
Crucially, these insights shouldn’t be locked away in dashboards for analysts to pore over once a month or quarter. They need to be surfaced in real-time and made accessible to the right people. This is one area where AI will certainly make a big impact by automating analysis, flagging at-risk users, and suggesting actions that service teams can take.
Building Better Experiences
Reducing churn is not a one-time initiative. It is an ongoing commitment to understanding the evolving needs of viewers and responding with thoughtful, integrated solutions. When streaming providers move beyond surface-level metrics and take a deeper look at how quality, support, and engagement intersect, they unlock new opportunities to strengthen loyalty. This starts by connecting siloed data, empowering teams with real-time insights, and using automation and AI to detect issues and recommend clear, actionable steps.
It also means recognizing that every viewer interaction, from account creation to playback quality, shapes the overall perception of the service. Providers who adopt this mindset are better positioned to deliver seamless, personalized, and satisfying experiences that build trust over time. While some churn is unavoidable, a strong customer experience strategy makes it possible to anticipate problems, respond quickly, and retain more viewers in a competitive market.
[Editor's note: This is a contributed article from Bitmovin. Streaming Media accepts vendor bylines based solely on their value to our readers.]
Related Articles
The latest Video Developer Report compiled by Bitmovin shows sustainability slipping down the list of priorities, a casualty of the search for short term profit. The report made clear the industry's top priority is monetizing content more effectively with ad tech, which can be attributed to the huge spike in AVOD and FAST business models.
19 Feb 2025
Moving streaming video services from 'serial churn' to 'pause and resume' in today's subscription economy.
16 Sep 2024
Companies and Suppliers Mentioned