Device sunsetting: When the numbers aren’t enough.

Supporting an app across a wide range of devices means eventually having to decide when some of those devices have reached the end of the road. Older platforms can continue consuming engineering and testing resources long after most viewers have moved elsewhere, while newer devices compete for the same teams’ attention. At some point, continuing support may become difficult to justify.

For many media companies, it can be tempting to treat that as a numbers-driven decision. Device analytics can show how many people still use a platform and how quickly usage is declining, while engineering teams can estimate the burden of keeping it running. But those numbers only capture part of the decision. Knowing when to sunset a device also requires understanding the audience and business context behind the data.

What happens after support ends?

Consider two streaming services that each get 5% of their viewing from older Samsung Tizen TVs. That 5% might look identical on an analytics dashboard, but the business implications could be very different. In both cases, that might look like a strong argument for keeping support in place, especially if losing the entire group would have a meaningful impact on the business. But the percentage doesn’t tell you how many of those viewers would actually disappear if the app stopped working on that device. Some may move to another screen without much trouble, while others could find the change frustrating enough that they watch less often or leave altogether.

That uncertainty is difficult to resolve through device analytics alone. You can measure how many people currently use a television model, but predicting how readily those viewers will change their habits requires more context about the audience behind the number. That’s why Mark McKenna, Director of Client Engineering at REDspace, sees analytics as the beginning of the decision rather than the end. “Start with the math,” he explains. “But follow it up with a realistic understanding of what might not be visible in the numbers.”

Think about a retailer considering whether to close a store that accounts for 10% of company sales. That location may generate 10% of revenue today, but closing it wouldn’t necessarily erase all of those sales. Some customers would shop at another branch or move online, while others might disappear if the alternatives are inconvenient enough. The useful number isn’t simply the share of sales flowing through that store. It’s how much business the company would actually lose once the doors close.

Media companies face the same kind of uncertainty when they remove support for a device. Usage data can show where viewers are watching today, but the business impact depends on how those viewers respond once that option disappears. The numbers can inform the decision, but they can’t make it for you.

The same percentage can carry very different risks.

Go back to our two broadcasters example with 5% of viewing on an older generation of smart TVs. The first primarily serves younger viewers who regularly move between televisions and mobile devices, with many households already using several ways to access the service. Losing the television app may create some inconvenience, but a large portion of those viewers could migrate without much difficulty. The second service has an older audience that tends to keep televisions longer and has relied on the same app for years. Asking those viewers to buy another device or learn a different way to access the service could create much more friction. Both companies see 5% on the dashboard, yet the potential audience loss behind that number could be very different.

Demographics, however, are only one source of context. A device with relatively little usage might still play an important role because of a distribution agreement or another commercial relationship. Elsewhere, natural churn could be doing some of the work for you as households gradually replace older televisions and move to newer platforms. Additionally, the burden of maintaining the device also affects the decision, since an aging platform can require an outsized amount of development or testing compared with the audience it serves. Those resources have an opportunity cost as teams respond to new platforms and changing device requirements. As a result, leaving support in place indefinitely can carry its own business consequences, particularly when legacy work begins crowding out higher-priority development.

This is where sunsetting becomes a judgment call rather than a threshold you can set once and reuse across every platform. A company could decide that 5% usage is acceptable on one device because those viewers are difficult to migrate and the platform requires relatively little maintenance. Another platform with the same usage could become a stronger candidate for sunsetting if support consumes substantial engineering capacity and most viewers have easy alternatives. The percentage still provides valuable information, but its significance depends on the circumstances surrounding it.

Questions to ask before pulling support.

Analytics should still be one of the first places teams look when evaluating an aging device. Usage trends can show whether an audience is shrinking, while engineering teams can develop a clearer picture of the resources required to keep the platform running. But before making the final decision, several questions can help expose the assumptions sitting behind those numbers:

• Who is behind the usage number? Look beyond the percentage to the audience using the device, including demographics and how comfortable those viewers are changing the way they watch. A relatively small audience can create a larger retention risk if moving to another device introduces significant friction.
• Is the audience already migrating naturally? Older televisions eventually get replaced, and broader changes in the device market can gradually reduce reliance on a legacy platform. Waiting may sometimes lower the audience risk, although continued support during that period still consumes resources.
• Are there commercial reasons to maintain support? Distribution relationships or other corporate agreements can give a device importance that isn’t obvious from its share of viewing. Those commitments need to sit alongside the usage numbers when the company evaluates the platform.
• What does continued support require from the organization? Development effort and testing demands can vary significantly between devices, especially as platforms age. Teams also need to consider what other work could receive those resources if support ends.
• What are those viewers likely to do next? Consider whether affected households already have another practical way to watch and how difficult the transition would be. The existence of another supported device doesn’t guarantee that every viewer will use it.

Where the numbers stop.

At the end of the day, the hardest number in a device sunsetting decision may be the one you can’t measure directly. Analytics can tell you how many viewers use a platform today, but they can’t tell you with certainty how many will stay with you once that platform disappears. That future behavior has to be estimated from the audience, the available alternatives, and the business context around the device. The data can narrow the decision, but some uncertainty will always remain. Good judgment begins with recognizing exactly where the numbers stop.