In 2012, an OpenSignal study found that there were about 4,000 different Android device models on the market. Within a couple of years, that number had risen to 12,000, and it’s likely only gone up since then. As a device tester, you already know that there’s considerable diversity among your customers, and that their needs are going to vary on a case-by-case basis—but who knew there was so much diversity just in the devices themselves?
This degree of fragmentation has more than a few implications for testers. Most importantly, it means that ensuring a high quality of service for customers with a wide variety of different devices is a more Herculean task than ever. Because there are so many mobile devices with which users could be trying to access your network and take calls, run apps, answer emails, etc., service verification is at once more important and more daunting than ever. This begs the question: how can testers approach this changing landscape with an eye toward improving customer experience and ensuring a high quality of service? One potential answer is to ditch manual device testing.
Challenges in Mobile Device Testing
Like we alluded to above, the shear degree of fragmentation within the mobile device market makes device testing a challenge. The 12,000 devices running Android’s OS only make up a fraction (albeit a large fraction) of a market that also includes iPhones, IoT (internet of things) devices, etc., which means that covering a majority of your users with your device testing involves running through thousands of different smartphones and performing functional tests for each one, potentially including 4G, 3G, 2G, and Wi-Fi verification. Considering that the average human tester is able to manually run through about 6-10 use case tests per day, this task can quickly seem impossible.
For many businesses, the solution of choice is to run simulated tests or perform tests on “rooted” or “jailbreaked” devices, rather than utilizing out-of-the-box end user devices. Unfortunately, because these aren’t actually the devices that are being adopted by your customers, you open yourself up to potential gaps in verification. If, for instance, you’re verifying a network update meant to reduce latency times for mobile broadband service, small differences in calibration will have a big impact on customer perceptions. As customer expectations with regards to things like latency and uptime become more and more stringent, the small differences between a simulated device and real one will become magnified. As we enter the 5G era, simulated tests will be even farther removed from reality, making it nearly impossible to verify the granular differences in network quality that many new technologies could come to depend on.
Customer Experience and Network Quality
Okay, so manual testing and simulated testing both present testers with real disadvantages when it comes to service verification in the modern marketplace. But how much should we really care? To what extent do your testing workflows directly impact your customers? After all, if these issues didn’t correspond directly to customer experience, we might be able to skirt some of them a little more easily. Unfortunately, network quality and network verification tend to have a pretty direct impact on customer experience. Why? Because device testing is the only way to uncover the bugs and service gaps that are viewed with increasing scorn by customers whose expectations for network quality seem to grow every year. As those expectations increase, the pressure to test effectively increases along with it.
What this means, essentially, is that running simulated tests, for instance, directly influences customer experience, often for the worse. By the same token, manual device testing—given the tremendous diversity of devices—is often too time consuming to hit desired test coverage levels, meaning that it, too, has the potential to negatively impact customer experience. By committing to a testing framework that you know can’t achieve high levels of coverage, you are essentially committing to leaving many potential gaps in service quality unexamined, boosting the likelihood that your average user will encounter service failure or bugs trying to access your network with their device of choice.
Enter Device Testing Automation
This, perhaps unsurprisingly, is where automated testing comes in. Here, it’s crucial to draw a distinction between simulated tests and automated tests: simulated tests on virtual devices leave open the possibility that, when the rubber hits the road, your users will encounter issues that you weren’t able to replicate in your test lab; automated testing on out-of-the-box devices, on the other hand, alleviates this issue. By reproducing something much closer to the actual conditions under which the devices will be used, you greatly increase the likelihood that you’ll find the bugs that are most relevant to your users. By the same token, because automation can empower you to run hundreds of tests per day instead of only a handful, your testing coverage suddenly increases significantly. Rather than focusing on only the most common devices, you can work to verify service on the latest technology as it emerges, potentially resulting in a competitive advantage in the eyes of early adopters.
In this way, automated testing begins to have a direct impact on customer experience. Rather than hoping that you’re able to cover the most common use cases and devices with manual testing, you’re able to strive for much more complete coverage—leaving no stone unturned, so to speak. Thus, the odds that a given customer will pick up the latest Samsung smartphone only to find that VoIP simply refuses to work with your service decreases considerably. This has a potential cascade effect, improving customer satisfaction and thus bolstering your customer retention rates. After all, the more effectively you can verify service across a wide variety of devices the less likely users are to jump because of network issues. Beyond that, automation makes continuous regression testing feasible (when time constraints might otherwise make it a hard thing to prioritize), meaning that as new devices enter the market you can continue ensure that they fit into your service offerings. Trust us, your customers will thank you in the long (and short) run.