Mobile backend-as-a-service (MBaaS) provider Kii has launched a first-of-its-kind A/B testing feature for MBaaS customers, which can also be used as a stand-alone tool for mobile app developers. Kii CEO and Cofounder Masanari Arai and Senior Vice President and Chief Product Officer Phani Pandrangi spoke with ProgrammableWeb about what the new features mean for mobile app developers and how to make use of their REST APIs and SDKs to better analyze the commercial potential, usage patterns, and growth opportunities of new app products.
This month, Kii has launched Kii A/B Testing as part of the platform and as a stand-alone service. It is built for iOS, Android, and Unity apps and enables developers to run simple or complex experiments. Developers can integrate the service into their app product development using REST APIs or SDKs.
The four layers of MBaaS product development
“What MBaaS is providing is a tech layer that can develop mobile applications without server-side coding,” Arai told ProgrammableWeb. “So, the technology is there to develop an application very quickly, and you don’t have to operate the server. So scalability and everything are provided by us.”
Arai sees this as the primary layer of product development services that MBaaS can offer their customers.
“But if you look at mobile app success rates, only a very small proportion of the products developed are building an audience and generating an income. So our business [as an MBaaS provider] will not really be sustainable without customers developing successful apps. So we see that the next layer that is required is around end-user acquisition and optimization.
“For end-user acquisition, if you develop a mobile app using MBaaS, you need to understand the end user. So from that viewpoint, we provide analytics as one function. But there is also the need to optimize the end-user experience. Now we are offering A/B Testing so our app developers can better understand what their users need. So when you come up with applications, you have much more probability to succeed.”
Arai then sees two more layers of successful product development sitting on top of these two foundations: Once an app is generating end-user customers and providing a pleasurable user experience, an app developer can monetize their product. Arai sees this as a connected but separate layer to user acquisition and optimization. Finally, he adds distribution as the top layer in his model.
“We also do a distribution service. For example, we have a Chinese distribution service. So what we are doing is we take that application, we do the translation service and the testing, and we have partnerships with big distribution markets in China. We can connect with the local payment system, and we can also plug into the local advertising network,” said Arai.
A/B Testing for mobile app developers
Pandrangi explains why A/B testing is crucial to an effective app product development user acquisition and optimization strategy:
“We see A/B testing as part of our second layer, that is, in the optimization experience, but it could also be in the monetization layer, like if you are testing certain price points and things like that. So instead of codifying app components like colors, price points, or layouts, a developer can use our A/B Testing framework and externalize these functionalities and make changes by actually testing different things and making decisions based on actual data rather than gut feel.
“Thus, instead of a color being green, you can say in the code of the application that you want an experiment to decide. ‘For 95% of users, I want to use green -- which has been the color I have always been using -- but for 5% I want to test purple.’ After a while, you can see in the developer portal when it becomes statistically significant, and then you can choose that you want to use purple for all users.”
Integrated with MBaaS or a stand-alone service
Pandrangi urges mobile app developers to consider using Kii A/B Testing as a stand-alone service, even if they do not need a full MBaaS provider.
“You can use our A/B Testing as stand-alone. So there are customers who are starting with a very new app, and it takes them a while to build a user base. And we have customers who may already have an app in the productivity space. We have one customer who already has 75 million users, and the company is using our MBaaS to add more back-end services.
“You can imagine for game developers, in particular, that this is a very important feature. Games developers tend to focus a lot more on analytics than any other developer.
“A lot of developers who are in a position to think about A/B testing are typically people who start thinking about this after they have built an app or a game and then they start thinking about these things. So there are a lot of customers who will just want to use A/B testing at that stage. When you have something you want to test in the app, you simply externalize that, and, to make that happen, you simply put in the SDKs and then you have access to this experimental construct,” Pandrangi said.
Pandrangi and Arai indicate that unlike other analytics services attached to app usage, Kii Analytics features allow deep insights.
“The beauty of our Kii Analytics is that it really lets you customize your end-user app data. What we have focused on is the specifics about your app or your game. For example, as a developer, you would want to collect specific analytics suited to your game; or if you have a productivity app, you might want to see analytics about the locations where people are taking notes. So if you want to segment on free versus pro, or location, you can set up your analytics that way.”
Pandrangi and Arai point to one customer who used A/B testing to discover that price point was not a deciding factor in user conversion rates, letting the app developers scoop up up a bigger pile of the money on the table.
“Price point is a typical test to conduct with A/B testing. Is $2.99 okay, or does $5.00 get me the same conversion? This is what you want to know. What features should I include in the $2.99 version or in the free version?” Pandrangi added.
“One of the customers we saw is one that does a premium service, and the customer tested $1, $3, and $5 as price points. They found out that the conversion was the same. So the price point was not a deciding factor. And this was in a subscription, recurring pricing model, so it is an important revenue discovery,” said Arai.
By Mark Boyd. Mark is a freelance writer focusing on how we use technology to connect and interact. He writes regularly about API business models, open data, smart cities, Quantified Self, and e-commerce. He can be contacted via e-mail, on Twitter, or on Google+.