Analyzing 20+ mobile apps and 75K+ reviews with AI text analytics

In early 2019, Wonderflow started working with a major consumer electronics and hardware company. The scope included coverage of over 20+ local language apps and platforms like Amazon. We speak to Marissa, the then Product Marketing manager for various apps. We discuss what her work looked like before and how it has changed since she started working with Wonderflow’s AI-driven Natural Language Processing (NLP) solution.

Product or consumer marketers, customer care/experience managers, business analysts, and customer insights managers can benefit from the use cases shared in this case study. It clearly shows how AI-driven customer feedback analysis can help you build a better experience for your clients and serve many levels within your company.

Being a core source of customer insight for a global consumer-focused software provider is challenging, even on a good day. Marissa, Product Marketing Manager for an entire range of mobile apps, knows this more than anyone. She was asked for insights and reports daily and was tasked with finding ways to continually improve the customer experience and the insights gained from it.

Marissa started working with the AI-based NLP solution from Wonderflow in early 2019. We spoke with her to discuss what her work looked like before and how it has changed.

“I had to cover 1500 to 2000 new monthly reviews in the app stores. I used to collect, categorize, and analyze the reviews when I only had one app to manage. My most innovative solution was a scraping tool that exported reviews to an Excel sheet. Unfortunately, that didn’t work for the Google Play store.”

Marissa recalls countless days when she was scouring app store reviews. The fact that she speaks multiple languages made her job a bit easier, but that was an advantage most of her colleagues did not have.

When Marissa started working, a lot of valuable feedback slipped through the cracks. “I looked at reviews for four hours per week at a minimum before using the Wonderboard, and those hours did not properly cover everything. I didn’t feel it was sufficient, especially country by country, as I wasn’t actually able to do proper analysis and give feedback to the team.”

“Going through the reviews was also a pretty depressing task. There was an issue with the migration process, so our ratings dropped. People write nasty things when they do one-star reviews…” Multiple solutions were considered to replace parts of the existing process and manual labor.

The decision for the Wonderboard was one with an eye on the future. The most significant improvements were the ability to dramatically increase the number of business users and scale the tool throughout the enterprise. Digesting such a large amount of reviews from such a diverse group of sources has given the enterprise a significant advantage.

Marissa is now the lead for a range of consumer apps. With this new role, she is now responsible for more strategic decisions, leveraging the Wonderboard data.

“Ramping up the number of reviews we analyze to include over 20 different local language app stores, that’s a step change in capability we can’t even quantify. We can now spot trends, and the predictive analytics capability shows us how much app sales can potentially be improved if we improve specific functionality.”

“We have added many new users to the Wonderboard from different teams. Each new user is onboarded with care, allowing them to use the tool when they need insights. The variety of users brings forward many use cases, too. One of these is that we leverage the Wonderboard to track how competitors’ products are performing. We have learned so much from new feature launches by competitors, giving us insights for product development.

It is also the leading tool we use when launching a new app or functionality because we get deep market feedback quickly. Distributing this information to the relevant people within our company in a form they understand matters a lot. We have, in general, become much more market-driven and deal-driven.”

Different departments within the company use the Wonderboard. Users already save time by eliminating unnecessary or duplicate processes. Other key results:

• The company has since then been able to analyze all customer reviews rather than picking samples
• Increased coverage of sources to include 20+ local language app store sources (as well as Amazon for their hardware business)
• Increased product sales through predictive analytics, highlighting potential improvements
• Streamlined market feedback process for product launches

Future plans

• Increase the number of local language app stores
• Ramp up the number of trained business users by 200% in the next 6 months
• Ramp up the competitor analysis to cover all products from their 10 biggest competitors
• Closing the feedback loop with the hardware and software development teams