Malpa Games. Customer feedback management software for a mobile games publisher

The story of how we organized the full-cycle customer feedback software, including role separation, filtering, flexible delivery of analytics and response automation using artificial intelligence. And the prospects of becoming a full-fledged SaaS product.

Our client is a mobile games publisher that needs to manage user feedback on its apps in the App Store and Google Play stores, as well as from Gmail.

To do this, Malpa Games needed its own ticketing system – a program for registering, tracking and managing support requests. Such services track feedback from various sources, automatically create requests (tickets), distribute them to specialists, and record the entire process of handling responses and resolving issues. They also collect valuable analytics – each ticket contains technical data: information about the device from which the review was made, the application release version and other parameters.

During the project we had to solve the following tasks:

• Aggregate user reviews of mobile applications from all sources in one program;
• Provide different scenarios for working with tickets and convenient filtering tools;
• Distribute functionality between different roles of ticket system users;
• Collect and visualize analytics by required parameters;
• Automate part of ticket processing through AI integration.

The project started as an MVP, but with each new release new features were added, turning it into a full-fledged helpdesk for customer support staff.

In the early beginning, Malpa Games managers used to process feedback manually. Then the company realized that this was time consuming and inconvenient and started using off-the-shelf ticketing systems. However, a problem arose. There are now two types of service on the market:

1. One automates the tickets processing, but has poor support for collecting reviews from the App Store and Google Play.
2. Others provide good analytics to promote mobile apps, but don’t include convenient software for working with feedback.

That's why Malpa Games needed a more comprehensive solution: feedback management software that collects the maximum amount of technical data, automatically creates tickets, analyzes and distributes them to responsible managers, assigns tags, manages statuses and working groups.

• 195

  Number of tickets processed per day

• 56 k+

  Total number of tickets processed

The CEO of Malpa Games had been in contact with our CEO Max Bantsevich for several years, so he was well aware of our experience in developing custom systems and process automation tools. At the right moment, the company approached us with the project idea and the main problem.

We had a great synergy with Malpa Games, thanks to which we were able to constantly improve the product and continue to do so. After each new release, the managers who use the ticket system give us their feedback and requests. This allows us to quickly make improvements and add new functionality.

Where the Malpa Games ticketing system started and how it works now.

Let's take a look at what happens inside.

There are several roles within the ticket system:

• Read Only – can only read the text of the feedback, but without the ability to respond to it or change the status of the ticket;
• Agents – work with tickets and respond to feedback;
• Lead – manage projects within the ticket system (heads of support departments);
• Admin Manager – have full access to tickets, full access to everything that happens with tickets within the company, including management of automations and integrations.

The entire list of employees is located in the Clients section, where you can add new employees, delete accounts and change their roles.

How to process the tickets using the customer feedback analysis and response software.

This section displays all the tickets for the project, collected from all sources. For each ticket you can see:

• Status (set by the manager himself);
• The source of the ticket;
• The text of the review;
• The rating given by the user;
• The release version for which the review was left;
• The operating system of the user's device;
• The language in which the review was written.

Gradually, more and more managers joined the system, which made it necessary to track all actions on tickets: who assigned a particular tag and when, who sent a reply to a review. In addition, it was necessary to track edited reviews (Google Play users have this option).

Therefore, from the moment a ticket is created until it is replied to, all details and technical data are stored with full details of all changes. You can also use filters to track the work of each manager. For example, you can see what status is most often assigned to tickets by this or that specialist.

Our approach for process automation.

In order to partially automate the work with tickets, we also connected an artificial intelligence to the system, which should solve the following tasks:

• Recognize the emotional color of reviews - negative, positive, neutral, etc;
• Determine the reasons for negative reviews - for example, a large number of advertisements in the game;
• Assign a tag.

Main challenges and useful tools.

To develop an admin panel where you can manage accounts, companies and projects, we used our own product Admiral – a front-end framework for back-office development on React. It has a lot of out-of-the-box components and tools that make developing the admin panel interface easier and faster.

App Store API passes all reviews by default, but taking the entire account history is not necessary. Therefore, we take reviews in order from newest to oldest. If the date of the next review we write to the database is older than 14 days, the download will stop. This reduces the load on the integration.

By default, the Google Play API only passes reviews from the last 7 days (new or updated). We simply take all available data. Also, in Google Play, there is only one review-response pair at a time. When a user updates their review or someone changes their response to a review, Google doesn't change the review and response IDs, but updates the date and time.

Another interesting point related to mail. Since Malpa Games has a common corporate e-mail, we created alias addresses for it – additional addresses to the managers' accounts. With their help we were able to differentiate the distribution of incoming feedback - the entire chain of emails is filtered by the recipient (alias address) and the feedback goes to the corresponding project.

Frontend: React, Next.js

Backend: Laravel

Libraries and services: Docker

We started working on the ticketing system in May 2024. Dashboard development took about 3 weeks, integration with AI - about 4 weeks. We stay in touch with the client and are always ready to add new functionality to the product.

• 10+

  months