Guiding investments and development through natural language analysis

The project

In the gaming industry, you have thousands of anonymous users that you know very little about. For example, you don’t know how they feel about your game, your latest release and what they like or dislike about it.

The gaming community does, however, have a culture of reviewing games on gaming portals, and this makes it possible to use actual user reviews as a guide for game development investments and to provide structured feedback post-investment to studies on a publisher’s portfolio. It is, however, practically impossible to manually assess the general user sentiments across millions of reviews.

Implement cooperated with a publisher of games on using natural language processing (NLP) analysis of game reviews to guide the publisher’s investments and development. NLP is letting artificial intelligence process (in this case e.g. normalise “gaming slang”) and analyse text and speech as data, and it contains a huge potential for generating – and acting on – insights such as topics and sentiments from e.g. written reviews.

The final solution used NLP to analyse millions of reviews every day across platforms from Steam, Microsoft Store, Google Play and Apple App Store. Following the natural language analysis, the publisher displayed the results in a dashboard across a range of topics like graphics, performance, bugs, monetisation model, cheating and aesthetics.

The impact

As a result of the NLP solution, the publisher as well as all the studios now have much deeper insights into the user base of their games and would-be investments. Each studio can track opinions across topics in response to updates and compare with close competitors.

The publisher can follow performance of their investments in the market on a more subjective basis and better invest in future prospects to identify not just high-performance candidates but also thematic matches and synergies with the existing studio portfolio.