In a new study from the University of Skövde, the University of Virginia, and Lancaster University researchers have developed a model that the computer game industry can use to predict competitor sales, and plan future launches and investments in their own games. The model is based on own past product sales data and competitor's prerelease Google Trends.
In terms of market value, computer games are one of the largest sectors in the entertainment industry. Many games require very large investments and can be under development for several years. Predictions of a game's sales are very valuable information to determine which resources that can be allocated to produce and support game titles during development and after launch.
Online behaviour improves predictions
As the computer game industry becomes increasingly competitive, companies need to gain insight into the actions of competitors and new product launches. Today, many predictions are made based on data from own sales and the product characteristics of computer games. By taking in more available data from Google searches and other online behaviour, the researchers have created a model that improves the accuracy of predictions by 18 percent, which is very valuable for computer game companies, to coordinate marketing activities and actively manage player acquisition and retention.
“The online buzz surrounding a game before its release is found to be a very good indicator of its sales, both during launch and over its total lifetime. We develop mathematical models that enable us to translate publicly available information on how consumers search for computer games into sales predictions.” says Nikolaos Kourentzes, Professor of Informatics at the University of Skövde.
Available for smaller game studios
The model can provide invaluable insights for a computer game company that gains insight into the potential success of competitors' products and how these are positioned against their own. It makes it possible to optimize the company's plans and make better use of available resources. It’s a relatively cheap way to get very valuable information, says Nikolaos Kourentzes.
“The results offer a low-cost and effective way to obtain competitive intelligence. Conventionally, this is a very resource-intensive task, with limited success. Our approach is therefore offering a new way to obtain competitive intelligence, which is often not available to smaller studios.”
The model can be translated to more industries
The research is directly applicable to the computer game industry, which is also confirmed in the research article by the game company 2K Games, which states that the improved model developed by the researchers can, among other things, help them with launch strategy, to retain players and to get an idea of the expected return on their investment. But Nikolaos Kourentzes sees several possible areas of application.
“As it is possible to source the aggregate online behaviour of consumers for different sectors and areas of interest, the proposed methodology can be beneficial for other sectors or stakeholders. For instance, one could potentially track the interest and potential adoption of governmental policies, use of incentives, etc.”
More about the research
The journal article "Predictive competitive intelligence with prerelease online search traffic" is published in Production and Operations Management and is open access. https://doi.org/10.1111/poms.13790