What is in the name? Analysis of the names of 5820 games on Steam

“What's in a name?” More than you might think. If you have been working on your game for a long time or are in the planning stages, then at some point you will still have to choose its name. The names of video games can stimulate the imagination and cause nostalgic memories.

In Steam, more than six thousand games are released annually, but how many of them cause the players interest? Let's look at a study in which we studied the names of games released last year.

Why are names important?

Before moving on to numbers, let's briefly discuss why the name of a video game is very important. Although it may seem like a small part of the overall feel of the game, in fact, the name is the key, especially when it comes to marketing. In the frantic world of digital marketing, the name, advertising slogan and one or two images may be the only thing that a potential buyer will see.
')
In the Steam store, a name and a small picture are all that a game can attract. In the early stages of digital marketing, there were a much smaller number of games released. The increased shaft of manufactured games with a reduced marketing space makes the name more important than ever.

In addition, many other aspects should be considered. Marketing "word of mouth", places in the search engines, the ability to confuse the game with others and much more - these are the reasons for making additional efforts to ensure the name of the game "protection from the fool."

Now let's get to the data.

All statistics

To collect statistics, we analyzed 5,820 games released last year (excluding free-to-play). We considered only unique games released as separate products, therefore DLC and other extensions were excluded from the data set. The graphs are followed by a brief description of the data and methods.

Name length

For the 5,820 games analyzed, the average title length is 17.15 with a standard deviation of 10.26. That is, in general, the length of the names of games released last year ranged from 7 to 27 characters, not counting emissions.

The average name consisted of 2.8 words with a deviation of 1.7, which gives a total interval from 1 to 5 words.

This demonstrates a general tendency to more concise names, which allows us to choose one of the options: to follow the trend or intentionally avoid it. Both approaches in the current state of the market have their advantages, and they need to be carefully analyzed when choosing your own name.

In addition to the length intervals in characters and words, many of the games analyzed demonstrate similar naming structures. From the very first stages of the analysis, there were fewer than two dozen name structures in which more than 90% of the games fell. It is noteworthy that the most common structures of names were "{noun} of {noun}" and "{noun}: {adjective} {noun}" (for example, "West of Loathing" and "Divinity: Original Sin 2").

Multi-part titles

15.14% of the games in the entire data set had a multipart name. In the analysis, the name was considered to consist of several parts, when one of the following symbols was present: ":", "-" and "|", which thus created two separate pieces of information. For example, “Yooka-Laylee” is not considered to consist of several parts, and “Steel Division: Normandy 44” is considered.

Use unique words

One of the most important aspects of the data collected was the use of “unique words” in the name of the game. For purposes of analysis, a word is considered unique if it does not exist as a dictionary definition in the ten most popular languages ​​used by Steam players. In general, the unique word in the title contained 27.93% of the games.

References to other intellectual property

Having searched our data set for references to other intellectual property (Intellectual Properties, IP), we obtained a new subset that requires further analysis. 11.96% of the games in the data set to one degree or another referred to the already existing IP (previous game, film, book, or other media).

If you use this subset for further research to find the relationship between the sequels of games, games based on films and books, you can create your own interesting project, if someone is interested.

We evaluate the best

Let's see how these games relate to bestsellers. By taking the top 10% of the games and comparing them with the general data set, we will be able to create a clear picture of what is important in the names of the video games.

Name Length Comparison

The best-selling 10% of games released last year have an average name length of 17.63 characters. Compared to the 17.15 total data set, it can be said that bestsellers have on average a little longer titles.

Calculating the average lengths of titles in words shows similar results. The names of the best games consist of an average of 2.92 words, (2.8 words in the full data set). This is enough to show that the best games tend to be more verbose.

Multi-part titles

19.17% of games from the best-selling 10% have names from several parts (15.14% in the full data set). The difference in these two values. shows that it may be more profitable to use longer names. Perhaps this is due to the more optimal use of limited space in which developers can attract potential players.

Use unique words

As for the use of unique words, bestsellers differ dramatically! At 48.19% of the titles, at least one unique word is used (in the total data set of such games, only 27.93%). This is a huge difference in data; she suggests that using unique words makes the game more memorable, or more likely attracts a potential buyer.

References to other intellectual property

Top selling games contained references to other video games, films and books in 20.73% of cases - almost twice as many as the figure of the entire data set (11.96%). This demonstrates a clear tendency to the fact that the expansion of existing games usually show themselves better, or at least have a greater chance of success.

Data sets, sources and analysis techniques

The dataset used in this article is fully obtained through SteamSpy and the official Steam Web API . If you want to conduct your own research or try other algorithms, it is best to start from these sources.

When reading data, keep in mind that the limitations of the data sets and analysis methods may leave gaps that can affect the results. For example, there are fatal problems, such as the impact of free distribution of games on sales data and games with sales close to the statistical error.

Although this data is still useful, such problems are hard to control.

Data interpretation: what does it all mean?

We have processed a lot of data, but what does it mean specifically for you? Here are the clearest tendencies we found:

Make up titles of 1-5 words

Most games are in this range, and the best results that have been shown fall into the range of 3-4 words. Try to carefully select the words and make as detailed a description as possible in this confined space. If you have already found the perfect name, and it is a bit shorter (1 word) or a little longer (6-8 words), then do not worry - small variations are not bad.

Multi-part titles attract more interest.

Players usually prefer multi-part titles. Most likely, the reason is that they allow you to transmit more information in a limited space. A good plan would be to use a multi-part name that evokes additional emotions or contains a description.

The name should be unique and interesting.

Titles with non-vocabulary words are more likely to have good results, so something unique in the title allows it to stand out from the list of standard titles.

Sequels and franchises sell better

If you base the game on an existing job, then you will get a greater chance of success. The reason for this is most likely that the players are familiar with the previous IP. But this is good news for indie developers who want to turn one project into a series of games.

Further research for data lovers

Want to take your research to a new level? Here are two ways in which you can expand your analysis and get deeper results.

Smash IP Subsets

We talked a little about the percentage of games that refer to already existing intellectual property, but this subset itself may contain interesting data. How many games refers to previous games? Films? Books? Another interesting line of research can be a comparison of the results of different games in the series. For example, were the fourth games as successful as the third?

What about the worst games?

We compared the best 10% of games with the total data set, but what about the worst 10%? Segmentation of this part of the data may lead to interesting new discoveries, especially in which names should be avoided.