| Churned Y/N | Number of Players |
|---|---|
| Yes | 4652 |
| No | 8896 |
Advanced Data Analysis
1 Introduction
The purpose of the analysis contained in this html output is to explore patterns and multiple factors correlated with player churn within the gaming environment.
Since keeping existing players is frequently more cost-effective than sourcing new ones, ‘churn’ - the term for players who stop playing a game - is a crucial metric in the gaming sector. Using the data provided, we will analyse several player attributes, game engagement metrics, and potential churn predictions to understand how to retain the players.
2 Outputs Included
- Churn Count
- Average Play Time Per Day/Churn Status
- Churn Status by Day
- Churn Rate by End Type
- Daily Play Time Distribution
- Retention Rate Over Time
- Interactive Table
- Buy More Moves by End Type
- Trends Across Game Levels
3 Exploratory Data Analysis (EDA)
In this next section, we will break down the dataset into graphs, tables, and charts etc. to better understand the data. Upon completion of the analysis, we will be able to provide a deeper insight as to why players are churning, and how to possibly reduce the number of players churning.
3.1 Churn Count
3.1.1 Interpretation
In the table above, we can see the total count of players who have churned while playing the game, and players who have not churned. This gives us a better understanding of the the churn status, to which we can later refer back to.
3.2 Average Play Time Per Day/Churn Status
3.2.1 Interpretation
Moving on to our line graph, this gives us information surrounding average play time of players, combined with churn status, and day of the week. The line chart is interactive, meaning if you wish to view the data on a certain day, proceed to hover over the chosen day and the chart will provide the information.
A chart containing this information can be highly beneficial to understand the trend of what day of the week the players are playing the most, while also showcasing what day of the week has the highest churn numbers.
3.3 Churn Status by Day
3.3.1 Interpretation
The bar chart above visualizes the player churn status by days of the week. We have highlighted the players who have churned, while adding labels to show the count of each individual day of the week.
Understanding the individual count per day of the week will help us analyse trends throughout the week, such as what days players are most likely to churn on, weekday and weekend analysis, and play times on specific days.
3.4 Churn Rate by End Type
3.4.1 Interpretation
This barchart shows us the churn rate based on player end type while playing the game. To make the graph more legible, I have included percentages above the bar charts, to show the segment percentage of win/lose.
Drawing the users attention to end type labelled quit, showing over 52% is a worrying number. It begs the question, why are people quitting without finishing the level. This could be down to the fact that the game is not giving more start moves, and users ultimatley not wanting to pay for more moves, so they churn.
3.5 Daily Play Time Distribution
3.5.1 Intrepretation
The scatterplot above shows data obtained from daily play time distribution. As the week goes on, play time by seconds rises amongst players, slightly increasing on Wednesday, then rising rapidly for Thursday and Friday. This could be down to users work patterns, or social events.
There is however a drop off on Saturday, when amongst the gaming community, would be a prevalent day for spending time playing the games you love.
3.6 Retention Rate Over Time
3.6.1 Intrepretation
The line chart above shows data surrounding player retention over time. Upon inspection of this graph, we see that player retention over time is constantly moving. This could be for a number of reasons such as frustration from losing, or repetitivness. As we can also see, players that reach over 6000 days played, tend to become less, meaning that the players that have reached over 6000 days on the game, are bouncing on and off.
A side note to notice is that the retention rate over days played only drops below 50% once, suggesting that players do enjoy playing the game, however, this could be improved among the higher levels.
3.7 Interactive Table
3.7.1 Intrepretation
The table above shows the level of the game, the churn rate per level, average play time per level, and the total players per level. This table contains alot of data but can be a useful tool to sort through to understand what level of the game has the highest churn rate, hence why we have coded the churn rate red, to draw the readers attention to it.
This also indicates that total player counts decrease steadily as the levels progress, and this reflects natural attrition.
Using a table like this can give valuable insight into possible opportunites to improve player retention.
3.8 Buy more Moves by End Type
3.8.1 Intrepretation
Looking at the bar chart above, we can tell that players who are winning, have a higher average of buying more moves. This would suggest that players are more likely to succeed when buying moves. This could indicate there is a level of determination among these players.
Players who lose have a lower average of purchasing more moves, and this would indicate that they would rather give up than buy more moves.
Lastly, looking at quitting players and players who restart show near zero engagement with buying more moves in the game.
3.9 Trends Across Game Levels
3.9.1 Intrepretation
From the line graph above, we can see that as the used moves rise, so does the buy moves from players, meaning that the average number of moves provided by the game, is not sufficient to the number of moves to complete the level.
We have group the levels in increments of 1000, and as we look towards level 3000 - 4000 as the gap increases between used moves and start moves, there is also a sharp increase in buy moves to combat this. This shows that the hardest levels that require the most moves to complete, are between 3000 - 4000.
4 Key Insights
4.1 Churn Patterns
Churn levels vary significantly by different day of the week, with the weekend with some of the lowest drop off rates. This would indicate potential issues with player engagement with players during weekdays. We also notice that the churn rate is highest when players quit, and this shows critical player behaviour. Players quitting during a level within the game suggests either frustration or lack of motivation to continue.
4.2 Playtime Trends and Engagement
The “Average Play Time by Day and Churn Status” demonstrates that plays who have churned generally have a shorter playtime within the game, compared to retained players. This would give an indication that keeping players engaged longer during each session could reduce churn. If we then turn our attention to our “Daily Play Time Distribution”, this does show a slight increase in playtime during the week on Thursday and Friday but we notice a drop off when it comes to the weekend, contrary to expectations that gamers have a higher weekend engagement.
4.3 Buy More Moves and Game Level Progression
The “Trends Across Game Levels” shows that higher level in game require more moves to advance to the next level, driving an increase of players buying more moves. Between the levels 3000 - 4000, the gap between moves that are provided by the game, and used moves widens, and this would suggest that players do not have enough moves to complete the level, prompting them to buy moves. Players who also win have a higher engagement with buying extra moves, showcasing their determination to progress through the levels. However, quitting and restarting players show very little engagement with this feature, and this could be a missed opportunity for profits.
5 Conclusion
The analysis hihglights several flaws and gaps in the game that can be optimised to increase revenue. By addressing the issues with the game in churn patterns, daily engagement etc.. the game can create higher retention, and also create a better experience for players.