How to select Key Performance Indicators for an Online Game?

Key Performance Indicators (KPIs) are very important in evaluating how your game is doing.  Some people believe that it is important to have every indicator measured as a KPI while some believe it should be a set of a few indicators.

In my opinion, KPIs should be high level numbers that encourage research and discussion.  If it was up to me, I would use the following KPIs:

• Peak Concurrent Users (PCU):  PCU is the most basic of online KPIs because it shows many things such as the following:
• Technical Issues:  Sudden drops in PCU indicate there was a major service issue or if there was a small technical issue.  For example, a small drop hints that one shard / server may have crashed while a total drop means the entire service went down. 
• High Level Player Engagement / Retention / Acquisition:  PCU over time tells you how your player base is trending.  If you have an acquisition program and your PCU changes at a rate higher than before, then it was successful.  If it remains the same, it was not successful.  If it drops, well you had a campaign that hurt the user base.  If you are not doing any acquisition and your PCU goes up, maybe a new consumer base found your game.  Who are these people how do you get more of them?  If your PCU is drop rapidly over time, your player engagement and retention is horrible.  If your PCU stays flat without any acquisition, your player engagement is good and your retention might be good.  You will need to look deeper to see if you have high organic acquisition and low retention which also might cause this.
• Informs Service Strategy:  PCU tells you at what time people play and when you would except the most service calls. PCU also dictates the amount of servers you need to support your game.  Not everyone is online at the same time so you need to make sure you do not over commit on server hardware.
• Competitive Impacts:  PCU will also change based on what your competitors are doing.  If you get a high drop in PCU, that means your competitors really hurt your game.  You will need to play for a counter attack.  If your PCU is not affected, that game was a different audience than your user base.  Should you check to see what type of user plays that game and why?  That is up to your longterm strategy.
• Operating Profit (OP):  OP is the most important financial measure besides ROI.  ROI is not a KPI because it should be campaign, program, or development based.  OP is a daily, weekly, or monthly measure that tells you if you should keep your game alive.  OP is more important after the initial launch of the game because it might lead to bad decisions before the launch of the game.  Only revenue and costs that can be made or saved by turning off the game should be factored in this number.  OP is an indicator of the following:
• Sustainability:  Remember that the investment in the game and all other prior investment are SUNK COSTS.   Keeping this in mind, this metric shows you how profitable the run rate of the game is.  Even if you do not cover your fixed costs, you are covering the costs of doing business and making enough to support other projects.
• Right Sizing the Business:  Once you launch, you will be able to see how successful your game really is.  Based on this, you can project your monthly OP.  If your OP is negative, you can lower some of your support costs.  Remember the vicious circle of Online Game Failures, cutting marketing to save money = lower acquisition.  Cutting Service to save money = lower retention.  Cutting new content to save money = higher churn.  Cutting everything to save money = no users.   Right size your business so that you have enough marketing, service, and content to maximize your OP!
• Economic Costs:  Even if your game is OP is positive, your OP% might be next to nothing.  You will need to compare keeping the game alive versus investing that monthly run rate into another project.  If you are doing your job well, the game should have a 70-80% operating margin or higher for an MMORPG.   This should make the economic costs of keeping the game alive very low.
• Churn:  Churn is very important metric because it shows you how well retention is and what average life is.  This metric is similar to PCU but is more focused on retention and average life.  Remember using the change in the subscriber base IS NOT CHURN.  Churn is the % of players leaving who had the chance to play the game!  The key ways this KPI is used is as follows:
• Retention Measure:  While PCU tells you how your overall user base is doing.  Churn, when used as Volume Churn, tells you how retention is.  Volume Churn is defined as follows Total Players Who Stopped Playing / Total Players Who Played the Game That Period.  Volume Churn controls for acquisition.  If you use just change, acquiring users may hide the any lose in players.  This metric does not do that.
• Average Life Estimate:  1 / Volume Churn is an estimate of your average player life.  If a service has a 33% churn.  1/ 33% =  3 months.  As you can see, this metric does work.
• ARPU:  As discussed before in a previous post, ARPU is a key indicator of your overall profitability by user.

Other KPIs that should be considered are based on your type of game:

• Conversion Rate:  Subscription based MMORPGs use Trial to Playing conversion rates versus Free2Play games use the % of their active player base who play.
• Daily Users:  A measure of player usage of your product.  PCU is a better indicator because of all the other factors it helps measure.
• Daily New Users:  A measure of player acquisition.  More important for Free2Play games that subscription based games.
• Average Weekly/Monthly Playtime:  A measure to tell you how longer your players are engagement.  Based on amount of content you have, you can project how long it will take people to complete the game on average.
• CS Satisfaction Scores:  How well is you Customer Service team is doing is very subjective.   Player Surveys or Satisfaction metrics are good at keeping your CS Support on the proper cost level.  The goal is to balance costs with good CS Support.  Free2Play games should have minimum support.  Statistics that should be included in this calculation are average response time, average resolution time, and player CS scores.

These are overall good high level metrics that after one glance can help you quickly evaluate how your game is doing which is the purpose of a KPI.  Any deeper metrics is a way to analyze the games performance and should be done in a more in-depth manner.  KPIs belong in a dashboard for all key stakeholders to review.  Once a problem is highlighted, then metrics to help explain the issue should be reviewed in EXTREME detail.  Metrics without context lead to BAD DECISIONS.  Every game is different so remember to keep that in mind.  Not all metrics are applicable to all things.

In addition, I personally like to use dashboards that list all key marketing and production activities that occured so you can visusally see the impact!