There are countless articles and docs on the web arguing what are the most important KPIs to look at how this KPI is the single most important vs that other KPI. While these articles and lists are all good, I feel a key element is missing from the discussion – and that is that different stages of a game (or any product) requires focus on different KPIs. I’ll try to be brief and to the point so this won’t end up a long post:

Pre Soft Launch

It should be obvious but I’ll say it anyway.
KPIs *doesn’t matter* before you have some real users.
Your colleagues don’t count, your friends don’t count, the 100 people you let in on hockeyapp/testflight from god-knows-where don’t count.

What does matter:
Prioritize development of core game mechanics (= make it FUN) and features that affect retention (= reasons to return to the game often – including timers/bonuses/refills/whatever)
Validate the app is reporting all metric events correctly.
Prioritize development of balancing tools for the retention features.

Soft Launch:

Primary Metric: Retention.
KPIs: D1 -> D7 -> D30 Retention

Secondary Metric: Monetization.
KPIs: Conversion (Free -> Paying Users)

Targets: Retention numbers vary between app categories and between game genres. As a base rule we aim for these target numbers:
D1: 40%-50%
D7: half of 1D
D30: half of 7D.

Actual targets differ by genre and platform. Some genres have lower D1 retention (strategy games for example), some higher. You need to find the correct benchmarks for the type of game you make and you still need to aspire to D7 being half of D1, and D30 half of D7.

Also, these numbers are for mobile. If your app is on Facebook – a good healthy game should target 20% to 25% D1.

Soft Launch is a period where you should aim to have no new features introduced. Unless you’re extremely lucky, or have a very big team, you’ll need the people during this stage to optimize and change existing gameplay mechanics/features in order to reach the target KPIs. If you find most of the team busy developing new features instead of fixing & balancing the current ones during Soft Launch, you’re plain and simple shooting yourself in the leg (business-wise).

At this stage you should focus on balancing core gameplay and tweaking existing features in order to improve the retention funnel and conversion from free to paying users.

Retention must be high enough at global launch since Retention is the hardest metric to optimize. Getting it right is top priority and can make or break a product. Monetization, in comparison, is a lot easier to optimize.

Soft Launch should last as long as it takes to get the target KPIs right. The optimal is 2-3 months. Some games need/can afford even 5 months. A really bad (but unfortunately too common) decision will be to have Soft Launch period end in 4-5 weeks.

Please notice I don’t mention other KPIs here. It’s intentional. The relevant people should take care these of course, but all other KPIs should be prioritized lower during Soft Launch. They also shouldn’t be a requirement when determining the end of the Soft Launch period.

Global Launch – first quarter

Primery Metric: Growth – eCPI, New Users
Secondery Metric: Monetization – ARPU

Hopefully we got the base retention funnel working correctly by this stage, and some good conversion numbers. If, for some reason, you went global with bad retention then your KPI priorities should remain identical to the soft launch period.

Now it’s time to focus on immediate growth metrics, mainly eCPI and New Users.

The high retention is key so your marketing team won’t waste huge amounts of money acquiring users. This is also the period your CPI is probably going to be the lowest in your app lifetime, so good retention and an ok conversion rate will save you a *lot* of money (and headaches).

Monetization is also an obvious KPI to optimize, I don’t expect you to be profitable at this stage, of course, but you should aspire to get break-even with the marketing spend.

Ongoing Live Ops

Primery Metric: Growth – eCPI
Secondery Metric: Monetization – ARPPU, Conversion

Now we mostly focus on eCPI vs Monetization, where monetization is the real important thing – unless we have problems with retention, of course.

We measure KPIs on two levels:
On going Balancing (AB Tests, numbers, 3rd party ads, etc)
Specific Features, Sales, Events etc.

The first is where the app is on avg, without any special things happening, and KPI improvement here is very difficult, takes long time and changes are small. This is where a Business Analyst or Data Analyst can make the big difference in spotting hidden patterns.

The second is the search for a game changer. Is a special feature changes KPIs significantly? Running Sales? Real Time events? This is how you find ways to create spikes in the KPIs, though they hardly translate to a constant improvement. It is finding how many times a week you can do special sales or operate a key feature.

Last point:

I’m obviously only touching the surface, but in general these should be your priorities.
Still, different companies can be at different market conditions. For example, I assume most of your users will arrive via paid user acquisition, which is the case for most developers. If you’re one of the top developers/publishers, discovery and X-Promo is a huge factor and a game changer. eCPI might not be as important for you, but releasing a more mature game to soft launch – is.

TL;DR
Different KPIs matter at different stages of a game/app lifetime.

Soft Launch:
Retention: D1 -> D7 -> D30 (Primary KPI)
Monetization: Conversion (Free -> Paying)

Global Launch: (3-4 first months)
Growth: eCPI, New Users
Monetization: ARPU

Ongoing Operations:
Growth: eCPI
Monetization: ARPPU, Conversion
Retention: Retention of Paying Users via introduction of new content.

Further reading:

Some quick links to check are this analysis of 415 free to play games done by GameAnalytics
A nice list of KPIS that matter for apps collected by Anshuman Bapna (which was posted on Quibb and actually triggered this post)
Actionable vs Vanity Metrics and How to Spot the Difference
Why Kabam chose LTV over ARPDAU
The most underrated App metrics for Game Developers