Elder Game: Retention and Rebound

This Blog Post was part of the Elder Game blog. It was posted by Citan on January 31, 2011.

Previous Post: Elder Game: Graphical Upgrades Are Dangerous

Next Post: Elder Game: Who’s In Charge of Quality?

Additional Blog entries can be found on the Developers page or in Category:Game Blogs

Rage-Quit vs. Yawn-Quit

Long ago, when my friends and I were playing EverQuest, one of our party members decided to do some soloing after everybody else had gone to bed. He fell into a pit and died. In EverQuest, you lost everything you owned when you died, so all of his gear was now at the bottom of a pit. He then (foolishly) tried to get his gear back by himself, and subsequently lost all his backup belongings. The next day we offered to help him get his stuff back, but it was too late: he’d already rage-quit. He never came back.

If your last memories of the game are anger, then it’s much less likely that you’ll ever be back. But you might also quit just because you got bored, or some fancy new game caught your attention, or because you ran into money problems. In these cases, you may very well come back later.

No game will keep all their players forever, but when players do quit, you want to make it likely that they’ll come back again later.

Metrics Tell You What’s Going On In Your Game

As a game producer, you’ll be keeping an eye on certain key numbers to help you steer the game. Here’s some basic ones.

If people keep paying after the first free month (or whatever freebies you offer), they’re “retained” in the game. Obviously, you want your “retention rate” to be high — otherwise you’re in big trouble. A game that launches with a poor retention rate is likely going to die: nobody wants to play it.

But assuming a large percent of your players are retained past the first month, your next goal is to keep them around as long as possible. The length of time they keep playing is their ”retention length“.

People will inevitably quit at some point. Do they quit forever? Hopefully not. The number of people who come back to your game eventually is called your rebound ratio (or rebound percentage, or so on), and how long it takes them to return is the rebound duration. A game that’s doing really well might have a rebound ratio of 25% or more, with a rebound duration of eight to twelve months.

While priority #1 is to have the longest retention you can, you should also put a priority on making the rebound ratio high. This is the key to keeping your game going for a decade or more: your best fans keep coming back year after year to help keep the servers lively.

When a game damages the rebound cycle, they shoot themselves in the foot. SWG’s “New Game Experience” update did that for a lot of people — no matter what SWG does now, they aren’t going back. EverQuest burned the bridge for many because their death penalty was so high.

On the other hand, I don’t rage-quit most MMOs. Instead, I just get bored, either from lack of content or lack of friends or lack of compulsion to play. Those are all fixable things. In Lotro I simply couldn’t find people to play with; now that it’s “free”, I intend to go check it out. They didn’t burn their bridge. I didn’t rage-quit, I yawn-quitted. Much better for rebound. Knowing how many people are coming back is really valuable info.

One last warning about these metrics: don’t make the mistake of jumbling your new players and returning players together. If you use one single statistic for “retention length” for the game, it won’t mean as much.

A great game will have an initial retention length of 6 months — that’s how long new players stay. The best I’ve ever seen is Asheron’s Call 1, back when it had very little competition. Its initial retention was nine or ten months. (That’s incredibly high, and I can’t imagine a modern game beating that average, even WoW… there’s too many other choices to switch to when you get bored.)

But even the best game will have a lower retention rate for returning players. If your initial retention rate is six months, your rebound retention rate might be three months. Keep these separate so you can tell what’s going on.

Using Your Metrics to Make The Terrifying Changes

When your game is live, it can be hard to figure out what to change. No matter what the topic, somebody will tell you that it is ruining the game. Others will tell you it’s the reason the game is still afloat. It’s easy to become paralyzed.

Suppose you know that some people are quitting because the death penalty is too high. But other people say they’re only playing because the death penalty makes the game challenging. (Or any of the other reasons that MMOs might be designed with a high death penalty.) Should you change the death penalty? Hard question.

The above metrics will get you started in answering this question. Is your retention length high and your rebound rate fantastic? If so, no — definitely don’t change the death penalty. The people who are quitting may say it’s because the death penalty is too high, but apparently they’re still coming back anyway, so this isn’t a deal-breaker for them.

On the other hand, the more likely case is that your rebound rate sucks. And it’s up to you to figure out why it sucks and fix it so that the game will continue to have players in the years to come, long after all advertising money has dried up. Returning players are really important in the later years.

Whatever metrics you have, use them. They will at least help you narrow it down. If in doubt, address the issues that will alienate the existing playerbase the least.

If you have to change the game significantly, I would always err on the side of making the game easier. That tends to cause people to bitch, and some to even rage-quit over it, but mostly they still come back. Making the game easier may hurt viralness (the odds of someone telling their friend about the game) but it won’t hurt the game’s retention or rebound ratio, in my experience.

It’s all a gamble. It’s all making moves with very limited information. It sucks, basically. And no matter what you change, you’ll hear about it from angry vocal players. Worst of all, you have to wait a long time to find out if you’re right — typically up to six months before you can see how the changes are working.

But don’t make the decisions randomly, and don’t try everything at once (as I’ve seen many teams do). Unless you’re about to go under and are doing a Hail Mary to save the game, you need to make changes slowly enough that you can tell what the results are.

Not Just For Game-Balancing

When we say that games should record a lot of metrics, people seem to assume it’s for balancing the game. Sure, metrics can help balance a game. But instrumenting your game is really time consuming and expensive! (I hope I didn’t make it sound easy. Collecting the data correctly and turning it into useful reports can easily be a full-time job.) If metrics were just for balancing, the cost may not be worth it.

But it’s almost always worth it to heavily instrument your game and billing services, because that data helps you figure out how to grow. Statistics like retention rates and rebound ratios are just a start.

(If you’re stuck on how to get your rebound ratio up, here are a few ideas.)