Postmortem: Appy Entertainment's Animal Legends
May 6, 2013 Page 4 of 5
What Went Wrong
1. Not enough time in soft launch/beta
The soft launch -- launching in a small limited market prior to going worldwide -- is starting to become a standard practice in mobile game development. For a period of a few weeks (or a few months) the game is released in a single market. American developers often soft launch in Canada. During this time, community managers can poll players for bugs and developers can fix any issues before the app is released into the wild globally. A soft launch is an extremely powerful tool -- especially for apps like Animal Legends, ones that are "always on" and built to be easily modified remotely.
Animal Legends was soft launched in Canada two weeks prior to the worldwide launch. While it allowed us to identify and fix several major problems, it was just not enough to get all the bugs within our advertisement services sorted out. These bugs mostly affected international inventories but we just didn't have enough time to get a fix in, resubmit, and hit the date we had penciled in with Apple.
Ultimately, the quality of the gameplay experience was there but some of our secondary revenue services were not, so we opted to move ahead with our launch date. In retrospect, we did lose some revenue in international markets. However, had we delayed the launch, we could have lost our Featured slot with Apple.
In the future, we'll pencil in a much larger testing window before we give Apple a date. If we don't find any major issues, we can keep the app in the wings until the pre-set launch window.
2. Lack of extensive testing of remote data
Animal Legends (and SpellCraft before it) were built to be updated remotely. All aspects of the economy, combat, heroes, and gear are placed into "plists" which can be changed by updating our remote servers. This allows us to fix bugs or issues that players run into within just a few hours instead of having to resubmit the app.
Though we tested dozens of variables to make sure they propagated right, Animal Legends features thousands of variables. As you might expect, some of those (specifically, three of them) didn't update in the app properly when changed remotely. One of the variables simply refused to deploy; another would apply the change and, every time that specific game actor would update it, roll back to the original data.
Once they were identified we were able to resubmit and patch them within a short period of time. However, we lost precious time. In the future, we'll test all our remote variables before releasing the game.
3. Lack of preparation for emerging markets
We invested in a large amount of translations (English, French, Italian, German, Spanish, Simplified Chinese, and Japanese) for Animal Legends so we were ready for international markets. But we weren't fully ready for emerging markets -- especially China.
SpellCraft enjoyed a good amount of traction overseas with roughly 60-65 percent of the total number of downloads coming from places outside of North America -- mostly Europe. Animal Legends, however, has pushed that number up as high as 75 percent, with downloads mostly coming out of China, Brazil, and other emerging markets. To put the number in perspective, FaceFighter, SpellCraft and Trucks and Skulls had roughly 10-15 percent out of China. For Animal Legends, China represents roughly 30 percent!
The Great Firewall presents problems with some of our online features (like Google's AppEngine, which isn't welcome in the country in any form). Other services -- such as Amazon S3 -- are very slow. While the Great Firewall didn't stop the game from functioning, it did occasionally stop us from deploying a patch or sending messages to our players.
We didn't integrate Sina Weibo or QQ either, which meant that Animal Legends in China lacks any sort of sharing to social networks.
We are currently making substantial changes to the game for China, including placing servers on Chinese soil and enlisting the aid of native Chinese companies. Fortunately, a number of companies have emerged to help western developers bring their titles to China, like Yodo1.
The Chinese market is by far the most alien for western game developers. What the Chinese consider appealing in terms of design, art style and business models are quite different from the world beyond the Great Firewall.
The Chinese market has always been important -- but it's now the top mobile market worldwide, surpassing even the U.S. Fortunately, companies are beginning to emerge to help Western developers not just get into China, but to have a fighting chance in their competitive app market. With the help of Yodo1 we've "deep localized" Animal Legends. Updating our art, marketing assets, localization and even our UI to meet Chinese players' expectations. The company has done a tremendous job and their CEO Henry Fong has been a great ally for our studio.
4. Mana economy
Every hero's abilities are driven by mana. Some of the weaker abilities can be cast for very little mana, while the more powerful abilities consume a good amount. Mana regenerates over time, but it's a key driver for combat, and the main throttle for progression. While the mechanic works well in combat, if the player leaves the app and heads back to their town they can't tell how much mana is left for each hero, whether they're currently in or out of combat.
The feedback the player receives on mana regeneration isn't really an issue if he or she is playing nonstop, but most play Animal Legends six times per day, leaving the app and getting back in multiple times.
The relevant issue seems to be to know for sure when heroes have full mana (energy) and if they are ready to adventure again. We are working on it for a future update, which will add a feature that communicates each hero's mana from the main town view.
Page 4 of 5