As we announced earlier, our virtual world of Spelldom finally closed its doors this month. It’s been almost three years since we launched back in 2015. We’ve all spelled a lot of words, built some elaborate towns, forged some great relationships with our teammates, and crushed some noobs with less majestic towns and/or word skills. Unfortunately, we’ve looted the dark elves for all they’ve got, so there’s no more gold to cover the server and maintenance costs.
Let’s kick things off by congratulating some of the most accomplished users. Here are the top users in Spelldom’s competitive categories.
The Daily Champ
1. Dante 412
2. AdamPlays 159
3. JosephdePalmy 126
4. Leela 86
5. Bandor 31
Scrolls
1. Dante 4,279
2. Baba 3,331
3. DavidRicky 3,239
4. Rat 3,234
5. Pandemonium 3,041
Team Scrolls
1. Water Lizards 15,814
2. Village Orphans 11,473
3. Repus Dren 10,353
4. Del Ray 9,528
5. Winterfell 8,375
Congratulations to everyone!
Lessons From Spelldom
Spelldom was the most complex game we’ve ever built, both from a technology standpoint and a game design perspective. Now that we’ve wrapped up the magical woods, we wanted to share some of the lessons we learned along the way.
Choosing the Right Backend
Starting with technology, we built Spelldom’s backend with Kinvey’s mobile backend-as-a-service (BaaS). The BaaS market was and still is highly fragmented, so it’s often tough to choose the best option. Parse was likely the largest player in the BaaS market at the time, followed by Kinvey and StackMob. Gaming centric backends like GameSparks hadn’t taken off yet, so we had to build out all of our team mechanics, matchmaking, and leaderboard code from scratch. Ultimately, we chose Kinvey over Parse because we needed to make a ton of server calls to power the game. Parse charged per call, while Kinvey charged a flat fee. Also, we didn’t want to risk using a smaller player, for fear that they would go out of business. Over the life of Spelldom, it seemed nearly everyone disappeared except Kinvey. Parse was bought by Facebook, then shut down. StackMob closed up shop. Countless smaller players went under. We think we made the right decision given the options in 2014.
If we were doing it again today, however, there are many other compelling options. Amazon Web Services (AWS) has built out a more mobile friendly set-up, which allows you to save costs compared to BaaS companies that are running their own interface on top of AWS. Couple that with Kinvey’s shift in emphasis toward higher cost enterprise plans, and the cost savings are significant.
GameSparks, who Amazon just bought last month, specifically targets games by offering lots of prebuilt gaming functionality. This allows you to save time building out a lot of multiplayer features yourself. There are some drawbacks. It’s nice to have the granular control that comes with writing all of your own server code from scratch, plus you can more easily move it to another platform if necessary. Nevertheless, the time savings are hard to ignore.
Finally, Google Firebase’s real-time database is ideal for powering team chat. It’s definitely an easier set-up than the intermittent querying approach we wrote for Spelldom. Regardless of which platform you use, you can build a multiplayer game backend cheaper and more quickly than you could just a few years ago.
On-Boarding Users
Day one retention—otherwise known as the percentage of users that continue to play your game the day after they download it—is key for mobile games. While console gamers have a monetary investment that makes them more likely to give games a thorough chance, mobile gamers will casually move to the next free-to-play game if the experience doesn’t immediately suit them.
Consequently, a lot of thought and testing goes into the on-boarding process. Mid-core strategy games like Spelldom require a good deal of explanation for users to grasp all of the rules and nuances of the game. You want to make sure users learn everything quickly so they can enjoy the game to its maximum potential. On the other hand, you don’t want to bog users down in a tutorial for too long, or they’ll get bored. It’s helpful to make the tutorial as interactive as possible to keep players’ interest high, but you’ll still need to grant the player more freedom quickly or they’ll move on to something else.
Our original goal was to make sure the tutorial was no longer than the tutorial for Clash of Clans. We accomplished that, and had plenty of guided player interaction. Unfortunately, day one retention was still a little light of what we wanted. We had already built a lot of analytics into the app in order to check for bottlenecks in the game flow. Oddly enough, there wasn’t an obvious churning point where players were more likely to leave the game. Departures were spread out pretty evenly throughout the tutorial. It was simply too long.
After a lot of experimentation, we shortened the tutorial dramatically, leaving out some of the less critical information. To offset this, we added a series of early game achievements that players could pursue at their own speed. These goals effectively taught players how to upgrade their towns and improve their attacks without the heavy-handed nature of a formal tutorial. It worked too! When in doubt, keep the tutorial as short as possible, and let the players learn the other rules in a fun organic way.
Critical User Mass
One of the most challenging parts of managing a multiplayer game is the need for a large user base. You need players at all stages of progression, so that you can match players of similar skills for attacking. A competitive team tournament scene requires a lot of teams, which requires a lot of players. Whenever you make changes to the on-boarding process or some other aspect of the game, you’ll want a lot of new players to test whether your changes made sense. As soon as the player base starts to dip, players don’t get attacked as much, which lessens their incentive to build their defenses. They also don’t have to log-in as often to collect their resources. What’s the rush if nobody is going to steal their gold!
Indeed, Spelldom players launched fewer attacks as the active player base shrunk in the latter days. The majority of player activity started to center around the Daily Champ competition, which didn’t require as many players to foster a competitive leaderboard as the team competition.
The moral of the story? You’ll have to acquire a lot of users on a continuous basis to make your multiplayer game work. Unlike a single player game that can randomly catch fire without a large user base, a multiplayer game needs to already be on fire to attract more users organically. Test which ads get you the best quality users for the price—you’ll need all the users you can get! Also, be careful not to build a multiplayer game that is too niche. Spelldom appealed to fans who liked both word games and strategy games. This is a smaller population than fans of more vanilla strategy games, which made it harder for us to keep a large player base.
Daily Champ Competition
The Daily Champ competition was arguably the most popular aspect of Spelldom, although we didn’t even have it at launch. Our original daily game—the Gem Grab— was a purely single player experience that allowed players to get some extra gems. It was more of a side game than a major attraction like the Daily Champ.
There are a few takeaways here.
First, don’t underestimate the power of a great daily game. Anything that really encourages the user to open your game everyday is great. Honestly, this is probably even more true for a single player game, considering a successful multiplayer game should have other incentives for daily logins, such as team chat.
Second, the enduring popularity of the Daily Champ beyond the main multiplayer game probably suggests that Spelldom’s core users who played for months, and in some cases years, were more passionate about the word game aspect than the strategy. Yes, the Daily Champ had an advantage over our main game, as it didn’t require as many users to create good competition. Nevertheless, even the users that played the main game heavily for the longest time largely had incredible word game skills.
Spelldom was challenging. We’ve written about it in previous posts, but it’s always easy to misjudge the difficulty of a game that you play for countless hours during development. Mark that as a bonus lesson. Spelldom did a good job of retaining serious word game players. In the end, it just didn’t capture enough of the casual audience to keep the player base sufficiently high.
Conclusion
Thank you to everyone who played Spelldom. The game was a huge undertaking, and we’re humbled by all time people spent in our game. We hope to provide you with more entertainment in the future. Until then, keep your skills sharp. We usually like to make things challenging!