Author: by Paul O'Connor

[Appy Entertainment, the developer behind Trucks & Skulls and FaceFighter, recently launched its first free-to-play iOS game. Studio brand director Paul O'Connor picks apart the successes and failures in getting to grips with the casual and F2P space simultaneously.]

Appy Entertainment committed to developing SpellCraft in late 2010. Trucks & Skulls launched as a premium hit that November, enjoying Apple's top banner support on iTunes, but even during that period of record earnings we knew the iOS market had shifted. And we were right. After a strong run through the holidays, Trucks would drop out of the App Store Top 100. The market was speaking with its wallet -- it was freemium or bust. As of this writing, 70 percent of the U.S. App Store "Top Grossing" games are freemium titles, and there's little reason to believe that trend will reverse any time soon.

Short-term, we responded by pivoting both FaceFighter and Trucks & Skulls from premium to freemium monetization, but these games were like battleships hastily converted to aircraft carriers in World War II. Players rewarded our changeover with improved reviews and revenue, but neither game had the kind of data-driven, collecting-oriented gameplay that best takes advantage of the freemium format.

We needed a purpose-built social-mobile game -- something innovative, fun, and with plenty of room for expansion. We were fans of games like Zombie Farm and Pocket Frogs and had spent a lot of time on Facebook with Zynga games like FarmVille, but as a studio we had never built a social-mobile game from the ground up.

Our first concept was kind of like Dungeons & Dragons meets Ravenwood Fair, but we quickly realized the idea was too big and too ill-focused for our first foray into this new genre.

We simultaneously elected to reduce scope (by focusing just on wizards, rather than the "full" role playing experience) and broaden our market beyond D&D players by appealing to fans of the most potent fantasy property of our time and also embracing the farming mechanics of Zynga's top game.

Thus armed with the killer high concept of "Harry Potter meets FarmVille", we donned our wizard robes and repaired to our tower to brew spells and blow stuff up.

What Went Right

1. Outsourcing

Appy is a young company with a small staff, and a major challenge in SpellCraft was to amplify our internal capacity by outsourcing production wherever possible. As was the case on Truck & Skulls, much of the engineering was done out-of-house by Quinn Dunki, but where SpellCraft broke new ground for us was in the extent to which we outsourced our art.

We'd first started searching for outsourcing assets while producing updates for a previous title (Trucks & Skulls), so we had a running start in putting together our list of candidates to work on SpellCraft. Production design remained in-house, as did game design, UI design, animation, sound, and some asset creation, but 65 percent of SpellCraft's in-game art was produced by outsourcing firms in India and China under our tight direction.

(click for full size)

We vetted each firm with test art assignments before committing to them for SpellCraft, to make sure they could follow our direction and adopt our house style, and were pleased we could get such high quality results without having to embed anyone at the outsourcing studios. It was a significant step for Appy to increase our art bandwidth with external resources while maintaining the high quality of our previous (and mostly internally-produced) titles.

2. Canadian Soft Launch

An unusually fast approval by Apple (we got the green light on the Sunday of Thanksgiving Day weekend, hats off to the Cupertino guys for working the holiday weekend!) meant our game was available a full 10 days prior to our final launch date.

This let us do a "soft launch" for SpellCraft in limited territories -- in this case, Canada -- to test the game with real players before going worldwide. While we weren't always able to act on it, the data we got from actual player behavior was invaluable, and we've already determined that a longer and more wide-spread soft launch will be built into our future games.

3. W3i Partnership

One of the things we know at Appy is what we don't know, and we knew we had little expertise in launching and analyzing social/mobile games. We addressed this need through partnership with W3i, which invested in the game and consulted in SpellCraft's development.

W3i is a Minnesota-based software company with a deep background in monetizing desktop software which is moving into the mobile space in a big way, and we'd been in discussions for several months about how we might do business together. W3i specializes in app marketing and user acquisition -- two areas where we really appreciated their expertise. W3i was especially helpful in the run up to launch and our test market period, helping us sift through mountains of player data to make the game more fun and "sticky."

Post-launch, W3i provided guidance on how best to market the game with targeted ad buys putting the game in front of the right audiences to find new players. It is true of all iOS games that your job is only just getting started when you ship a game, but it is doubly true of social/mobile games, and we've been fortunate in picking good partners to navigate these waters in W3i.

4. Customer Support

Appy's made a significant investment in customer support, maintaining open channels on our blog, Facebook, and Twitter, and inviting player feedback via email and "Livebar" messaging in our apps.

It is basically my full-time job to interact with our fans and listen to their concerns. So far. I've been able to do it largely on my own, but if the response to SpellCraft is any indication, we may need to build out our customer support program soon.

Every single communication Appy receives is personally answered -- sometimes within hours or minutes, even on weekends and holidays. Personally connecting with players is one of the ways Appy has been able to punch above its weight and compete with substantially larger companies in our space.

For SpellCraft, it's been an immensely positive experience to have a dialogue with players, both because it has helped to shape the development of the game going forward, and because it has helped accelerate our understanding of the new kind of player attracted by SpellCraft.

We've had some crossover from loyal players of Trucks & Skulls and FaceFighter, but for the most part, our SpellCraft players are different than our existing player base, with more women and more casual players coming into our fold for the first time.

These players have different expectations than our pre-existing maniacs who want to punch faces or smash skulls with trucks, and having all those touch points open to our new customer base has helped us quickly get oriented to our strange new world. Our customer support channels would also prove a critical lifeline in sorting out a nasty bug during SpellCraft's launch (about which more shortly).

5. Polishing Time

We made a conscious decision to delay SpellCraft's launch (and possible support from Apple) because we felt the game just wasn't ready for prime time. We were divided internally about the decision to delay, but ultimately decided that we only had one chance to launch, and that we were invested in SpellCraft for the long haul, which meant that we couldn't go to market without additional UI polishing and gameplay balancing.

We swallowed hard, cancelled our marketing spend, informed the sites that had previewed us that we were slipping our release date, and reluctantly told Apple that we just weren't ready, then extended our crunch to kill more bugs and make sure that our "onboarding" process (the tutorial experience and first three hours of gameplay) was nice and shiny. Every project, it seems, can benefit from more time at the end, but in the case of SpellCraft this was an especially difficult inevitability to accept, because of...

What Went Wrong

1. Late Launch

Actually, in the scheme of things, SpellCraft was on time -- we were only two or three weeks late by our own reckoning, and while weeks = months in iOS development, for a project with so many moving parts we pretty much shot directly through the hoop of fire.

The problem was that, owing to the vagaries of App Store approval times and shutdowns, this pushed our release date back from November 17th to December 8th, which both put us square in the Christmas crunch, and also may have cost us an opportunity for an Apple feature at launch (they were engaged with us in the run-up to the 17th, but after that I think they had other priorities). Launching as late as we did meant that our guys were also eager to head off for the holidays immediately after we went live, which was problematic because of our...

2. Reset Bug

We developed the game for five months, tested internally for six weeks, and tested in a live market for a full ten days, but still ended up shipping our game with a nasty bug that erased player progress and purchases if the game crashed during startup. We just missed it.

Actually, we had two reports of this bug during our soft launch, but I dismissed them as user error because they came from kids (one of which was my own scatter-brained son!) Regardless, no sooner had we shipped the game and the team had crawled off to their caves to recover, reports started flooding in of our reset bug.

We launched wide on Thursday, fixed the bug on Saturday, and (thanks to an Apple expedited review) had our update live by Tuesday, but we still disappointed thousands of players, resulting in an avalanche of customer support mail (the answering of which became a full-time job, inhibiting our ability to promote the game at launch).

We reacted as fast as we could, made things right with the people who wrote us, added a 24/7 live bar message informing players of the bug and how to update before purchasing, and answered each and every email, giving replacement gems to all of our affected customers that contacted us. Luckily for us, our players were understanding and didn't let our bug interfere with their affection for the game, giving SpellCraft an average user rating of 4.5 out of 5 stars since launch.

3. Undercooked Pets

Pet meat is stringy if you don't cook it right, and the same could be said of our SpellCraft pets. We knew we wanted pets in the game; we knew what they should do; we spent a lot of money designing and animating the pets; but we didn't bring it all together in the form of a fully-cooked feature for the game's release.

The pieces are all there -- you find pets in the dungeon, feed them and play with them, and they increase your character's defenses -- but there's really no way to understand this from the way pets are presented in the game. We're fixing this in an update, but considering the time and effort required to even under-deliver on pets, we would have been better off introducing those features later and applying that effort to a more polished 1.0 version of the game (and the same could be said of our multiplayer dueling feature).

We've made iOS games for three years, but we do still sometimes misjudge the scope of our projects, and pets definitely fell into this category. At the same time, our players have responded to pets with affection well out of proportion to our implementation of the feature -- pets are fun, and polished, they're just not as fully-featured as we would like, and players are confused about how to use them. We are definitely on to something here, and we wouldn't have such a firm idea of where we need to take the game if he hadn't included pets -- in however imperfect a form -- in our first release.

4. Not-So-Daily Reward Feature

We had the right idea -- encourage players to engage with SpellCraft every day, to get them hooked on the game -- but we made the requirements of our "Daily Gold Reward" too stringent, and wound up with a pack of unhappy players writing in to complain about not getting their gold.

Our original system required that players open SpellCraft and perform 10 game actions, such as fighting monsters or brewing spells, meaning it excluded players who just entered the game long enough to harvest a plant or two.

Players were also confused by our clock, not understanding that they needed to perform their required actions within a specific twenty-four hour block of time -- a distinction lost on a player who opens SpellCraft at 2:00 PM on Tuesday, and returns to it at 1:00 PM on Wednesday, and really has every right to expect that he's fulfilled the requirement of "daily play."

Bottom line: if you are going to give something away... just give it away. Furthermore, don't be concerned if players try to game the system. These are retention tools, not a final exam.

We've since patched the system to give an increasing bit of gold each day for 10 days basically for just entering the game, but we are still receiving complaints that the reward isn't arriving, or that it should extend beyond 10 days. We've learned that if you promise players a gift, you must not merely deliver -- you have to over-deliver. Even then, expect a few squeaky wheels. Don't make your job harder with Byzantine redemption requirements!

5. Confusing Store (At First!)

Our experience with creating virtual currency and setting prices was limited when we launched SpellCraft. Our conversions of FaceFighter and Trucks & Skulls to free-to-play monetization let us develop our store code ahead of time, which was an immense help, but we were still flying pretty blind when it came to actually constructing an in-game economy.

SpellCraft has a two-tier currency system, with a low-value currency (gold) earned through gameplay, and high-value gems acquired primarily through in-app purchase. Our initial store had a means of converting gems into gold, but no way to purchase gold directly, which frustrated our players and limited monetization.

We also made a mistake in hard-coding the quantities of each purchase package into our game, meaning that we couldn't adjust prices on-the-fly with a wireless patch, but instead could adjust our store only by resubmitting our game through Apple. We've since corrected both these problems, but launching this way limited our options by needlessly delaying our ability to act on the data we received about player behavior.

Conclusion

Looking back on SpellCraft, it might seem many of our difficulties were related to scope. We did significantly reduce the scope of the project from our original design -- focusing just on wizards, and cutting dozens of ancillary features before writing our first line of code -- but if we had it to do over again, we might have dropped our pet and multiplayer features from the first version of the game, particularly if it would have saved us a couple weeks at the end of the project.

In fairness, those elements help flesh out a pretty lean design, and neither pets nor multiplayer impacted us in the final weeks of the project -- our polishing time was primarily devoted to the onboarding/tutorial portion of the game. Removing those elements would have taken time out of the middle of the project, and probably would not have moved up our end date in any case.

I also don't think that additionally reducing scope would have made us any more likely to catch our crash bug, which didn't originate with our crunch for completion, and eluded detection even with the luxury of the longest and most wide-spread testing program Appy's ever had for a game. Sometimes the puck just gets past the goalie.

If we'd caught that reset bug (and launched on time with Apple support), this would have been a perfect project for Appy. As it was, we got our noses bloodied a bit at launch, but the game has recovered, player ratings are stellar and we have found an audience for SpellCraft.

Even without the benefit of Apple promotion at launch, SpellCraft is off to a faster start than was Trucks & Skulls, and thanks to freemium monetization, SpellCraft appears to have a profitable long tail that we can build upon, as opposed to Trucks, which basically crashed out of the charts following its "Elvis Month" of Apple favor.

Emboldened by this success, we are now a full-time freemium, social/mobile game publisher, with two new iOS games in development that fully leverage our learning from SpellCraft. We're cooking up new spells, taking better care of our virtual pets and moving forward with our plans to accept investment and expand our staff as we stake claim to our part of this new gaming frontier.

Data Box

Developer & Publisher: Appy Entertainment (in partnership with W3i/Recharge Studios)

Release Date: 11/28/2011 (Canada), 12/8/2011 (worldwide)

Platforms: iOS Universal app for iPad and iPhone/iPod touch

Number of Developers: 6

Length of Development: 6 months

Development Tools: AppyEngine (Appy's proprietary game engine), Versions, XCode, Objective C, Ruby on Rails, Interface Builder, Motion, Photoshop,

Other Interesting Information: Apprentices at the SpellCraft School of Magic have cast 58 million spells, cleared 20 million dungeon rooms, purchased 400,000 sacks of fertilizer for their magic plants, and pet their animal familiars 11 million times.