[In this postmortem, Neil Sorens, creative director at Zen Studios, looks back on the two year-long development of October 2010 downloadable title Pinball FX2, the studio's "most successful game to date."]
Pinball FX2 was originally conceived as a title update to Pinball FX, our successful if unrefined debut title on Xbox 360's Live Arcade service. We had several features we wanted to add to the game, and Microsoft was eager for us to support its "NXE" Dashboard update in any way possible.
After much discussion, we agreed to create a full sequel, rather than trying to cram a sequel's worth of features into a very size-limited title update.
As time went by, the feature set grew and the release date was pushed back. Normally, the next thing you'd read here would be, "...and that's when the wheels really started falling off the wagon."
However, as in the case of Toys for Bob spending an extra six months to add the features that made Star Control II a masterpiece, the extra time gave us the opportunity to put in some additions that made a big difference in how the game was received.
After almost two years of planning and development, Pinball FX2 was released in October of 2010, and it's been our most successful game to date.
Although good news for us, it does mean that we don't have all that much on the "what went wrong" side of things, which is always the more instructive/juicy of the two. In contrast to much of our previous work, Pinball FX 2 was based on a proven game, had an experienced team behind it, and used mostly existing tech -- all signs of a low-risk project. And despite a months-long delay, that's really what it turned out to be.
What Went Right
1. We Picked the Right Business Model
Since many potential customers have only a passing interest in pinball games, we wanted to make it as easy as possible for them to experience the game and all of its content.
To that end, we allowed players to download the game for free, to play trial versions of each table (rather than just a single table, as in Pinball FX), and to purchase content à la carte. We also priced content at impulse-buy levels, with a single table costing $2.50 (in the USA) -- a relative bargain given the complexity of each table and the fact that a single table takes a team (designer, artists, and code support) about six months to produce.
However, some of those players decided that they liked pinball and wanted more content -- which, of course, we are happy to provide. Our large and ever-growing library of content also lifts the traditional business model's cap on per-user revenue -- although production of that content is certainly more difficult and expensive than cheat/unlock codes or the infamous horse armor.
Although it is certainly not applicable to many types of games, a DLC-based business model is perfect for this one.
2. We Kept the Pinball FX Community Intact
During early discussions with Microsoft, we agreed that if we did make a sequel, we had to do so in a way that did not alienate our existing customers. We did not want a situation where, for example, Pinball FX owners could not enjoy a new table because it was only available for Pinball FX2, or where one player was still playing FX1 while a friend had moved on to FX2. Not only would a fragmented community be bad for table sales, but a sequel perceived as a money-grab would also damage our reputation as a consumer-friendly developer.
In order to encourage Pinball FX players to make the jump (and ensure new players started with FX2), we had to make sure that there was no downside for them to the release of a sequel. To do so, we planned several key features:
- Free and easy importation of FX1 content into FX2
- Geometry and texture upgrades for FX1 content
- Free download of FX2 platform (no table purchase required)
- Delisting of FX1 from Marketplace, to ensure that new players started in the right place
Implementation of these goals turned out to be quite complex, both for us and for Microsoft, as there had to be dozens of different items on the Marketplace to allow for trials and imports of all nine FX1 tables. In fact, the most notable item in "what went wrong" deals with the problems we experienced as a result.
3. Working Closely with Microsoft Paid Off
In January of 2010, about 10 months before the release of the game, it was essentially in a final state. However, the XBLA team at Microsoft sent us some ideas for new features. These features would greatly enhance the social aspect of the game, they said.
We were pleasantly surprised by how well their ideas fit within our vision and existing design. Normally, when I get a big list of ideas like this from a first party, I find one or two reasonable ones, a bunch related to their latest peripherals or initiatives, and some that are well-intentioned but not very timely.
However, as I went down the list, it was just one great idea after another, all completely cohesive and feasible. In fact, they were good enough to get us to hold off on releasing the game and to spend another nine months on it.
Because pinball, like many games with arcade roots, draws much of its appeal from competition for high scores, Microsoft's ideas centered around how to enhance that competition: a cumulative score for all tables, which extended the competition beyond just individual table leaderboards; a score that included friends' scores and offered avatar items as rewards, which encouraged players to populate their Friends leaderboard (competing with friends is more enjoyable than with strangers); a "challenge" feature that allowed you to send a message to a friend when you beat their score; in-game notifications when you were getting close to your or a friend's high scores; and several others.
These social features they suggested are mentioned by reviewers and customers alike as reasons to keep coming back to the game. We owe quite a bit of our success not only to Microsoft's ideas, but also to the extensive work the company did to accommodate our "pinball platform" model on the back end.
In turn, our willingness to put in the extra time and effort led to Microsoft's inclusion of our game in the Game Feast promotion, which included a few days of Dashboard placement -- a huge boost to visibility that no doubt contributed heavily to our success.
We continue to exchange ideas with the XBLA team and coordinate development as closely as we can with the marketing team.
4. We Did Our Own Marketing and Press Tours
XBLA is a tough market because the competition is extremely strong. Gone are the days of cheap retro ports and shallow original titles. In recent years, Microsoft has been very proactive in bringing quality content to the platform. Today's best XBLA games, like Shadow Complex, Super Meat Boy, Splosion Man -- and too many others to list -- boast incredible style, depth, and polish.
In this environment, it is not enough to create a high-quality game. As most forms of advertising are cost-prohibitive for XBLA developers, the game also has to have enough visibility to attain the critical mass that will sell the game through word-of-mouth. This is especially true for FX2, which relies on competition with friends for a large part of its appeal. If your friends aren't playing, that part of the game's appeal is lost (for this reason, one major goal of the "Wizard Score" feature is to get players to populate their Friends list with other pinball players).
One of the moves that paid off in this regard was to hire a full-time senior PR/marketing guy with extensive experience in the game industry to help promote the game. Although Microsoft gave us a few days of Dashboard placement at release, we had our own PR campaign, from communication with hundreds of gaming sites to visits to major gaming press. By getting our game in the hands of hundreds of sites, we made sure that potential customers searching for information on the game would find plenty of facts and opinions to guide them, rather than a wasteland of press releases and speculation.
It became evident after the initial response to news of the game's existence (thanks, Australian Classification Board!) that one of the biggest potential complaints we had to address was, "Why don't they just release more downloadable tables?" In order to justify a sequel, we had a number of points that we had to communicate clearly, such as the free "platform" approach, the ability to import old content, the aforementioned social features, removal of the 250 point Achievement cap, etc.
Although it was a tall order, by and large we managed to get across the most important points: that it would cost you nothing to try out all of the content and to import all of your FX1 content, and that there were some cool new features to make it all worthwhile.
5. Timely Workflow Improvements
After it become clear that a DLC-based business model was ideal for a pinball game, we restructured our workflow and table scripting system to allow for easier production of tables and integration of new hires.
Once we had several tables' worth of experience under our belts, we had a better idea of how much time and how many people it would take to create and test a table that met players' quality expectations. Based on that knowledge, we established a set process, task list, and schedule that could be applied to each table we created.
We also changed the organization of the team such that each pinball table had dedicated resources, rather than resources being spread among multiple tables. This change allowed for a greater focus on each table and a sense of responsibility for the quality of each piece of content.
We created a custom scripting system to replace Python, the scripting system we used for the first batch of FX1 tables. This system simplified the scripting environment, making both the learning and development processes faster, and made the scripts run more efficiently on the target hardware.
These changes took place not long before pre-production on FX2 started, preventing the inefficiency that results when content development and workflow changes are happening simultaneously. This improvement paid off during the production of content for FX2, as three of the four initial tables were created by newly-hired designers and were completed well before the launch of the game.
It also allowed us to complete production of the Marvel table pack in time for release during the critical six week post-launch period.
6. Partnership with Marvel
Speaking of Marvel tables, our partnership with Marvel was also critical to the success of the game. While we had created tables based on licensed properties before (Street Fighter, Rocky and Bullwinkle, Ninja Gaiden Sigma 2), all of them were one-offs, and none were of the same stature as the Marvel brand, which is known worldwide in games, movies, comics, toys, etc. and has seen a major resurgence in the last decade.
With Marvel, we signed a long-term deal for multiple tables, bringing some badly-needed brand recognition to the stable of FX2 content.
With the release of the Marvel four-pack in December, about six weeks after the launch of FX2, we cemented in players' minds the notion of FX2 as a platform for new premium content, rather than just a clearinghouse for the old stuff.
The Marvel brand also opened some doors that may otherwise have been closed. Press who were not particularly interested in pinball alone were more willing to dedicate time and copy to a product with an appealing brand, especially since these tables were something other than obligatory, half-baked movie cash-ins. And we gained new customers who were on the fence about pinball but liked the Marvel characters.
The Marvel pinball table pack quickly became our best-selling FX2 DLC (and is on pace to surpass even DLC available for years starting with FX1) even though the bundling of the tables meant a higher total price than that of single tables (200 to 240 points, $2.50 to $3).
Although Marvel is no doubt one of the tougher licensors to please, because of how strictly it protects its properties, the work that has gone into the partnership paid off -- and also gave us valuable experience in working with brands of this caliber.
What Went Wrong
1. Supporting Old Content Was More Painful than Expected
Since our old tables were designed around an outdated physics model, to support them in FX2, we had to create new table geometry based on the new engine. The tables also needed to be re-textured, as they had not aged well.
Several of the old tables had also been outsourced, or created by designers who had since left the company. Because we had moved from Python to a custom scripting system (see "timely workflow improvements" in what went right), the old tables needed to be re-scripted. Without the original scripters available, this task became extremely time-consuming.
Implementing and testing the importation process was also difficult. Because of the way the Marketplace works, we had to prepare a title update to FX1 and work with dozens of pieces of content, while keeping the process simple for the player.
Furthermore, with the addition of a "Superscore" that combined scores from all tables, we had to rebalance the scoring systems of the old tables against each other and against the new content. Not only was this a lot of work, but it also caused problems post-launch when we had to update the Rocky and Bullwinkle table and reset its leaderboard.
Finally, we were not able to add many of the Pinball FX2 features, such as local multiplayer, local high scores, split-screen play, etc., to the imported FX1 tables because of their legacy design.
This issue also created a dilemma in how we sell that old content, which appears alongside the latest, greatest tables in the store at roughly the same price. We wanted the tables to be available for those who wished to play them, so removing them entirely was out of the question. We did not want to price them lower, because that would encourage players to buy the earliest, lowest quality tables we made, giving them a less than optimal experience. And yet it felt wrong to charge the same amount for old tables that were clearly inferior to the new ones.
With no perfect solution feasible in the available time frame, we ultimately decided to charge about the same price for all content, to place the newest/best content higher in the list, and to hope that the demo available for each table would help players make a good decision on what to buy.
Although this problem still persists today -- we get occasional feedback, questions, and complaints about it, but less than we expected -- we probably made the right decision, though it is impossible to say with certainty. We hope to address it eventually with a combination of feature upgrades to the old content and interface features, such as player ratings for each table, that will help to guide players to the content they will most enjoy.
Would we still make the decision to allow the importation of old content today? Yes, because it was such a critical piece of our pitch to FX1 customers. Plus, the additional content, although lower in quality than the newest tables, helped tide players over until we could start releasing new content again this year.
2. We Didn't Have Enough Content to Release Regularly After Launch
After the Marvel pack in December, we had nothing to release until the Mars table (a port of content previously released on the PlayStation 3) in April. The first completely new table, Fantastic Four, did not arrive until mid-May. All this despite the fact that we're employing the largest pinball design group in any company's history!
In an ideal world, we would have had enough content to prevent that five-month drought, especially since our business model is based on delivering new content.
However, the lead time for new content is several months long, and pinball designers don't grow on trees; even if they did, it would have been quite a gamble to hire the necessary people to create that much content long before we knew how the game would fare.
Since the game's release, we've expanded our pinball group -- though we still need to expand further -- and staggered the expected release dates of the content, with the goal of a table release every month and a multi-pack every now and then.
3. Budget and Schedule Busting
As mentioned previously, we ended up extending project development by about nine months. Naturally, this long delay cost a lot of money beyond what was budgeted and delayed an expected source of revenue.
In addition to implementing the features Microsoft had suggested, we needed to redesign (again, with feedback from Microsoft) and rewrite the entire UI to support those features, as the basic system we had in place -- something we had figured would be good enough to get the job done at the start of the project -- was not capable of doing so.
Although this kind of delay was definitely a risk, and such cost and schedule overruns might represent a hardship for some studios, the new features represented the kind of opportunity that a private, self-funded developer like Zen can pursue without the worry of angry shareholders or a trigger-happy publisher. As such, this issue was more of a footnote for us than a major concern.
4. Updates for Individual DLC are Cost-Prohibitive
Bugs and balance issues are a small but recurring problem for us, and they are difficult for us to deal with efficiently. Because our pinball tables are enormously complex, and because it is impossible to test every possible event sequence and bounce of the ball, we, like pretty much every other developer in the world, sometimes unknowingly ship product with bugs.
What makes our situation especially problematic is that each table we release is essentially a self-contained game, more prone to bugs than most types of downloadable content. Although the tables tend to sell reasonably well, it is still prohibitively expensive to have a patch for a single table tested and certified by Microsoft, since the cost of testing/certification is very high relative to the revenue from a single piece of DLC, and we already paid it once the first time around. Thus, we must let some bugs live on until we can release an update that addresses multiple tables at once.
The upside of that approach is that all the numbers add up in a way where it still makes sense for us to produce this content at current prices. The downside is that players often have to wait for quite a while after a table's release before problems are fixed.
5. It's Pinball
Although PR generally went well, there were a few notable failures. IGN didn't even bother to post a review of the game at all, a symptom of a more widespread issue: many outlets simply weren't interested in writing about pinball. Other games that were released during the Game Feast promotion (and indeed, most high-quality XBLA games released during any time frame) have much more numerous reviews on Metacritic:
- Pinball FX2 -- 24 reviews (88 score)
- Comic Jumper -- 45 (74)
- Super Meat Boy -- 50 (90)
- Hydrophobia -- 58 (59)
We saw a similar trend in our review scores. One of the biggest complaints we saw in reviews of the game (and in reviews of subsequent content releases) was "It's pinball."
This type of complaint -- and the lower review scores and reduced coverage -- is understandable, given that only a fairly small percentage of gamers have any interest in the genre. Pinball games are never going to be huge blockbuster hits. It is quite common for trailers of our Marvel-themed tables to receive more dislikes than likes on Youtube, with comments generally along the lines of "Pinball? WTF is this? How come this character isn't in MvC3?"
The other side of this coin is that because pinball has been an underserved market, our efforts to revitalize the genre with inexpensive, high-quality, and feature-rich content has attracted a core of very enthusiastic and supportive fans who buy every table we release (thanks!)
Pinball FX2 represents a high water mark for the company in terms of game quality and depth, review scores, positive customer feedback, and sales.
A variety of factors contributed to this success: experience with two previous pinball games, a good working relationship with Microsoft, our partnership with Marvel, an extensive pre-launch PR effort, good feedback from our players, a willingness to take risks when appropriate, and correct identification of the most important attributes of the game.
Despite the delays, hiccups, and assorted problems, we're proud of what we've achieved and hoep to build on our success with a raft of new content and features over the coming year.
I'd also like to give a shout out to Ryan Peterson, our biz-dev representative, who was the guy behind the scenes making all the right decisions and partnerships happen.