4    Fair Play: Scaling Teams for Impact

Most academic games are made by a few students or faculty members working together, perhaps contracting with an artist. This approach works well (enough) for making technology demos or answering (some) research questions. But what about taking projects to market, scaling them in situations in which they don’t have “captive” audiences (e.g., in school or paid for in a research study)? Can we make games for impact with the production quality and processes of entertainment games? How can we set up projects from the beginning to be scalable? How does management in such a project look?

This chapter details experiences in building and scaling a team that spans academics and commercial developers from large studios and small independent teams. It articulates a prototype-driven approach to design—suitable to a noncaptive audience with multiple unknowns.

The narrative begins with a grant awarded to Dr. Molly Carnes to explore how digital games might remediate implicit bias. Dr. Carnes, director of the Center for Women’s Health Research at the University of Wisconsin–Madison, studies how implicit biases and stereotypes color our actions, especially in relation to women and minorities. For example, scientists rate papers and résumés from applicants with female names lower than identical applications with male names, perhaps because people associate being a quality scientist with being male (Moss-Racusin et al., 2012). These biases are inequitable and detrimental to society; every year qualified women unnecessarily leave the sciences, which leads to shortages of skilled scientists, to speak nothing of the ethical considerations.

Dr. Carnes’s grant hypothesized that a game based on her gender bias workshops could mitigate bias at scale. Conducting face-to-face workshops is expensive. Carnes’s team hypothesized that an online game could deliver role-playing experiences more cheaply. Creating such a game for faculty and graduate students was exciting, if not harrowing (for a lengthier description, see Paiz-Ramirez et al., 2011). What genre would work for scientists, most of whom are nongamers? Why would scientists, who likely think of themselves as fair and objective, play?

We wanted equity issues in our portfolio, and the chance to work with Dr. Carnes was exciting. Our team had not yet tackled issues with a driving social purpose, which was personally disappointing. We wanted to champion these issues within the Wisconsin Institute for Discovery (WID), which had few senior female, Latinx, or African American scientists. Social justice was not a commitment shared by all team members or, necessarily, by WID as a whole, as it would be in education. This project created an opportunity to define group values (see table 2.1).

Developing a Learning Game for Science Faculty

A game about bias raises production questions as well. Could we represent people with a range of emotions (e.g., surprise, disdain, guilt, anger)? What art style works for this audience? This project began before the Gamergate controversy, but we were aware of potential backlash.1 We also had concerns common to serious games projects: how to give the player real choices while also ensuring that learning occurs, how to bridge the cultural differences across academics and developers, and how to manage player failure and transgression in sensitive areas. These concerns color almost every games for impact project to some degree, but they were especially salient here.

We committed to a rapid prototyping ethos of “failing early and often,” rather than overdesigning on paper and hoping it would work out in the end. We planned to uncover design challenges through cycles of design, testing, and reflection, rather than arguing endlessly. Second, we committed to involve developers from diverse backgrounds. We wanted the game to be built by people with lived experiences of bias events and an active interest in the topic. We wanted to advance the careers of populations underrepresented in games and science.

Prototyping

We built a technical prototype to test early concepts while the full team came on board. The first idea to test was Research Lab Story—a takeoff on Game Dev Story, an iOS game made by Kairosoft in which the player manages a game development studio. We imagined that players would manage a lab, deal with conflicts between nonplayable characters, and see how not attending to bias incidents could hurt the lab’s productivity. We liked dealing with bias in utilitarian terms; players could “experience” the impact of bias events on team morale and productivity, rather than arguing on a purely moral basis. We also liked that the game would simulate research labs, which could have secondary positive effects (contemplating how research labs work). Narrative events could be taken directly from research literature that has documented the impact of bias.

User studies of our first prototype (see figure 4.1) revealed the following (see Paiz-Ramirez et al., 2011):

  1. Scientists liked seeing their work environment as a game. Playing a game set in a research lab interested faculty members (much as game developers like playing Game Dev Story). We entertained fanciful scenarios, but a realistic setting ameliorated the concern that scientists would find the game irrelevant (see Keller, 2009).
  2. Scientists liked the objectives: writing papers and producing research.
  3. Bias was presented clearly. Players inferred conflicts on-screen between characters.
  4. Players min-maxed the game to the point of distraction. Players employed strategies that maximized research production to the point of abusing lab workers. Anything else (including bias) became secondary.
  5. Players ignored bias events. Bias events were important but not crucial to lab productivity. We would need to either rework the underlying model or amplify the severity of bias events. Otherwise, bias would fall into the background of player experience.
  6. The “God game” presentation detached players from the experience. We wanted players to feel the emotional sting of bias, for aesthetic reasons (why miss an opportunity for games to provide vicarious experience?) and practical ones (emotionally identifying with bias is important in overcoming it; Todd et al., 2012).

Figure 4.1

Prototype 1: Research Lab Story.

We kept the setting and goals but shifted the player perspective and interface.

This target audience was not literate with games. User tests revealed that most scientists were uncomfortable learning by trial and error. Faculty often sat awaiting explicit instructions and were unable to discern the “deeper meaning” of the game. In response to the user interface challenges, we developed design pillars: always have a clear goal on-screen, simplify navigation and choices whenever possible, and put the game content front and center. To this last point, in “Theme Is Not Meaning,” Soren Johnson (2012) described how it is important to distinguish between a game’s theme (for example, the fantasy of World of Warcraft) and its meaning (evolutionary battle between designers and players in World of Warcraft). We needed theme and meaning to be consistent.

Emphasizing the Story in Research Lab Story

With Fair Play, we needed to consider simultaneous projects and a studio model with dedicated services across projects, including a 3D art pipeline with modelers, animators, texture artists, and postproduction staff. We hired industry-trained programmers and level designers to build robust systems capable of scaling. Seemingly overnight, the team grew from about 12 faculty members, students, and developers to about 30, including 11 developers (5 artists, 6 programmers), 2 postdocs, and 12 graduate students. Under the advisement of Dr. Rock Mackie, we hired industry veteran Brian Pelletier to lead development. Brian was highly regarded by students and staff, and we wanted someone with experience in managing development teams to help spin out companies.

We recruited Erin Robinson to be the design and production lead for Fair Play. Among the many reasons we recruited Erin was that her first game, Puzzle Bots, featured players controlling robots to solve puzzles that were reminiscent of rats trapped in a psychology experiment maze. Parts of the game were motivated by psychology experiments that she studied as an undergraduate. Puzzle Bots oozes personality and charm, something that we can say of few educational games.

After reviewing the research literature, prototype, and feedback, Erin developed several iterations of game ideas (see figure 4.2), but nothing stuck. We spent weeks discussing ideas. In retrospect, I realize that I hadn’t sufficiently empowered Erin to lead the team. I wanted a relatively flat management structure, but instead I was getting design by committee. Frustration built because of the lack of clear roles.

Figure 4.2

Fair Play storyboards by Erin Robinson.

Erin next developed a seventy-slide series of mock-ups of a play experience. Unlike design documents, design mock-ups can provide pacing, timing, and tone so that the reader can get a feel for the game experience. She described this approach in Robinson (2012):

If you’re lucky, a series of mock-ups like this can do more than explain your goal: it can energize and inspire the team to do their best work. These particular mini-presentations were popular enough that sometimes a few of the senior faculty would sit in on our meetings. The goofy placeholder art and the informal nature of the presentation invited questions and discussion. It was a real boost for everyone—and reminded us that we were making something fun.

The mock-ups read like a comic book story of Jamal, a young professor attempting to establish a research lab. Erin discarded related ideas, such as like gender bias, and focused on a narrative of building a career with a secondary theme of facing racial bias. Erin targeted key environments where bias incidents occurred (such as academic receptions), which later became chapters. The mock-ups addressed a majority of concerns, they were feasible, and we wanted to play it.

Erin’s redesign immediately struck me as the right approach, but not everyone agreed. Some felt ownership of previous ideas or wanted to tweak the previous game. The team had, until this point, worked by consensus, using an agile process of two-week sprints with weekly reviews. In reality, there was no consensus, but there were strong wills, conflicting views, and then leadership weighing in on what to do. In an effort to facilitate team ownership, I had been passive in leadership style, which was a mistake. Regardless, the strength of Erin’s presentation made the next steps easy, as it was better than what we would have come up with.

I started to see flaws in my management approach (see “Building Academic Game Development Teams” later in the chapter). Roles and responsibilities were not sufficiently defined for Erin to truly set direction, direct production, and lead the team. Any project has multiple overlapping forms of power, status, and control, and these are even trickier in academic contexts, in which freelancers, advisors, temporary staff, students, and professionals are working together. The key takeaway, which probably applies to academics more generally, is to ensure that all members of the team have clear lines of responsibility and domains that they control. Even the most egalitarian models (such as Valve’s design cabal) typically give ultimate authority to leads in a given area.

Games Come Together: Storyboarding, User Stories, and Vertical Slices

The game that evolved, Fair Play, tells the story of Jamal Davis, a graduate student at a large midwestern university. As Jamal negotiates graduate school, he confronts bias incidents. Fair Play seeks to induce players’ empathy for Jamal through their observing and experiencing these incidents (drawing from social learning theory applied to games; Gee, 2003). Bias incidents begin with his arrival on campus, where he is mistaken for an athlete, and continue to the final chapters, in which he confronts allegations of plagiarism. These bias incidents are recorded in an almanac for review. On the basis of our feedback, we put players on relatively constrained rails so that they were guided throughout and confronted at least two bias incidents within five minutes of playing.

With the core idea created and the team coalescing (and with her own game, Gravity Ghost, underway), Erin handed leadership to Dennis Paiz-Ramirez, a graduate student. Dennis further focused on deepening player identification with Jamal, creating triggered events to elicit bias, and polishing to deepen player narrative engagement.

We observed differences in the ways game designers and medical educators approached learning. Game designers’ instincts were to immerse players in bias situations to induce emotional reactions. We wanted players to reflect on those reactions. Medical educators were concerned that exposing participants to bias incidents could produce deleterious effects. The concern here is that exposing people to antisocial effects in the game could be a negative stimulus, thus creating further bias. If games are going to include transgressive play, learning through failure, or exploring consequences of actions, designers will need to have freedom to explore such ideas and give a degree of trust to the player, which many are uneasy in doing.

We ultimately introduced bias events to players in a relatively benign way, with clear markers of a “bias radar” (red exclamation marks) to signal that the player experienced bias. By labeling bias detection “bias radar,” we hoped to explicitly encourage players to practice identifying bias. Players then practiced reflecting on these biases and naming them. The final chapter, in which Jamal is accused of plagiarism, addresses bias in the direct, emotionally intense way we wanted.

Readers will observe an evolution in art style. We loved the stylized “cute” characters (large heads, larger eyes), but they did not appeal to scientists, who perceived them as childish. The scientists reported, however, being impressed by a 3D game with motion captured animations, which signaled a “real game.” The game was set on the UW–Madison campus (although it isn’t explicitly named), which provided a Easter egg for locals. They could wander through local landmarks such as the Memorial Union Terrace and WID. Future games might explore using aesthetic approaches, such as setting the game in a highly fictional setting, or including more extreme bias events.

Figure 4.3

Fair Play scene set on the University of Wisconsin’s Memorial Union Terrace.

Research results of Fair Play    Playing Fair Play reduces bias for players who show high empathy for Jamal. In an experiment (see Kaatz et al., 2017), participants were chosen either to play Fair Play (experimental condition) or to read a text (control) before taking an implicit association test (IAT). High-empathy players who took Jamal’s perspective displayed the lowest implicit bias. Many of these high-empathy players were women. We don’t know whether this was due to play style or to the experiences of women facing similar biases in the sciences, or to some other factor. In retrospect, we probably should have designed a female character, given that most women prefer to play female characters in games, but it was encouraging that the game still worked best for female players (Yee, 2017).

A number of follow-up questions could arise. Does Fair Play work best for female players because female players in academia are more empathetic? Were they most receptive to this message having perhaps experienced similar bias events in their careers? Is there an underlying phenomena in game-based learning in which people with the most knowledge or direct experience interpret game events more deeply than do those with fewer underlying experiences?

Fair Play has won awards and is used in training. It won the People’s Choice Award at the 2012 Meaningful Play Conference, and Dr. Carnes’s team’s won the Adolphus Toliver Award from the National Institutes of Health in part for this work. Fair Play situates experience and enables and provides a safe context for players to discuss experiences and make connections to their lived experiences. Fair Play workshops across the country use the game to introduce bias, and then players draw from game play in discussions. Workshops include vignettes and leverage PDFs included in the game.

This success suggests limits for using games to scale ideas without very careful consideration of context. Many grant programs hope that games for impact will scale because the content will be more interesting presented as a game (and, I suspect, many program officers are less interested than the public in entertainment games generally). There is little evidence that people will play Fair Play (which is freely available) without incentive or coercion. Fair Play could replace required video training sessions. The issue, then, is that most faculty want efficient, quick training, and administrative leaders want to minimize legal exposure (and perhaps address the problem), whereas advocates want effective programs.

Developing expertise in a domain often leads to new opportunities, and our team’s expertise in implicit bias led to a collaboration with the University of Oklahoma to build an implicit bias game for the Central Intelligence Agency, funded by IARPA, the Intelligence Advanced Research Projects Activity, which is responsible for forward-leaning research projects. This project adapted Fair Play to the CIA’s context, reusing the underlying basic research and knowledge gained from prototyping, but retooled for the CIA. Working with CIA agents was a fun change of pace for team members such as Meagan Rothschild, whose expertise is in early childhood education. The game, based on a prototype directed by Dennis Paiz-Ramirez, won Best Business Game and Best Adaptive Force Game at the Serious Games Showcase and Challenge at the major military training conference I/ITSEC 2013.

Building Academic Game Development Teams

A vision for the Games+Learning+Society (GLS) Center emerged to research and develop games with a broad audience, demonstrated impact, and supported scale. Whereas most groups focus on either research or development, or either education or entertainment, we wanted to explore intersections among them. The center-wide research questions in this phase were as follows: Can games diversify income streams to the university? How should games fit with the teaching mission of a university? How do we leverage games to amplify academic impact? These questions also drove the work at our peer institutions. The following sections detail opportunities and challenges for centers trying to innovate in higher education, with implications for higher education more broadly.

Game development within academic centers. Most research labs and groups organize around tentpole projects central to the group’s mission and identity. Our identity was researching and developing science games to increase participation in science, with satellite projects from learning analytics to game design. A key decision for any group is whether or not to develop games internally. Five strategic factors drove our decision to scale up internal development:

  1. A development team is an asset for research. Studying learning through games (from analytics to identity formation) requires games to study. On-site development teams remove the hassle of finding partners, aligning values, and instrumenting for research. Internal teams remove other faculty needs to manage student-developers.
  2. A development team is a recruitment tool for students. Students studying game design (or the mindset of developers) can leverage an internal team’s connections with industry.
  3. A development team is a teaching tool. Students entering both academics and industry benefit from experiences on professional teams. Professional developers provide enculturation, training, and contacts.
  4. Maintaining multiple projects enabled the development of (a) a full 3D art pipeline, (b) a professional web team, and (c) a community outreach team, which were assets. These functions are difficult to fund through government grants (which prioritize research) but are valued by researchers. They require roughly $1.5 million annually in projects to support.
  5. Interest in learning games was accelerating, and we had an opportunity to make an impact through research, products, and companies (see Squire, Gaydos, & DeVane, 2016). Hype for games has waned, but public interest in a technology, paradigm, or approach can be leveraged to diversity funding streams.

These drivers suggest reasons to scale up from a lab, but scaling introduces challenges.

From a Research Lab to a Research Center

Teams of fifty or more are organized differently from teams of ten or twelve, and most academics do not have experience working on large teams that require full-time managers (see table 4.1). Research lab cultures emerge through sustained, directly collaborative work (analyzing data, designing materials, or writing); decisions can be made by consensus and disputes resolved with everyone present. A research center culture requires more top-down intervention and layers of management. Most academics are not experienced with layers of hierarchy and disputes involving layers of management. The work of a research lab falls largely within one’s own skill set, whereas game development (particularly once 3D pipelines are involved) extends beyond one person’s skills.2

Table 4.1

Contrasts between research labs and research centers

Research lab

Research center

Size

2–12 (small group)

20+

Faculty role

Maker-manager

Manager

Primary management technique

Sustained interaction

Vision setting, supporting, and enabling

Layers of management

1–3

3+

Managing growth    Despite our best efforts to manage growth, we experienced pain.

Our strategy was to encourage students and staff to form companies so as to stay around twenty people. Rock Mackie, who had successfully launched multiple companies, advised us to (1) encourage interest groups to form around projects, as those groups are proto-companies, and (2) staff teams so that core business roles (CEO, COO) are fulfilled. Just as Filament Games originated through staff (Dan Norton and Alex Stone) collaborating with students (Dan White), we hoped that more teams might form. Two more teams spun out of the lab this way.

Game management literature distinguishes between teams under and over 15 people (see Spaulding, 2009). With groups around 10 people (which is also how academic labs look), leaders split time between “making” and “managing.” Smaller companies have a 15:1 ratio of makers to managers. In contrast, a midsize game company that runs multiple projects will have 70 or more people. Here, the group leader should be managing, not making. Large teams typically have a 10:1 ratio of developers to managers, so a group such as GLS can have 4 or 5 full-time leads without specific production tasks. Enter middle management dynamics. To quote Spaulding, someone needs “to devote their full time and attention to department-management issues like career management, scheduling, hiring, and training and mentoring new leads” (2009, p. 32).This is not something that academic managers are experienced in doing.

Academic Game Development Production

Creating and maintaining a culture cutting across any large group is difficult. Our challenges were in spanning research (both education and science) and development (both independent and AAA), including fault lines within these sectors.

Scientists and Gamers versus Educators

Within academics, game studies (the study of how and why games are engaging), learning sciences (how people learn), educational areas (e.g., science education), and scientific domains (e.g., disciplinary fields in WID) have different cultural norms. In fact, there are subfields within those, but we will hold off on that for now. A weird fault line emerged between those identifying with science and games (heavily male-dominated fields) as compared with education (which can be gendered female). Participants were of different genders, but patterns formed that aligned roughly with gendered ways of constituting acceptable talk, ways to disagree, and ways to behave. Postdocs trained in other fields exacerbated issues of clashing value systems and cultural norms.

AAA versus Indie Game Development

Seasoned AAA developers tend toward waterfall development (a linear process of progressing from step to step), well-defined roles, and hierarchy. In general, indie developers lean toward prototype-driven development with minimal documentation, flatter hierarchies, and riskier ideas. The “artists versus programmers” issues that define many teams were replaced by different values about games altogether. Academic development falls between indie and AAA development. As with indie development, academics ask, “What should be built and why?” Academic games change direction during development, and marketing departments don’t dictate features. Like indie games, academic games are largely driven by passion. However, unlike indie games, academic games exist in larger institutions with HR policies and norms. The norms of academics are unlike those of AAA games. AAA jobs are highly competitive, eighty-hour weeks are routinely demanded, and employees are terminated easily. Try applying the management techniques of AAA in academia and see what happens. Thinking in terms of an integrated technology that understands academic and contract development works better; game development requires highly functional teams with a shared understanding of work processes, flows, and personalities and trust. If a good game dev team who knows how to work with academics forms, the last thing you want to do is lose it.

Individualistic versus Team Cultures

The extent to which academia is individualistic and built on a “star system” becomes evident when compared with game development. Academics are judged as individuals and, in even the most collaborative team environments, become associated as authors of ideas. Game developers, in contrast, work in teams. Work goes into enculturating identities as team members. Individuals taking credit for a team effort can poison game development. PhDs and postdocs look for ideas to own.

Many professors feel that their groups have a team culture because the culture is built around them. Some groups surely have a collaborative culture, but most academics work on a star system and have labs named after the professor to reflect that.

Commitment to Growth

Managing employee growth is a full-time job in game development, which turns over technology about every five years. Companies invest in individuals as leaders and create career ladders that exceed the pace of academic labs. A game employee might grow from entry programmer to advanced programmer within five or six years and earn multiple promotions and increases in job responsibilities. When our studio director developed advancement ladders and growth curves for the university’s HR department, they thought we were crazy. Academics do (hopefully) commit to students’ lifelong growth. Academic advisors look out for students (again, hopefully) in ways that are atypical in game development, although informal networks of developers are powerful (and difficult for outsiders to crack).

Different Workflow Rhythms

A key challenge to game development in an academic context is the rhythm of workflow. Academic calendars (conferences, meetings, classes) require teams to meet once or twice per week and work independently, which is not good for game development. In contrast, most game developers have daily stand-up meetings, work concurrently, schedule daily play sessions, and work at a faster velocity. Game development requires making a million small, interrelated decisions that come together to form user experience. If I left the office for two or three days, projects drifted as they responded to new challenges. The idea that a graduate student would “disappear” for a week or two is anathema to game design. Students who are integrated into core development need to be in the office twenty or more hours per week (and hopefully compensated on overload).

Lessons for Teams

To help the team grow, I eventually took leave from UW–Madison to work at WID full-time. My presence on the ground all day, every day provided leadership clarity and feedback for teams, which I believe is required to lead such a center. Academics considering similar efforts might also do the following:

  1. Expect conflict. Managing interpersonal conflict is not a routine part of most academics’ jobs. Passion-driven creative enterprises generate conflicts over vision, priorities, or personalities. Resolving them is a large part of the job.
  2. Empower and rely on strong team leaders. Game development benefits from strong leadership. There are many viable answers to development decisions, and many of them end up being wrong. Reasonable processes and good leaders manage this (Spaulding, 2009). A collaborative, flat model may emerge (such as Valve’s design cabal), but the number of moving parts and sheer number of decisions to make requires delegation.
  3. Promote team leaders on the basis of disposition. Team leaders can come from academics or industry, but they must be good at integrating perspectives, ensuring that voices are heard, and managing clients. Leaders must have expertise in game development and in the subject domain and a clear sense of research goals, methods, and outcomes. We rotated team leaders for professional development and group cohesion.
  4. Empower self-sufficient subdisciplines. Teams that spanned projects (web services, art, analytics, Institutional Review Board procedures, statistical analyses) were slammed and constantly balanced competing demands. Designating a strong, self-managing team lead for such disciplines to meet with project teams, plan services, and allocate team time is essential. Such groups are best when they self-manage, which requires exceptionally strong and trustworthy dispositions.
  5. Design for redundancies. Games for impact teams include specialized skills, and losing any one of several employees can tank a project. Performing work for private funders, from foundations to companies, is demanding because if a funder wants a feature (such as web functionality), the funder expects it to be done regardless of team composition or university HR policies. Developing secondary skill sets, such as training a game programmer to also do analytics programming, helps.
  6. Help teams prioritize efforts. Games for impact tend toward too many things at once. A good game for impact must simultaneously (a) interest target users, (b) appease funders or partners, (c) satisfy domain experts, and (d) generate research results. Design decisions might suffice one condition (e.g., satisfy domain experts) at the expense of another (appeal to users). Leadership helps team members understand priorities, particularly with regard to resource allocation.

Managing teams or, more accurately, managing the managers of teams, is time intensive, and academic developers must be careful to have clear management, roles, priorities, and, critically, time, such as twenty or more hours of face time per week.

Game Development in Academic Institutions

Game centers are at the forefront of attempts to remake universities for the digital age. Viewed from the top down, universities need to evolve in a context of decreased government funding, increased competition for students, and competition from nontraditional entities and support games for impact centers in testing ideas. Viewed from the bottom up, academics see game development as opportunities to attract funds, support research goals, educate students, and maximize impact. Efforts to launch innovative centers such as game centers will face three types of problems: (1) accepting money on contracts, (2) spending money to develop games, and (3) investing in proactive business development. Anyone seeking to grow such a center might use this list when negotiating with administrators.

Problems Accepting Money in University Settings

Private contracts are challenging for most universities, which do not process them in a routine way. Most universities have policies in place for patents, copyrights, and indirect rates but fewer processes for royalty negotiation or contract arbitration. Negotiating contracts is laborious and can involve months of negotiations. As an example, one contract negotiation with a UK-based company became stuck over which country’s laws would adjudicate breach of contract. The university does not live on contract work, so honestly it was easiest for legal staff to let the project die. If a negotiation stalls or fails, no one in the legal department loses their job. Jobs are lost, however, if a project is approved without proper safeguards, so as a whole, the university is incentivized to dismiss them. It is difficult for a university to assure stakeholders that private grants and contracts will be processed in a timely fashion with sensible copyright-sharing and indirect cost policies.

Problems Spending Money for Game Development

Spending money for game development is equally difficult. Hiring game developers challenges an human resources processes for nontech companies. University job titles and descriptions make little sense, especially for artists, who are hired as instructional support specialists. Game developers may not have college degrees, which might require administrative approval. Pay scales are significantly less than even in private nonprofits with hard pay ceilings (such as $62,500 for senior programmers at UW–Madison, which may have changed). Game development frequently requires contractors for short-term (e.g., sound design) or emergent (e.g., technical artist) needs, whereas in public universities, paying contractors is laborious, for good reasons. Funds for other unforeseen expenses, from power tools to subscription services, can be difficult, if not impossible, to procure. Partners in private contracts have little patience for bureaucratic hassles.

Investments in Business Development

Working with private entities requires spending money to (1) identify and procure business leads, (2) develop and maintain relationships (which means unplanned travel), and (3) follow-through on partners’ requests. For example, one partner required us to fly to the Caribbean and present work before its executive board on a week’s notice, which the university was not wild about. Private funders, from foundations to companies, do not care about an organization’s rules about expenses; if they’ve given money for a job, they expect it to be done if it’s legal and paid for. Setting aside money and negotiating expectations with administrators is imperative.

Supporting a Skunkworks through Private-Public Partnerships

Our group developed a private-public partnership with the Learning Games Network (LGN) to support game development. This arrangement is explored in some detail because more research labs are creating nonprofits to accept gift donations and perform work that falls outside the scope of the university. Our model, established through a memorandum of understanding with the university, enabled the university to partner with LGN to support games for impact. The university led research grants, with subawards to LGN, and LGN led development projects, with subawards to the university. LGN agreed to rent space (a renovated nightclub affectionately called “the purple building”) and hire administrative staff. We continued the vision of the Morgridge Institute for Research but in a lean, creative, and sustainable way fit for games.3

A professional game development studio in the heart of the UW–Madison campus was a boon for creativity. It drew faculty, game developers, graduate students, postdocs, and undergrads from different disciplines to work side by side in integrated teams. Unlike Morgridge or university spaces, which have policies appropriate to large organizations, such as which employees get desks, LGN was flexible and responsive. Undergraduate programmers set up camp and joined teams. We prohibited hoarding of resources (a common university practice) and created an ethos of “use it or lose it.” We openly encouraged a culture of “asking for forgiveness instead of permission.”

The space, location on campus, mix of talent, and culture fostered creative collaboration. For example, let’s return to the example of AnatomyTable. It began when Dr. Matthew Berland was awarded a National Science Foundation grant to build interactive table exhibits for the New York Hall of Science. The academic team ran into problems with contractors, and within days we had cleared space and built a makerspace. This table led to the aforementioned interactive table for veterinary care. Dr. Lonie Salkowski saw this table and improved upon it by developing an interactive mannequin (working with LGN) for teaching anatomy, which was given a grant for further development by the university. A game studio on central campus promoted the group internally and externally. Graduate and undergraduate students cited it as a deciding factor in their coming to Wisconsin. Alumni valued it as an example of applied work on campus. We hosted all kinds of such events, from ESPN Game Day to department retreats.

Flexible, unallocated space is important to innovation but nearly impossible to find at universities. For example, we created an audio recording studio with vintage equipment to attract musicians. The makerspace for artists to physically sculpt brought Jordu Schell, creator of the aliens for the film Avatar, to hold a daylong workshop for local artists. Artist Mike Beall created a Baby Groot for Nicole Perlman, who cowrote Guardians of the Galaxy. We encouraged these as hobbies, understanding that some might have tangible benefits later. Key to this space (and, I argue, to any good lab) is a large open area in the middle for hosting tours, play sessions, and design jams. We wanted end users to have a physical presence in the space so that they were in our minds (as well as available for “Kleenex testing,” the quick and dirty kind of play testing that you throw away afterward).

Taking a do-it-yourself approach to running a research and development lab was an investment in ourselves not easily understood across the university. When we started the experiment of renting the purple building, we had funds to cover costs for two years, which, at over $200,000 per year, was nontrivial but enough to cover the lease. Yet we were routinely asked, “What if it doesn’t work?,” as if a lab should continue indefinitely, which for me, highlighted the difference between an entrepreneurial, innovative spirit and a university. Should something living forever be a precondition to trying anything?

We tried this arrangement for two years to test the idea. We figured one of three things would happen: (1) the group would thrive and perhaps require even more space; (2) we would quit, either failing to generate enough income or just growing tired of it; or (3) the university would value these activities and provide on-campus space or funding.

A mix of all three happened. The arrangement created new business in the form of multiple contracts. Yet the stress of working with a separate nonprofit mounted. Accounting and management issues created liquidity issues. LGN was a small organization with an operating budget of a few hundred thousand dollars that grew overnight to more than twenty-five people and millions in annual expenditures, including federal grants. Federal grants bring with them substantial cash flow issues and require extra oversight. We spent a year reorganizing the business, responding to a mandatory federal audit, and fighting administrators at the university who did not want us to exist. This became my life for about two years.

Innovative work requires supportive leadership. At the end of our lease, WID director David Krakauer and Dean Julie Underwood left their positions, and within weeks, the university decided to audit LGN, which had already completed a mandatory federal audit. For six months, all payments to subcontracts at LGN were paused, which shut down LGN-university business and closed the purple building. Without an administrative champion at the university, there was no incentive to forge such an innovative partnership, and in short, I am now (happily) at the University of California, Irvine, and not Wisconsin. Our lab was subsequently hired by UW–Madison and became GEAR Studios, where it exists as of this writing.

Reflections on Innovative Academic-Industry Partnerships

The most critical factor for academic-industry partnerships is for leaders with decision-making authority on all sides of a project to want it to happen. This statement may seem self-evidently true, but it is worth restating. If top leadership does not want innovative partnerships, they will not happen. All incentives align for middle managers to avoid new, risky, or innovative partnerships, and they routinely terminate projects by slow walking or stalling them out.

Leadership support is essential for executing contracts. University-private partnerships usually involve firsts for both organizations. New legal situations are confronted and new policies adopted. Observing contract negotiations (inherently an adversarial process), one might reasonably conclude that legal teams would prefer that deals not be completed. Those with decision-making authority must interpret feedback from legal counsel, push to solve roadblocks, and nudge parties to flexible terms. Academic administrators, who are not known to be risk-takers, must be comfortable with risk exposure.

Academic-industry partnerships may be more difficult in older, more established universities, with guarded business processes, than in younger organizations that create new programs, procedures, and policies routinely. In many universities, administrative teams are autonomous and operate as their own fiefdoms; they make their own HR decisions and, critically, select, train, and enculturate the next generation of administrative leaders. Business decisions, such as the appropriateness of contract terms, are made within business offices (the deep state of universities), whose role it is to protect the university from faculty who will break the university or fly-by-night leadership that doesn’t understand the university. They aren’t entirely wrong, but they also prevent innovation.

Lifelong innovators report that the two biggest hurdles to innovating in big organizations are (1) the “culture of no” and (2) the related response “But we’ve always done it that way.” Most readers know what a culture of no is; it is the cultural expectation of arriving at an answer of no, to not try something new, to not take risks. In the words of Clay Johnson (2015), “not how to make something work but to ensure that it does not fail.” As Johnson observed, a culture of no is an authentic desire to protect the institution. Administrative staff (particularly those second and third in command) are what Johnson calls the “Be” team, meaning that they were there before administrative leadership arrived and they will be there after they leave. In universities, these are the program managers and business office directors who feed information to directors and deans and thus shape institutional perceptions. Pursuing projects that challenge existing practices means that you are now in the business of organizational change.

In fact, most games for impact leaders find themselves managing organizational change. Organizational change requires (1) explicit administrative support and (2) a mindful approach to the act of change itself. In Rebels at Work, Kelly and Medina (2014) distinguish between “bad rebels,” who are characterized as pessimistic rule breakers who alienate others, and “good rebels,” optimistic creators who build better systems through forming coalitions. Building on surveys of change managers across government and industry, Kelly and Medina articulated a tactical approach to organizational innovation. Rebels must, for example, be more trustworthy than their peers, must learn to pitch and sell ideas to leaders, must never openly criticize the character of leaders (particularly superiors), and must inspire and mobilize colleagues.

Universities are averse to failure. Within technology, the mantra “Fail early and fail often” is used to the point of parody. In this age of innovation and the “faculty entrepreneur,” universities may claim to embrace failure but are more often uncomfortable with initiatives closing. Even those institutions (such as WID) that are held as sites of innovation are expected to do amazing and interesting things with 100 percent success rates. In fact, most university innovations are just achieving existing goals more efficiently, such as improving student retention rates. Innovations that suggest new goals (increasing private partnerships, fostering student start-ups) or call into question existing practices are less likely to be valued. Tellingly, five years after the experiment of WID began, none of the leaders involved in its creation were on campus. To the extent there is any story told about GLS within the university, it is told by Clay Johnson’s “Be” team. I’d bet the story is framed as a cautionary tale.

Notes

  1. 1. Without getting into the details, Gamergate was a reactionary movement against progressives who questioned sexism in the games industry, masquerading as concerns for ethics in games journalism.

  2. 2. There are one-person design “teams,” particularly in the indie genre, who do all of the art, programming, level design, sound, and back-end data infrastructure themselves. But these rare projects are the exception that proves the rule. They take years to complete, target niche audiences, and rarely are 3D or include industry-standard artwork. They would not, for example, meet the demands of Fair Play.

  3. 3. This arrangement was made possible by the support of UW–Wisconsin’s dean of the School of Education, Julie Underwood; WID director, David Krakauer; and senior university legal counsel, Ben Griffiths, to whom I am indebted.