Wednesday, May 2, 2012


GDC 2012:  Art Director / Lead Artist Roundtable

Day 2 - Thursday, March 8

3 Day Attendance Stats:
Wednesday -116, Thursday - 90, Friday – 57

First of all, let me state once more how much I enjoy attending and speaking at GDC.  This was my fifth opportunity to speak at GDC, but the first time heading up this specific roundtable.  I want to take a brief moment and thank everyone who attended the roundtable series, especially those who put forth the effort to attend each session.  I also want to extend additional thanks to the Conference Associates who helped run these sessions.

On with the notes…

Day 2:  “This is more like it…”

The second day of the roundtable was much more successful.  Armed with the knowledge that I wouldn’t be able to post notes to the walls, I came armed with tape and was able to post the notes to the table (facing inwards, towards the center of the circle).  The slight downside is that this meant that not everyone could see the notes, and no one could see all of them.  However, it provided enough topics within easy view that individuals seemed to feel comfortable broaching topics and discussion points.

Beyond introductions, I started the session by asking any art director present to voice his number one frustration over the past year for the group to drill down and discuss.  The first topic voiced was Scope Creep.  From there the group then began brainstorming and sharing experiences from a variety of studios as to how they addressed this issue.  What follows in a list of key strategies and contextual solutions.

Problem Topic:  Scope Creep
  • Mini-milestones.  One manner by which scope can be managed is by setting more finite delivery dates and sprints.  Lengthy milestones can be tempting opportunities for exploration, both good and bad.  A fast pace keeps the team focused on immediate goals.  The scope is constrained by the necessity of stringent deadlines
  • Weekly Reviews.  Regularly recurring reviews provides opportunities to re-establish goals.  Also, the team maintains traction on priorities.
  • Milestone Bucket.  Be clear with how much is planned for a given milestone.  Either it’s in, or it’s out.  The bucket should have enough “buffer space” to allow for the unknown, but not so much that focus is lost.  Any in-depth exploration is then reserved for future buckets, which have to sacrifice what may have already been planned.
  • Prioritized Backlog.  Goes hand-in-hand with the previous point.  Succinctly, new ideas or avenues of exploration are added to the backlog rather than the current milestone.  It should be noted that the person who voiced this solution also commented that this largely works in theory, but little in reality as the backlog can just become a catch-all for ideas and then the priorities can be manipulated outside of the scope of the project.
  • Sprints.  Clearly related to the mini-milestones and prioritized backlog.

The next question posed:  What is the greatest contributor on your projects to scope creep?
  • Increasing Complexity.  This can be natural due to emerging gameplay systems coming online and having to interact.  This can also be the unfortunate consequence of pursuing complexity rather than simplicity.
  • Shifting Priorities.  During the development cycle, priorities shift as expected.  The problem arises when the shift in priorities result in pursuing gameplay, engineering or visual development that wasn’t previously planned or established.
  • From running to flying.  This was a concrete example of a foundational gameplay component that would have significant impact on scope.  In this instance, the way in which the game is authored and played has to be re-evaluated at a basic level.
  • Story Development.  Changes to story will also impact scope.  Gameplay components and engineering needs may be directly impacted by story hooks, encounters and even by cinematics – especially given the popularity of QTE sequences.
  • Competing Art Direction.  In this case, competing art direction referred to having to match stylistic and franchise (or IP) direction across a variety of platforms (console, PC, web, mobile).  It’s possible that limitations of one platform result in art direction changes across all SKUs.
  • Creative Director.  Understandably, whoever holds the “project vision” will directly impact the scope of the project, especially if the “vision” is unclear or prone to change/redirection.
  • Buffers.  An interesting point raised was that “buffers” are necessary to insulate the team from certain scope changes.  The idea being that the department director (and Leads) can prevent the team from becoming “distracted” by scope changes.  The counterpoint voiced was that the “buffers” can also negatively impact scope by preventing information from reaching the team for too long.

The next question posed:  What strategies have your studios employed to address scope creep?
  • “Baby Kills.”  In short, the team strives to engage in regular sessions where they kill ideas before they become anyone’s “baby.” 
  • Broad vs. Specific.  The idea is that the project’s or team’s broad goals are established and confirmed first, before the pursuit of specific elements commences.  Chasing the details before the high-level vision is established is the quickest route to scope creep.
  • Striketeams/Scrum/P.O.  Using some form of agile development or feature-driven development teams may prevent scope creep from happening by providing a broader cross-section of the team to investigate the element at an early stage.
  • Find Allies.  At the risk of sounding political (which is NOT the goal), working with others to help them understand why an idea is or is not worth pursuing can bring the issue to light and gain traction on a solution quickly.
  • Pick Your Battles.  Priorities.  Because some scope creep is inevitable and some can be beneficial, it’s also important to remain open to new opportunities.  Plans can change, and probably should.
  • Change Date.  Not much detail needed here.  If release dates move in, the scope of the project needs to be changed.  The problems really become exacerbated in the rare circumstance where dates move out, and the team defaults to adding more content rather than refining/polishing what already exists.
  • Communicating Consistent Vision.  If you know the “razor’s edge” or “hook” which defines the vision of the project or the core gameplay, changes to scope can then be measured against the consistent vision.
  • Good Project Management Team.  A good management team will develop and safeguard a clear change management approach.
  • Overscope Early.  The idea here is that crafting a large scope early dissuades the team from continuing to overscope and instead breeds a culture where the team is focused on trying to reduce scope.  However, I would say that this only works if you have a leadership group who understands that that they’ve intentionally overscoped rather than just having unrealistic expectations.  The two scenarios look disturbingly alike.
  • Broadcast for Ideas & Actively Cull.  In short, create an open environment for idea generation and then openly cull the ideas, closing the communication loop.
  • Know Team Strengths.  The idea here is that identifying the skills and abilities of the team can prevent the project scope from drifting into areas that do not play to those strengths and will invariably take longer to explore, learn and develop.

The next question posed:  Is scope /feature creep always a bad thing?
  • Feature Creep Can Be Good.  One attendee rightly pointed out that sometimes the most innovative ideas come out of exploring ideas that were not part of the original plan.  However, this needs to be carefully balanced against reality – not all gambits pay off.
  •  Art Leads Manage Process for Change.  Although this is true for the leads from any department, scope creep has the highest probability for success if the department leads actively manage the change rather than stepping aside.  If the leads are invested in the change (and moreover, willing to sacrifice other elements), the new element has the highest probability of success.
  • Tools Impact Viability.  Understandably, the capability of the tools can limit scope creep from happening.  However, changes to scope can also motivate tool development and additional (or more efficient) feature integration.

Having wrapped up this topic roughly halfway through the session, it was a good time to jump into the next major topic.  The next topic up for discussion was how to create an environment for effective art critique?

Problem Topic:  Creating Effective Art Critique
  •  Meetings.  The point here is to fine the meeting structure that best suits your team or group.  Some people prefer a group meeting structure, some face-to-face interaction.  Meetings aren’t a surprise, the effort comes in finding the process that works for your team, and then refining that process as the team evolves.
  • Dailies.  One participant suggested that daily feedback is most effective.  What should be kept in mind is, as above, the frequency with which feedback is given should be adjusted to match the team, not the other way around.
  • Artist-driven vs. Director-driven.  One point of discussion was whether or not to build a process driven by the art lead’s or director’s needs, or rather the feedback should be initiated by the individual artists.  While there was no sure-fire solution (again, build the structure around the team), there seemed to be a general consensus that the early stages of development are defined more by lead-driven feedback and the latter stages by individual-driven feedback.
  • Peer critique first.  One director shared that his system was built around ensuring that the artist has shared the work with his peers first.  The idea being that the person can share the feedback they’ve received from others when the director reviews the work.  Understandably, there is a chance that the artist may respond to conflicting direction.  However, the upside is that hearing the feedback an individual recounts from their peers provides the art director insight into how well the art direction of the project is understood by the whole team.
  • Animation – block out, checkpoints, asset viewer, game approval.  This was the process advocated by an animation lead for staging out feedback on animation development.
  • Get it in Game.  Despite the ease of reviewing work in your primary DCC tool, most attendees agreed that it was best to view assets in game rather than outside.
  • Group Meetings > Emails.  The consensus here being that feedback via email is significantly less effective than group meetings.
  • Space Planning for Interaction.  This point was that the art team be arranged and organized in such a manner as to create effective critiques, especially in cases where you want similar artists to hear similar feedback or learn from one another.
  • Let it sit; Use time.  While the natural tendency for most art leads and directors is to provide feedback instantly  and in-the-moment, one attendee suggested that allowing time for an asset to sit can be beneficial.  Obviously, this wouldn’t be the case for assets that are way “off mark,” but could be useful for having an artist move onto something new while the lead reviews a larger group of assets in context of one-another.
  • Remote Development -> Skype, Vid Conf ; Face-to-face > email.  The question of how to art a direct external team was posed.  While the leading tool was understandably some form of video conferencing, more important was the need to have face-to-face or real-time interaction rather than just emails and/or forums.
  • No Meeting days; Preserve time.  As time for critique always seems to come at a premium, clearly setting aside the time for proper critiques is instrumental rather than simply relying on critique happening on the fly and irregularly.
  • Impacted by Team Seniority.  As a key factor, the seniority of the art team directly impacts the effectiveness of art critique.  A more senior team can potentially stand in for the art director or lead at times, moreover than can help guide junior artists through the early stages of critique.
  • Time for playing the game.  The point voiced here is that setting aside time for regular play ensures the team maintains context for both their work and the project’s vision/direction.

Peer Critique:  The next topic that was discussed was how peer critique was leveraged in different studios.
  • 360 Feedback.  One posed that within a smaller studio or team, if the art direction is documented and clear, that the feedback can come from anywhere on the team rather than solely following a hierarchical structure.
  • Adaptable / Flexible.  The counterpoint to the above was that that required a team that was adaptable and flexible to changing direction.  The point here was that 360 feedback may require review and revision from a director if he or she disagrees with the peer feedback.  In short, this approach has to reflect the abilities and culture of the team.
  •  Safe but hierarchy for decisions.  Another attendee countered that this practice could create a more safe environment for critique, but that ultimately the responsibility for the decisions will still have to rest with the director or lead.
  • Start with technical checklist and then teach aesthetics.  This was a very interesting suggestion.  In the early stages of production, focus peer critique on the technical checklist of assets.  Over time, this can naturally segue into more aesthetic critique.  The upside to this approach is that it builds confidence and certainty over time.
  • Break the production pace.  The intention behind this suggestion is that setting aside time for in-depth peer critique breaks up the production pace and gives everyone time to review and critique rather than just moving on to  the next major milestone.
  • Too rigid a schedule for reviews.  The unsurprising counterpoint was that one attendee pointed out that most studios have far too rigid a schedule to allow for those types of critiques.  Naturally, the only way that idea would solidify is if it were built into the team’s culture from the inception and that the project planning took into account.
  • Show WIP.  Although obvious, one attendee pointed out that it’s important to build a culture of presenting work-in-progress rather than just whenever the artist thinks it’s complete.  The benefits should be obvious, but the reality of rigorous production milestones can erode away at that original intent.
  • Allow for iteration time.  Most would probably agree that trying to finish the game by completing cross-sections at a time can often create more problems than solve.  Rather, it makes more sense to get as much “beta-level” art into the game as possible allows for the gameplay to iterate across the entire experience without the art team wasting time polishing components that may get cut during gameplay iteration.
  • The next point was to ask what structure different studios used to facilitate reviews.  While some favored a simple network folder, other studio have dedicated forums or web galleries for displaying WIP.  In general, the attendees didn’t favor either system over the other, as the most important component was to ensure that the face-to-face time and time spent reviewing-in-game were preserved.
  • Lastly, it was posed that in multi-project studios, there is real benefit in having different teams review and provide feedback on the art that is in progress.  While they may lack the complete context necessary to provide an accurate critique, the other team is also not burdened by assumptions.  The critique can hopefully be honest and pose meaningful questions.

The last component of the critique discussion focused on:  Taking Non-artist Feedback.  Most attendees seemed to agree that taking feedback from outside of art is inevitable, but also beneficial.  However, there are some caveats to this process.
  • Reframing / Understanding Comments.  It’s important for artists to make sure they understand the feedback, as artists may not speak the same “language” as other departments.  Learning to ask good questions and rephrasing feedback is a good habit for artists to develop.
  • “Lunch & Learn.”  One attendee indicated that their studio would regularly invite non-artists to an in-studio lunch where they showed the current stage of art development.  This was good for both presenting art goals, and soliciting feedback from outside of the department.
  • Encourage whole team to participate.  Rather than creating an environment where non-art feedback happens erratically and only from a select few sources, invite the entire team to participate on a regular basis.
  • However, it is not enough to just sit non-artists in front of the work and expect usable or productive feedback to just pour forth.  First, the lead or director is responsible for providing the context behind the work before critique is offered.  Second, the art team is responsible for closing the loop on that feedback and clearly indicating which feedback they intend to integrate and which they will not.
  • One of the most beneficial upsides to organize critique sessions outside of art is that, naturally over time, you teach the other departments your vision, your values and your goals.
  • Clarity of Design.  The last point voiced was that, while the idea seems simple and alluring, it is also vital that the art be supported by clear design.  If the design of the game is unclear, then accurate and useful feedback is hard to judge or align with product vision.

Big Takeaway / Lessons Learned:
  1. Strategies for addressing scope creep.
  2. Ideas for building an effective and productive atmosphere for critique.

No comments:

Post a Comment