GDC 2014: Day 1 - Art Directors

As the first day of the roundtable was dedicated to art directors, I opted to start with an important, but otherwise easy topic:

How do you communicate vision to a team?

• The first attendee to respond spoke of "town hall" meetings.  The purpose of these was to gather the team to ensure alignment.  This type of meeting helped to maintain momentum and provide context behind the decisions which were being made.  Later on, this idea was further reinforced as "weekly share sessions" where various workgroups regularly demo their own progress (successes and failures alike).
• Another participant advocated the establishment of "core pillars."  These would become the central tenets for decisions.  Core pillars also served to establish "confidence" within the team insofar as granting artists the ability to make their own decisions.  When pressed for how the core pillars were transmitted, the participant offered the ideas of posters (visual reinforcement) as well as frequent dialog (both individual and team)
• It was also suggested that monthly milestone reviews were part of communicating vision.  While I agree, I think it is worth noting that milestone reviews are reactive and the most successful vision setting is actually proactive and comes at the front of a milestone rather than the end.
• Communicating vision was also accomplished through paintovers of prototypes.  Again, I think you risk becoming entrenched in reactive direction with this example.  Paintovers can be incredibly helpful in "fine tuning" the direction, but run the risk of wasted effort if they rely too heavily on throwaway prototypes.  Paintovers also need context, so regular interactions and context is equally important in this example.  However, developing concept paintovers from real gameplay is sometimes more effective than just a "beauty shot" which may not relate to the game itself.
• Another suggestion was to storyboard the game.  I've certainly heard of similar cinematic approaches to art direction for games.  Personally, I feel that this becomes more problematic relative to the scale and scope of the game.  Storyboards can infer a certain amount of rigidity and therefore run cross-purpose to the iterative nature of interactive development.
• Yet another example was the creation of key references or touchstones for direction.  In these cases, being able to refer to other media (film, artwork, other games, etc.) helps people to understand the vision more quickly than long design documents.  At the same time, I've also seen the collection of references cause inconsistent vision.  Thorough context must be given for why the reference is appropriate.  In the example of using a character from a film, should the reference be looked at for overall mood?  For materials?  For color?  Reference can be interpreted in lots of ways and most are incorrectly.  The AD must ensure that the reference is interpreted properly.
• Wisely, another participant commented that the technique used to communicate the vision is not the most important.  Rather, the director must ensure that he or she is "passing the bus test."  Vision must be appropriately delegated out to the team, whether that be through a formal leadership structure or informal ownership.  This goes hand-in-hand with the comment above about confidence building.  The director cannot (and should not) be on hand to answer every possible question.
• One new option I heard this year was the creation of screen recordings.  Using Camtasia, the director/lead would record a voice-over as they reviewed some aspect of the game.  They could then show specific examples of where direction or content was most successful.  These were then uploaded to a communal dropbox where all team members had access.  Moreover, the videos could then be watched by new hires as part of their onboarding process.  My only concern for this technique is the amount of up front time this may require from the leadership group.  However, if they are consistent in execution, then I would expect the cost to drop off significantly over time as the value increased as the team grew.
• Another attendee mentioned the sizzle reel.  This one comes up frequently.  For those who may be unaware, this is a video version of the reference/keystone option listed above.  The biggest upside to this option is that it can potentially generate much more team excitement; however, it is also subject to the same drawbacks if viewed without context and detail.
• Rather than spending so much time on collecting reference or concept art, another person suggested the focus be spent on developing real-time prototypes.  The downside to a lot of sizzle reels, paintovers and vertical slice demos is that they are often tightly scripted, hacked together and/or don't represent an accurate picture of what is reasonable.  The suggestion here was to, instead, focus on developing a real prototype and then find the style within the emerging technical limits.
• Another example was Workshops or Live Brainstorms.  Ultimately, this sounds problematic to me.  Direction by committee is rarely effective except potentially within very small teams.  Otherwise, I still advise for a central vision holder.  That individual can still use these types of meetings to solicit input from other members of the team, but the final decision making authority should still rest with him or her.
• Lastly, and most importantly, it was suggested that PLAYING THE GAME is the most effective way to communicate vision.  Obviously, this is more practical in full production than pre-production (when you are still building the core game).  Regardless, this should still be reinforced.  I've seen teams late in the production cycle still rely too heavily on broad concept artwork rather than directing the game itself and allowing that to naturally influence the rest of the team.

For the second half of the session, I wanted to try something a bit different this year.  It required a bit of openness as well as the desire to share hard-learned lessons.

Tell us about your failures.  Tell us about "unintended consequences."  Also, what did you learn as an art leader that you wish you had known before you started?

• One attendee spoke about creating detailed asset spec sheets when they were new to the lead position.  Unfortunately, as time and asset volume increased, these sheets grew into unmanageable excel lists.
•  Yet another person agreed to having a similar problem with a OneNote document growing to similar unmanageable extremes
• One AD admitted to the mistake of trying to solve too many things in concept art.  At some point, you have to abandon the concepts and start building to find out what works and what won't.
• Another interesting example came from someone who directed the creation of a large animated sequence.  Unfortunately, the sequence was designed and created without soliciting team input or buy-in.  The end result was a finished sequence, but one in which very few team members were invested or excited about continuing.
• One amusing failure involved the creation of a number of character concepts.  Unfortunately, these were all generated from the wrong point of view and didn't communicate the right information.  In this situation, you can imagine that either the content creators had to infer a great deal or the concepts had to be re-created with the gameplay POV in mind.
• One common thread to almost all AD discussion was the focus on content over the game.  This year was no different, and at least one AD admitted to becoming too obsessed with details and thus missing the big picture.
• At the other end of the spectrum, a director pushed the concept group to really explore exaggerated character silhouettes.  Unfortunately, the game had a more realistic (rather than stylistic) setting, so the character ended up being messy, blobby and lacked readability relative to other elements
• Similarly, another AD lamented not finding the "signature elements" until well into production.  However, I don't have the sense that this was by intent.  Sometimes a style takes time to develop.  Ideally, this is something that needs to happen sooner -- and it shouldn't be delayed by the failure to just make a decision.  Regardless, a delay in defining the signature style will result in a delay caused by iterating on all of those assets that had been developed prior to that point.
• One participant admitted to spending too long trying to fix the broken parts of the game.  This is not a failure in defining a style, but sometimes the style may not work in certain situations.  Rather than admitting this limitation and moving on, an AD may feel motivated to force the solution.  Sometimes it's easier to just kill these things rather than let them consume your time.
• By comparison, sometimes you also need to know when to do a "deep dive" on a good idea.  It's a balancing act, either way.  This is why I advise time boxing your prototypes -- there has to be some point at the future where you walk away if it still isn't working.
• One AD admitting to the frustration of working with artists who just want to do art rather than be game developers.  I'm sure many of us have worked with this stereotype.  I know that different ADs have different opinions on the pure artist vs. developer.  I know I have mine.  Rather than advocate one position over another, I suggest instead that you define which is important for your studio/project/team and allow that to guide your actions.
• As much as I said that artists are this way, sometimes the AD falls into a similar conflict.  One participant admitted to making the mistake of expecting every asset to be great.  The byproduct of this expectation is that it slows down the development process and frequently requires a large external team to support the internal team.
• Unsurprisingly, one leader admitted to micromanaging creativity.  That's a common topic and a mistake that new leaders are prone to making.
• Another participant pointed out the lesson that the art leader's responsibility is to make sure their team's work looks good.  However, more importantly, they must ensure that the visuals properly communicate.  In order to do that, the lead must ensure that clarity of vision is found and maintained.
• Another attendee admitted to sometimes being too accommodating.  The intent here being to isolate the team from distractions or keep the conflict out of the bullpen.  I understand the desire to minimize distractions, but the risk to the team in this instance is the appearance of changes being erratic and lacking context.  Accommodation is certainly one path to resolving conflict, but as another participant pointed out, you also have to learn when to say no.
• Lastly, there were a couple of over-arching key notes on how to effectively manage a team.  Ideas which most of the attendees seemed to support and espouse
• Fail quickly.  Try lots of things, but don't be afraid to walk away from the bad ideas.
• Help artists to find the answer instead of outright telling them the answer.
• Make your decisions about the game.  Let the art reinforce the intent of the game.
• Share the vision.  Allow others to mentor and own elements of the game.

Speaker Evaluation

Lastly, I wanted to do something new with the write-up this year.  Here is the raw, unedited report from the day's roundtable.  If I felt a comment warranted a response, I added it below in orange text.

Art Leadership Roundtable: Art Directors

Wednesday, March 19 from 3:30 PM to 4:30 PM

Room 112, North Hall

Total Headcount: 87

Percentage of Evaluations Returned: 34.48%

Percentile in Visual Arts Track: 85th

Percentile in Overall GDC Main Sessions: 77th

Session Totals (This Session)			All Main Sessions	Visual Arts Track
Response	Count	Percentage of Responses	Average Percentage of Responses	Average Percentage of Responses
Excellent	23	76.67%	62.42%	57.67%
Good	7	23.33%	31.02%	34.84%
Poor	1	3.33%	5.61%	6.59%
Terrible	0	0.00%	0.95%	0.91%

Comments

Not enough time. great insights and enthusiasm about helping each other and up and coming leaders of the art world.

 Got quicker into discussion than previous round table years so a positive improvement. Still too big to be a real round table of huge value.

 More time for discussion would be better, with some more explicit control and guidance about the topics covered so it doesn't bounce inbetween just the usual complaints or concerns. 

 Excellent discussion, ambient noise was a bit distracting.

 Round tables are GDC's best invention. These also happen to be very well moderated!

Overly gabby participants stole the conversation and steered it into irrelevant territory. Questions were great. It would be nice if there was a little focus on soft skills.

 By far some of the best and most informative time spent at GDC. Keith was a great mediator and always kept the flow of the discussion moving.

Keith did a great job with this roundtable. I think the discussions went really well and moved at a great pace.

These round tables were the best thing about my trip to GDC! Please keep doing them.

These roundtables need better hosting. They quickly devolve into a single topic discussion. It might be better to pick three topics chosen by the group and spend 20 minutes on each.

Unfortunately, I tried that one year and it went VERY POORLY.  I'm not sure how to get a group that large to reach anything close to consensus on only 3 topics.  Ultimately, I've decided to control the topics myself and shift as I perceive the audience shifting.

Would be a good thing to make these larger rooms as I reslly wanted to attend this session but cciuldn't

I'm sorry you couldn't attend.  The rooms have been filling up at an astonishing speed in recent years.  However, I'm concerned about the rooms becoming so big as to discourage interaction (which is at the core of roundtable design)

Invaluable for people involved with Art Leadership