Production Blog Postmortem - Shapeshift Dungeon

Hello and welcome to our final production blog post for Shapeshift Dungeon! This one will be a little different from all of the others; it is a postmortem blog focusing on the semester as a whole rather than just what we accomplished in a single Sprint. With that said, we had an absolutely incredible Sprint 7 even though it was cut short by a week compared to our previous Sprints. For us, this meant having our level designers working on new particle effects and a final detail pass on the population of environment assets in each of our levels. It meant having our programmers get one last layer of polish on our items system, our AI, and our audio integration. It meant having our 3D artists work on the most unique and creative item pieces that make the dungeon rooms just that much more special. It meant having our 2D artists create key art for our game and for all of our team to come together to get every last enemy animated for the final release. While this postmortem post might not be the right place to focus on individual Sprint 7 work, I want to highlight just how amazing our team is and how they never stopped pushing their best work out for our game, even when faced with other finals.

In total, we had completed a whopping 200 points in Sprint 7 which is our most ever. This leaves us with an average of about 127 points completed per Sprint and a grand total of 893 points completed throughout the semester. While we had planned for 1208 points, leaving 315 points cut, I am incredibly proud of all of the work we had accomplished this semester. Those 315 points were primarily composed of additional enemy types and stretch goal features. We reached our minimum viable product by the end of Sprint 3 and had a build we were happy with by the end of Sprint 5. Every point of work since then has been polish, making adjustments, and adding content. I think it's really important that we were able to make time in Sprints 6 and 7 for additional content because we were able focus on the fun and the creative tasks more than just the necessary ones. This allowed our team to really express their creativity and it really paid off.

What went Wrong?

No art background in leadership

Neither myself nor my designer, Sky, have an artistic background so when our artists asked us what we wanted for some of the assets, we did not have an answer ready for them right away. Nor did we always know what to suggest when giving feedback. We did not know what we wanted so when they gave us concepts, our response was all too often "oh wow, these are all great." Although this is something we worked to improve on over the semester, it led to a few circumstances where our artists, 2D and 3D, were going in circles for a few days without direction from either of us. Ultimately it came down to a lack of confidence from either of us and that slowed down production. I want to thank all of our artists for putting up with it and still producing such fantastic work.

Scope - Too many enemies and not enough time to test them all

While I'm sure every group will have some sort of scope problems, ours manifested in an interesting way. We had originally planned to work on a monster-of-the-week paradigm which would allow us to create, test, and iterate on one enemy each Sprint leading up to a total of seven enemies for seven Sprints. Ultimately, we made it to seven enemies but the sheer complexity of our AI system as well as the innate slowdown from sharing a single rigger with the other three teams delayed our production. This meant that we had our first two enemies implemented quite early and they got a lot of playtesting but the remaining five had to wait until the end when our systems and pipeline were scalable. We are ultimately very happy with having created seven unique enemies that all act very differently from one another, but we had to rely mostly on internal playtesting for them at the end. We even created a boss fight we were very happy with, but it went through several internal iterations and it felt irresponsible to release it to the final build without external playtest data. Our lesson here is that scope isn't defined just to what can be accomplished, but also what should be accomplished. There's a lot more to having a feature for a game than just creating it. A well designed game is one that has had a chance to respond and iterate after feedback.

Producer: "No Crunch"

One of the things that I'm most proud of from last semester in Fall was creating a game entirely with no crunch from any team member. That worked for us. So when I was selected as producer for this game, one of my goals was to continue that legacy. However, that is, quite frankly, not how our team likes to operate. Our programmers, lead by our designer, met online every weekend and jammed throughout the night to get as much done as possible. We came to calling them programming parties and they were a lot of fun and were very very productive. A big lesson for me as producer here was that I should not just implement my way of doing things onto the team. Instead, I have to recognize that my role is to facilitate so that each team member can perform to their absolute best. I had to set my own goals and my own ego aside and realize that the best thing I can do for the team is to enable them to succeed. Sometimes that meant being a parent and making sure that they take stretch breaks and are properly fed or hydrated, but other times its me being there with them and helping to solve issues with the code. This will change from project to project, but I think the real lesson here for producers is to be adaptable and to remember that the prime directive to is to enable. Everything you do should make something else easier for someone else.

What went Right?

No mandatory crunch

As for what went right, a big part of it was no mandatory crunch. We never explicitly told anyone that they would have to stay up to get work done for the build. Fortunately, our numbers were high enough that if someone didn't finish their task one Sprint, it's no big dal because they can just finish it next Sprint. I relied heavily on being honest with my team about my goals and expectations for them. Likewise, I also relied heavily on them to be honest with me about what they can and cannot accomplish each week. As you can see from our burndown chart, our team accelerated greatly towards the end. My belief is that this came from the team seeing our game come together and they became more motivated to keep making our game awesome with each Sprint. As a producer, I found that fostering an environment where I set clear expectations, allowed room for growth, and enabled them to be honest with me about what they could not do was a very positive thing for our game. Last semester, I was told that being a producer isn't about getting people to do more work. It's about making sure they are able to do the work they can do. In a student project, if they have an upcoming exam, a busy work schedule, or a family to take care of, these things have to come first and they need to know they can be honest with me about their priorities. Fostering a healthy environment is the first step to allowing the team to form a passion for the project they're working on.

Comprehensive Pre-production & Documentation

This was another big win for our team. Our game design document was started before the semester and was quite extensive already. In addition to that, Sky did a great job of updating it all semester long along with coordinating with me to update the asset lists, feedback docs, and bug logs. We streamlined so much work throughout the semester thanks to this documentation and every member on our team would refer back to it constantly. It should come to the surprise of nobody that having good documentation makes making games easier, but I would like to stress the importance of it. Every minute spent working on documentation paid back hours of time saved during production.

The Team

Something special about 495 that isn't quite true about all of the other classes in the major is that everyone is here proving that they're ready to do this specific job in the industry and it shows. I often forgot this semester that I was working with other students and just came to rely on them as experts in their own fields. We will all be entry level when we enter the industry, and for good reason - experience is important. However, our team in particular is incredibly talented, eager to learn new things, and every one of them just has this immense passion for making games that makes them a delight to work with. It has been absolutely incredible and a privilege to have been able to work with each of them this semester. I have nothing but respect and thanks for these future titans of the game's industry.

Justin Donato
Alexander Lopez
Joseph Warren
3D Modelers:
Devin Spencer
Jose Vazquez
Chance Villavicencio
Level Designers:
Marc Billones
Branden Kelley
2D Art & Animation:
Tanya Medina
Alexis Gutierrez
Armando Legaspi
Eder Ayala

What would we do differently?

Set clear expectations and provide a more clear art direction

We have far too many instances of beautiful designs and concepts from our 2D and 3D artists going to waste because we couldn't settle on a single design. As leads, it is our responsibility to hold the vision and we did not always have answers for our art team. Nor did we always have references for them so often times they would have to go off a short description in our documentation and a user story that was written for them back when we were paper prototyping. This is something we both need to work on if we want to be leads in the future because being a creative lead on a game means being responsible for all departments that are making the game. All of it comes together to make one unified experience for the player and this weakness that we have is not one we can afford to keep.

What would we do again?

Player Feedback

This semester, we focused on four types of player feedback for each of our 6 builds. The first is our major standard, a google feedback form that asks qualified questions to the player about their experience and demographic. This was good and helped us understand our players and allowed them room to give more detailed feedback of their choice for what they felt they needed. The second form of feedback we implemented was a level feedback UI form built into the game after completing each level. We set it up in Unity so that the  response, along with a small amount of data about the player's HP, potions, items, and progress, would be automatically emailed to a private email address with a 1-5 rating of our levels and an open text field. Filling this form was optional, but it turned out to be extremely valuable because our players would complete between 5 and 20 levels in each playthrough. Not only did this tell us about their progress, but it also let us collect their opinions while it was fresh from each level. Our level designers were tasked with writing a report each Sprint from the feedback and they were able to pinpoint every problem  with astonishing precision, turning each of our levels into nothing but 4 and 5 star reviews from happy players who felt like each one was well crafted. We also collected some data using Unity analytics but found that the data was not as useful to track until we had the right questions to ask from it. Towards the second half of the semester, we were able to use this data to create balance changes that ultimately lead to our upgrade system having a near even distribution of choices from our players with regards to upgrades. Finally, we got some video feedback from people playing our game and we got to see them react in real time. This proved to be the best feedback because it was very organic and we got to see where and how the players moved in real time. If we could have collected more of this, we would have, but it was a difficult thing to get during the pandemic. Thank you so much to all of our playtesters. We made a great game thanks to all of their feedback.

What did we learn?

People make games

As a producer it can be easy to get lost in the points and the numbers, but it is so important to always remember the human aspect of your work. Your team members are not assets, they are not resources, they are not points machines, they are people. Real human beings for whom every point of work is a story. It is a challenge, an obstacle overcome. It is a lesson learned, a journey taken, a task completed that they chose to dedicate their time to instead of spending that time with their family or pursuing a different career. Enabling these people to succeed, whatever that means, is the role of the producer and that will always change depending on the individual people and the project. Meet to gather consensus, not to make decisions. Remember to always be investing in your team. Work to make them better. Every piece of training and new resource given will pay for itself tenfold. And most importantly - we can do this. Our team was fantastic this semester and we made an awesome game. We're ready to do it again. We got this.

Thank you so much everyone for joining us on this journey. Making Shapeshift Dungeon has been an absolute blast and the game is genuinely really fun. Thank you so much to our team and to our playtesters for helping us get this far and to our designer, Sky, for all of his excellence and dedication to this project. It came out great and we couldn't be happier with our team and our game.

To any future CAGD producers and designers who might be reading this, feel free to reach out if you need advice. You can do this.

Shapeshift Dungeon is now finally Shapeshift Done-geon.

Arjun Gambhir
Game Producer

Get Shapeshift Dungeon

Leave a comment

Log in with to leave a comment.