MechDog is an endless runner created for mobile devices. In the game you control a dog in a mechanical suit who is racing across the desert. I collaborated on artwork with @stupaah, but everything design and scripting related was done by me. MechDog is also my first mobile release (I've only prototyped things in mobile before) so there were a lot of lessons and pitfalls during development.

3 Things that Went Right

Keeping an Updated Dev Blog
The dev blog for MechDog was first started to share development progress and demonstrate design knowledge on my portfolio. An unexpected benefit from running the dev blog was that it provided a chance to reflect on design decisions. I was able to take a perspective that wasn't a developer's perspective by trying to write for a non-developer audience. Many game creators warn of the fallacies in trying to provided useful feedback when you become so entrenched in development; the dev blog provided a good way to take a unattached look at my own game.
Going for 3D Lighting
The 3D lighting definitely adds something unique to the visual style of MechDog. I used the small scope of the game to my advantage and tried out this new visual looks that combines 2D assets with 3D tech. Although it took a lot more time and effort than I'd like to admit, the 3D lighting adds depth to the look of the game that I can't now imagine the game without.
Middleware service Integration
There was a bit of hesitation on implementing middleware. The hassle of having to work with live 3rd party services was new and intimidating. Looking back on it now, it was a silly trepidation to have. Using Google Play Services was an easy way to have multiplayer features and including Unity Ads enabled me to create an interesting feature for the in-game shop. Adding new middleware features were low-risk and high-reward.

3 Things that Went Wrong

Skipping Visual Concepting
I foolishly made the choice to forego any visual concepting during pre-production. Between designing and developing, I left the art style of the game to come together during development. Unfortunately, the cobbled together art style is present in the game.There are many mismatching visual elements that hurt the game's presentation. Creating a visual concept for how I wanted the game to look at the start would have gone a long way to make the game's visuals feel cohesive.
Ignoring Technical Limitations
MechDog was a project that definitely went past its set deadline. A large chunk of time was lost to reverting visuals effects to try and save performance. It was in the middle of development when I learned that perf tests on mobile are different than on PC. Feature additions/changes that are innocuous on PC performance can cause unexpected hits on mobile. Adding frequent perf tests to my schedule from the start of development would have probably saved me dev time in the long run.
Underestimating Content Costs
MechDog provided me the opportunity to really get into systems design (most notably the procedural level generator).Unfortunately, the time and effort I spent on systems design didn't factor in the associated content design. Game content is essentially fuel for a game system, and I made the mistake of only focusing on creating a good system. Procedural systems like the one in MechDog are great for stretching less content, but the game would have probably benefited from more enemy types and more upgrade options.

Outcome

Working on MechDog has taught me a few things.

Design new content at the same time as designing a new system
Keep an updated blog, it helps reflect on design decisions especially when working alone
Perform platform stress tests even if you are familiar with the engine
When working on mobile, plan for frequent performance tests rather than tabling them for the end
Don't be afraid to give middleware features like Google Play Leaderboards a shot