Context
During our weekly meeting with our publisher, we got the suggestion of working in shooter games, as at that time, cover shooter games were returning great LTV. I initiated an internal discussion with the team to assess the feasibility of developing such a game within a defined timeline. I put on my programmer hat, and after talking to my fellow programmer, we were confident in creating a structure for this concept. Generally, it takes us 4 weeks to produce a testable product, so we respected that timeline as our initial agreement
Hypothesis
I came up with the idea of Auto Cop, taking massive inspiration from Robocop. The shooter that was on the market was Agent Hunt. The idea for Auto Cop is to reuse what’s working for Agent Hunt, change the theme, and tweak the core to come up with something that feels different. Since our main character was a powerful being, I wanted the shooting to be very powerful, it means we sacrifice some precision and stealth.
Ideally, we wanted a CPI of less than $0.4 in the Meta platform on Android. If we can achieve that, we can aim to achieve an LTV of $0.8 at D7, taking us to 200% ROAS. It was the target of our publisher. They believed it was attainable. As they already had a title in the cover shooter category that was achieving it
Process / Experiments
We started the development. It was complex because the timeline was too short, and there were a lot of modules involved to get everything working. Our target audience was 90% male, 18-35 year olds, playing mobile games. Our design pillars were based on certain limitations.
- Portrait orientation
- Auto movement
This will keep us separated from the other well-established shooters. We don’t want to compete with those games as they are well saturated, and we don’t have the resources
We started prototyping. Our programmer used a spline for the path-based point to movement. We coordinated as a team. I prepared the GDD and started taking references from other games to get learnings. While my fellow programmer focused on building the core, I shifted to another project that I was also working on. After 2 weeks, we had a playable version. I hosted internal play sessions where we discussed and noted down things we wanted to try. We decided to keep enemy behavior simple. They patrol or stay idle -> after the first shot, they move to a random cover
After the core was ready, After 3 weeks, I joined the programmer to program. I added popups, camera shakes, smoothing transitions between cover and shoot mode, enemy force and fall, etc. These made sure the shooting felt on point. I also programmed the shop and added weapons that can be bought with earned coins.
I experimented with the aim button by trying an input where touching anywhere will start shooting, and the player will need to aim while shooting is ongoing. But it felt too harsh, although it gave a powerful feeling after playing for some time. To improve the powerful feeling, I added props that are destroyed when shot. I also made ammo infinite, and made it so that when shooting too much, the gun becomes hot and stuck. So users would need to stop and then shoot again when it gets close to maximum temperature. This improved the ‘power’ feeling
Observations & Outcomes
After testing in the US, it had a nice CPI of $0.37, the D3 cumulative playtime was 912s, and D1 Retention was 23%
It showed good marketability. But the gameplay numbers weren’t there. We were looking for a D1 Retention of 35%.
To summarize the v1
- CPI of $0.37 shows that people were loving it in the first impression, and installing the game
- The gameplay numbers show that it had some stickiness, but it was far from perfect
Because of the good marketability and with the hope of improving in-game numbers, we decided to iterate on it.
TAKEAWAY & NEXT STEPS
We had the CPI, which is generally hard to get. I did not estimate the timeline correctly going into the game, which wasn’t ideal. We put a total of 10 levels, which also could have contributed to the low gameplay numbers.
The game design focused on the core and fun. I did not put any effort into the progression design, weapon upgrades, etc, because of the time constraints.
After discussion with our publisher, we decided to iterate on the game. Now we aimed to ideally double the retention D1, and cumulative playtime D3 to increase the chance of the 200% ROAS in D7.
Here’s the link to the big design mistake I made on version 2
