The fine chaps of Action.Reaction Games in Finland recently were invited to an interview with GameMakers over at YouTube, and over the course of them talking, it dawned on me that while they are creating a really decent game with a very lean team, sufficient investment and all the right intentions, they are running into some usual pitfalls I had to learn while producing for and operating Planet of Heroes - an Apple-acknowledged Trend Game of 2017.
As I commented on this already in YouTube, I want to share the same thoughts on avoiding mistakes for all of you:
1. Monetization
This is the biggest. Vanity items (skins, visual perks...) will give you a global payrate of somewhere between 0.3 to 0.5%, but require Art assets clog down your pipeline a lot - concept, modelling, rigging, textures, animations, store material - , and you can only sell each item exactly once which will make you guys the prey part of the hunt.
Design for paid content that you can sell over and over. We have tried the fair approach without any Pay-2-Win mechanics, we have tried tried to use the Clash Royale system with gradual increase of power dependent on spending and/or winning, we have tried a slight to not-so-slight power increase at release of new characters. Frankly speaking, the last worked best on the balance sheet (by some x00% revenue) but it will burn through player base.
Note: The longer people spend in your core game - the splattering fun frag fest - the less time they actually have the opportunity to actually spend money on the game. Even more so: if the player is on the winning team, he or she is motivated to repeat that Dopamine release, and if they are losing, chances are that they close the game immediately - up to a point of churning. Longer core game time results in higher engagement, and thus a more drastic reaction to success or failure.
2. Concurrent Users
In a PVP game, you'd want to have matchmaking which cuts down your eligible player pool for creating a match.
You have essentially 3 ways to approach this; you have a steady flow of users coming in (unlikely for a limited budget company, but the reason why the well-known companies get away with it for a time), you have a very loose or flexible matchmaking bracket system (can lead to frustration and thus churn), or you invest time to design good bots systems that are very hard to distinguish from real players in their own skill level bracket. The last option is probably the way to go forward and most sustainable.
Also, you don't want to design new game modes; focus to nail down exactly one game mode to prevent splitting up your valuable CCU. Planet of Heroes had a story mode from the very beginning and split its user base right in half, with a big amount of players never touching the PVP game mode. You can even verify this issue being prevalent for the biggest boy on the block, League of Legends, with comparing queue times of Coop, ARAM and Ranked PVP games.
3. Networking
Multiplayer real-time is a fickle thing, and this time it's the techies to get the headache. While in soft launch in single countries or regions, you probably won't notice yet but at some point you will come to the situation that a South East Asian player - and they really love those PVP games! - would face a European player, leading to latency of 300-500ms.
The only viable strategy apart from user base segregation (see above for CCU issues) is a robust server-side prediction model that not only prevents rubberbanding (the observable warping of players) but also makes the game experience equally enjoyable for each latency-affected user.
That being said, watch out for scaling issues if using a server-side hosting model, especially at launch where you will naturally see a massive spike from launch featuring: plan early for a flexible system like AWS and a matchmaking that can dynamically scale regional segregation by linking neighbouring regions together (SEA+China, Europe+Russia+MENA, East Cost+West Coast, LATAM North+South etc).
Lastly, in latency analytics, don't trust client data. The worst offenders in packet drops will never even reach your analytics, skewing your analytics to the better side. In addition, think of overcoming temporary packet or even connection losses from cell-tower or network bridging in mobile games, additional load on the network from other network actors or background services and many other factors that affect not only the user that is subject to these issues, but also the users connected to this team mate - the skillshot that will miss, the take down that does not happen, the frustration about an AFK player sitting at base and dooming your success chances.
Lots of things to chew through but I really want to help not making the same experiences we had to 2 years ago - ping me for any questions.