Algorithm starting to recommend us to the wrong players?

So our game previously had between 1-2k ccs for 1-2 months straight. We suddenly had a spike that got us to 3.5k CCU out of nowhere, after which our engagement stats instantaneously dropped, for example play time dropped from 15.5+ to 13+ in a day. Following this we subsequently dropped to about 300-500 ccu, which was understandably a huge loss. We’ve since tried implementing tutorials, better onboarding, however these don’t have much effect. After the game died down to 300-500 ccs, we started hitting our engagement metrics, which then spiked us back to 1k ccu following which our engagement stats dropped again, which is leading me to believe that as of recently, the algorithm is somehow less efficient at suggesting the game to players who would enjoy the game.

3 Likes

yeah. I’ve noticed this too. The new algo sucks and is really volatile. Especially for new games that were released within the least 2-4 months, its really hard to maintain stable stats and ccu

2 Likes

Yeah, I hope someone from the support team can advise, it seems like every boost in CCU results in poor engagement, because the algorithm keeps pushing us to the wrong audience when our signals are good. We’ve had extremely unstable CCU and engagement metrics in the last month.

1 Like