Friday, 10 April 2015

On Wargaming behavior analyst team


Hello warriors,

you probably didn't know it - I did not - but Wargaming is not just about games, it's also about psychology. They have their own player behavior analyzing team working on the data obtained from live servers in order to make players as happy as possible - sorry, to make the most money of course :) Andrey Yarantsev (WG vice-president) and Sergei Kadomsky (player behavior data analyst) reveal how it works. Here is a summary of what is written in the Russian article above.

- analyst work consists of routine (reading reports, discovering issues and solving them, discovering negative trends) and special commission when a client comes and asks "it was like this and changed to this, why?"
- in game industry, analytics can be split into marketing (to achieve high return of investment) and development (internal)
- in WG case, the analysts segment the player-base into blocs (payers/non-payers, newbies/hardcore etc.) and treat them accordingly
- one of the latest reports was the report was an analysis of why players are leaving WoT and what players groups are the most risky
- on average, WoT player plays 20-25 battles in one session
- the amount of traffic on RU server is hundreds of terabytes of data per day
- analysts are collecting analyzing all kinds of data apart from the obvious (XP, grind speed, damage done etc.) , for example socialization (when the player plays in platoon, in company, when does he join a clan) or behavior in chat, purchases (consumer behavior of the player)
- marketing is also collecting and  connecting practically all the data available - for example how long does it take players from clicking on the ad banner they saw somewhere to buy their first premium tanks: even the smallest issues in this chain have to be remedied
- the most important data that are looked at first are the online numbers, new registrations, payments and expenses
- the most typical reason for various non-standard behavior in this sense (payment drops etc.) are technical issues with the game (server downtime), even smaller fluctuations however trigger an investigation
- WG analysts are also taking regions into account and studying each region separately as well
- the most important values for the analysts are CCU (concurrent users), new registrations, payments, CPI (costs per install) and LTV (life time value, eg. what the player spends during his gaming career in total)
- when WoT started, players were split into very basic groups (newbies/potential core/core + payers/non-payers). Later on, the analysis became more complicated (for example "how is the non paying core part of players behaving")
- another phenomenon that was investigated was the one of "paying newbies"
- regarding various groups of players, CW players for example are motivated "by the need to win and lead an army", e-sports players are motivated by the option to earn money and to accomplish something. The motive of random players is mostly "fun"
- developers are trying to stimulate team-play "from the bottom" - first incentivize players to play platoons, then companies, then clans, because socialization is what keeps the players in the game
- random players are incentivized by minigames, discounts, events and competitions
- one of the most important elements that keeps players in the game is user-generated content (social networks, videos, streams)
- stimulating players to buy stuff is actually pretty difficult, they have to start thinking that they actually need something, like premium account. One way to motivate them are various discounts - but there's a huge problem: if they come too often, people get used to them. That's why WG is trying to attach discounts to various very important events and tries not to make them like a regular thing.
- instructional videos (for example "How to play XX") also stimulate players to buy stuff because players that do not know the vehicle are afraid to spend money on it
- one of the cases where analysis failed were historical battles that expected much higher interest from players, one of the main reasons the format failed was "lack of balance and the absence of garage battles"
- another huge failure was the 9.0 improved graphics where the developers overestimated the desire of players to upgrade their PC, expected 30-50 percent of players to switch to improved graphics immediately and in result only 5 percent did
- one of the things that surprised the analysts is how long some memes can survive, some RU memes survive for years ever since the betatest
- 70 percent of RU players claim that WoT changed their attitude towards history (reading more books about it and stuff)