Data centric Design & Operation: A data-driven and scientific approach for game business

Data centric Design & OperationA data-driven and scientific approach for game business

Nguyễn Chí Hiếu - Japan Dept – VNG Corporation

Methodology of data-centric approachWhat is data ?DisclaimersUnit economy

Table of content

Methodology of data centric approach

Japanese Methodology & Principle buzzword:KaizenJust-in-time principle

Good at Math Bad at Math

Opposite of Good at Math is not good at Literature or Creativity

Data-Centric # Limitation of Creativity

Methodology of data centric approach

What is data ?

Is this the “data” we looking for ?

Option AOption B

We still need a good game to start with

Data-centric: Disclaimer

Profit = ( Revenue per User – Cost per User ) X Number of User

Number of User Acquisition

Revenue per User Retention Monetization Life Time Value

Unit Economy

Acquisition – User funnels Everyone Internet user Gamer Platform user base Target Segment Ads Awareness Interest Desire Action Registration Download client Chose character Tutorial Play Stay Regular player Payer Regular payer …….

K-factor measurement needs reliable viral mechanic. Viral is becoming less and less effective.

Acquisition – Viral K-factor

Profit = ( Revenue per User – Cost per User ) X Number of User

Number of User Acquisition

Revenue per User Retention Monetization Life Time Value

Unit Economy

Let’s have a look at how new users stay in our games.

Retention

Normalization chart for user retention over 1 month on daily basis.

0 2 4 6 8 10 12 14 16 18 20 22 24 26 280.00%

10.00%20.00%30.00%40.00%50.00%60.00%70.00%80.00%90.00%

100.00%

=

28-02-1227-02-1226-02-1225-02-1224-02-1223-02-12

Percentage of staying user/total user

Date

Retention

Normalization chart for user retention over 1 month. How do we keep user ? How did they leave ?

1 3 5 7 9 11 13 15 17 19 21 23 25 27 290

50001000015000200002500030000350004000045000

40945

14,318

6,883

Staying User

Staying user

Retention

Begging the players: “Don’t leave me, I can change for you” ?

Retention

Lock them up ? Any better idea ? Let’s stay by asking ourself: “Why do users stay ?”

Retention

User retention at a closer look. How users funnel into your game. How do you impress your player. “Don’t make me think” - KISS (Keep It Stupidly Simple). How can user understand “core design”. Do you have Retention Features in your game cycle. What is your Retention Feature KPIs. 1st Login to 2nd Login. Define your Hardcore/Reg user. ……

Retention

Profit = ( Revenue per User – Cost per User ) X Number of User

Number of User Acquisition

Revenue per User Retention Monetization Life Time Value

Unit Economy

How we frequently look at the most important part of our business: ARPU : Average Revenue Per User ARPPU : Average Revenue Per Paying User DARPU : Daily Average Revenue Per Paying User MARPU : Monthly Average Revenue Per Paying User Conversion Rate. Paying User Rate. Sale charts.Is that all ?Can we do better ?Why do user pay ?

Monetization

Profit = ( Revenue per User – Cost per User ) X Number of User

Number of User Acquisition

Revenue per User Retention Monetization Life Time Value

Unit Economy

Life Time Value = Total Revenue you get from 1 user until they cease to be your user.

Profit = (Cost per User – Life Time Value) X Number of User. Most reliable Life Time Value is historical data. Historical data = history, you need some way to predict, or project

your Life Time Value 2 most simple Life Time Value Models on Cohort basis:

LTV = ARPPU x Paying Rate x User Life Time = ARPU X User Life Time

LTV = ARPPU x Paying User x Paying User Life Time

Life Time Value

Normalization chart for user retention over 1 month.

1 3 5 7 9 11 13 15 17 19 21 23 25 27 290

50001000015000200002500030000350004000045000

40945

14,318

6,883

Staying User

Staying user

Life Time Value – User Life Time

Life Time Value – User Life Time

Projecting object lifetime is an old problem.

Life Time Value – Poisson Distribution

Segmentation criteria Daily cohort basis. Marketing Campaign basis. User source. User behavior. User Demographic.

Life Time Value - Segmentation

Banner A new users: 1,000 Banner A cost: 500$ Banner B new users: 5,000 Banner B cost: 1,000$ Banner C new users: 500 Banner C cost: 500$

Total new user: 6,500

Total banner cost: 2,000$

Banner A user LTV: 1$ Banner A LTV: 1,000$

Banner A Profit: 500$

Banner B user LTV: 0.1$ Banner B LTV: 500$

Banner B Profit: -500$

Banner C user LTV: 3$ Banner C LTV: 1,500$

Banner C Profit: 1,000$

Life Time Value - Segmentation

Profit = ( Revenue per User – Cost per User ) X Number of Users

Number of Users Acquisition

Revenue per User Retention Monetization Life Time Value

Unit Economy

Thank you !

My Contact: hieunc@vng.com.vn