Share
Embed
Olesya Kozlovskaya, Arseny Alikhanov, Google
2018 in Review: (a short) Google's Perspective on Mobile Gaming Marketing Trends
(White Nights Conference Moscow 2018)
The official conference website — http://wnconf.com
wnconf
November 06, 2018
Transcript
79 slides
-
Proprietary + Confidential Perspective on Mobile Gaming Marketing Olesya Kozlovskaya
| Analytical Consultant, Gaming Arseniy Alikhanov | Analytical Consultant, Gaming -
Proprietary + Confidential Proprietary + Confidential Part 1: The state
of Gaming in 2018 Part 2: UAC Optimization Strategies + Creatives Part 3: Revolutionize your strategy with Machine Learning Agenda -
Proprietary + Confidential Proprietary + Confidential Big numbers first Number
of gamers worldwide? -
Proprietary + Confidential Proprietary + Confidential Big numbers first Source:
NewZoo Global Gaming report Number of gamers worldwide 2.35 Billion -
Proprietary + Confidential Proprietary + Confidential Big numbers first Installs
acquired with Google? -
Proprietary + Confidential Proprietary + Confidential Big numbers first Source:
NewZoo Global Gaming report Installs acquired with Google? 10+ Billion -
Proprietary + Confidential Proprietary + Confidential Big numbers first Source:
NewZoo Global Gaming report Market capitalization? $137.9 Billion -
Proprietary + Confidential Proprietary + Confidential Three big things in
2018 Source: NewZoo Global Gaming report -
Proprietary + Confidential Proprietary + Confidential Three big things in
2018 Source: NewZoo Global Gaming report -
Proprietary + Confidential Proprietary + Confidential Three big things in
2018 Source: NewZoo Global Gaming report -
Proprietary + Confidential Proprietary + Confidential Three big things in
2018 Source: NewZoo Global Gaming report HYPERCASUAL -
Proprietary + Confidential Proprietary + Confidential 952M There were almost
1 Billion downloads of hypercasual games in 2016 Source: AppAnnie. -
Proprietary + Confidential Proprietary + Confidential 952M 1.52B +60% Whole
gaming: +13% And then, they start growing fast… More than 1.5B of downloads in 2017 Source: AppAnnie. -
Proprietary + Confidential Proprietary + Confidential 952M 1.52B 1.85B (E)
3.28B (E) +60% Whole gaming: +13% +116% ...like really fast. More than 3.2B downloads in 2018 is expected Source: AppAnnie. 3.28B (E) -
Proprietary + Confidential Proprietary + Confidential 952M 1.52B 1.85B (E)
3.28B (E) +60% Whole gaming: +13% +116% ...like really fast. More than 3.2B downloads in 2018 is expected Source: AppAnnie. Note: only apps, published by Playgendary, Voodoo, Ketchapp Studio, Playdots, Easybrain, Miniclip, Tap Tap Games, SuperTapx, Kwalee, TabTale, BitMango, BoomBit, FuturePlay, Green Panda, Homa 3.28B (E) -
Proprietary + Confidential Proprietary + Confidential Things behind growth •
CPM growth • Number of titles (tens and hundreds) -
Proprietary + Confidential Proprietary + Confidential Things behind growth •
CPM growth • Number of titles (tens and hundreds) • User Acquisition -
Proprietary + Confidential Proprietary + Confidential Things behind growth •
CPM growth • Number of titles (tens and hundreds) • User Acquisition • Cross-promo -
Proprietary + Confidential In 2018, mobile games will generate $70.3B
or 51% of Global Gaming market. -
Proprietary + Confidential Gaming already accounts for: • Over 90%
of Google Play Revenue. Gaming @ Google -
Proprietary + Confidential Starting on October 16th, all new app
install campaigns created in Google Ads run only through UAC Year ago... -
Proprietary + Confidential One tool to reach a multi-billion audience
through 7 major products & 2 OS / -
Proprietary + Confidential UAC became the most efficient campaign type
when the main goal is reach or when studios consider engagement and revenue as bigger priorities. -
Proprietary + Confidential App strategies & Goals Installs tCPI Actions
tCPA Value tROAS* UAC Installs UAC Installs, Advanced UAC Actions UAC Value -
Proprietary + Confidential Understanding the gaming user base segmentation Engagement
(Play Time) Low High Spend (ARPU) High Low Whales Sporadic Spenders Economical Enthusiasts Casual/New -
Proprietary + Confidential Looking for the right user segment mix
Casual/New Economical Enthusiasts Sporadic Spenders Whales Value Scale Having the right user segment mix is key for the success of the game The right user segment mix is function of • Studio strategy • Monetisation model • Game maturity -
Proprietary + Confidential Understanding your app’s strategy and goals User
segment Casual/New Economical Enthusiasts Spenders/Whales Goal Scale Engagement Revenue What are the associated KPIs? -
Proprietary + Confidential User segment Casual/New Economical Enthusiasts Spenders/Whales Goal
Scale Engagement Revenue What are the associated KPIs? - Max scale below tCPI - Max scale for a specific budget Understanding your app’s strategy and goals -
Proprietary + Confidential User segment Casual/New Economical Enthusiasts Spenders/Whales Goal
Scale Engagement Revenue What are the associated KPIs? - Max scale below tCPI - Max scale for a specific budget - Short term retention (D1, D3) - Med term retention (D7,14) Understanding your app’s strategy and goals -
Proprietary + Confidential User segment Casual/New Economical Enthusiasts Spenders/Whales Goal
Scale Engagement Revenue What are the associated KPIs? - Max scale below tCPI - Max scale for a specific budget - Short term retention (D1, D3) - Med term retention (D7,14) - Short term ROAS (D3, D7) - Med term ROAS (D14, D30) - Cost per purchaser - Cost per purchase at x days Understanding your app’s strategy and goals -
Proprietary + Confidential User segment Casual/New Economical Enthusiasts Spenders/Whales Goal
Scale Engagement Revenue What are the associated KPIs? - Max scale below tCPI - Max scale for a specific budget - Short term retention (D1, D3) - Med term retention (D7,14) - Short term ROAS (D3, D7) - Med term ROAS (D14, D30) - Cost per purchaser - Cost per purchase at x days Events to track Understanding your app’s strategy and goals -
Proprietary + Confidential User segment Casual/New Economical Enthusiasts Spenders/Whales Goal
Scale Engagement Revenue What are the associated KPIs? - Max scale below tCPI - Max scale for a specific budget - Short term retention (D1, D3) - Med term retention (D7,14) - Short term ROAS (D3, D7) - Med term ROAS (D14, D30) - Cost per purchaser - Cost per purchase at x days Events to track - Install /1st open Understanding your app’s strategy and goals -
Proprietary + Confidential User segment Casual/New Economical Enthusiasts Spenders/Whales Goal
Scale Engagement Revenue What are the associated KPIs? - Max scale below tCPI - Max scale for a specific budget - Short term retention (D1, D3) - Med term retention (D7,14) - Short term ROAS (D3, D7) - Med term ROAS (D14, D30) - Cost per purchaser - Cost per purchase at x days Events to track - Install /1st open - App open at Dx - Level X achieved - X sessions played - Tutorial completion Understanding your app’s strategy and goals -
Proprietary + Confidential User segment Casual/New Economical Enthusiasts Spenders/Whales Goal
Scale Engagement Revenue What are the associated KPIs? - Max scale below tCPI - Max scale for a specific budget - Short term retention (D1, D3) - Med term retention (D7,14) - Short term ROAS (D3, D7) - Med term ROAS (D14, D30) - Cost per purchaser - Cost per purchase at x days Events to track - Install /1st open - App open at Dx - Level X achieved - X sessions played - Tutorial completion - Purchases with value - 1st time purchase - Virtual currency spent - iAP shop visited Understanding your app’s strategy and goals -
Proprietary + Confidential User segment Casual/New Economical Enthusiasts Spenders/Whales Goal
Scale Engagement Revenue What are the associated KPIs? - Max scale below tCPI - Max scale for a specific budget - Short term retention (D1, D3) - Med term retention (D7,14) - Short term ROAS (D3, D7) - Med term ROAS (D14, D30) - Cost per purchaser - Cost per purchase at x days Events to track - 1st open - App open at Dx - Level X achieved - X sessions played - Tutorial completion - Purchases with value - 1st time purchase - Virtual currency spent - iAP shop visited Preferred UAC Mix UAC Installs UAC Installs Advanced Understanding your app’s strategy and goals -
Proprietary + Confidential User segment Casual/New Economical Enthusiasts Spenders/Whales Goal
Scale Engagement Revenue What are the associated KPIs? - Max scale below tCPI - Max scale for a specific budget - Short term retention (D1, D3) - Med term retention (D7,14) - Short term ROAS (D3, D7) - Med term ROAS (D14, D30) - Cost per purchaser - Cost per purchase at x days Events to track - 1st open - App open at Dx - Level X achieved - X sessions played - Tutorial completion - Purchases with value - 1st time purchase - Virtual currency spent - iAP shop visited Preferred UAC Mix UAC Installs UAC Installs Advanced UAC Actions (retention event) Understanding your app’s strategy and goals -
Proprietary + Confidential User segment Casual/New Economical Enthusiasts Spenders/Whales Goal
Scale Engagement Revenue What are the associated KPIs? - Max scale below tCPI - Max scale for a specific budget - Short term retention (D1, D3) - Med term retention (D7,14) - Short term ROAS (D3, D7) - Med term ROAS (D14, D30) - Cost per purchaser - Cost per purchase at x days Events to track - 1st open - App open at Dx - Level X achieved - X sessions played - Tutorial completion - Purchases with value - 1st time purchase - Virtual currency spent - iAP shop visited Preferred UAC Mix UAC Installs UAC Installs Advanced UAC Actions (retention event) UAC Actions (revenue event) UAC Value* Understanding your app’s strategy and goals -
Proprietary + Confidential How to stay in control of the
machine Bids Creative Event Selection Budget With less levers to pull, creative asset strategy is even more important to succeed with UAC -
Proprietary + Confidential Ad Format The specific format a creative
is rendered on (e.g. Interstitial AdMob or YouTube TrueView). Ad Formats are auto generated from different assets Asset An ad component such a text, image or a video. Can either be manual assets provided by advertisers or automatic assets pulled by UAC from either App / Play store. Definitions -
Proprietary + Confidential Shifting to an Asset Mindset in UAC
Serve via ad formats depending on network and inventory across all Google properties AdMob Interstitial AdMob Square AdMob Banner YouTube Native Mobile Web AdMob (all sizes) AdMob GIF YouTube TrueView AdMob Rewarded AdMob Native AdMob Portrait AdMob Square AdMob Landscape Mobile Web Text Input (Asset) Output (Ad Format) Play Homepage Play Browse Play Search Search Ads YouTube Native Gmail App Install AdMob (all sizes) Video Image HTML5 -
Proprietary + Confidential Maximize inventory and user reach via creative
asset mix 4 text ideas 20 videos 20 images 20 html5 -
Proprietary + Confidential Maximize inventory and user reach via creative
asset mix 4 text ideas Text lines in different order can appear as a separate ad, or they can combine with videos. As for images, the text lines can be matched only with landscape image (1200x628). Write standalone sentences in text lines because the system will automatically combine these to create the best text ad. -
Proprietary + Confidential Maximize inventory and user reach via creative
asset mix 20 images 320x50, 300x250, 320x480, 300x50, 728x90 (leaderboard), 480x320, 480x32, 320x100, 1024x768, 1200x628 (landscape image - recommended) - cover over 95% of AdMob inventory. Upload high quality (72dpi) images and HD ads (e.g. upload 640x100 for 320x50). -
Proprietary + Confidential Maximize inventory and user reach via creative
asset mix 20 html5 Recommended sizes (your assets will appear as full screen ads on the majority of devices): 320x480 (portrait) 480x320 (landscape) Mobile only (no Tablets) No audio / video media support Having a persistent “install now” button throughout is mandatory. -
Proprietary + Confidential Maximize inventory and user reach via creative
asset mix 20 videos We recommend to upload different versions of multiple assets across the top 3 recommended ratios. I.e. 10 landscape videos, 5 portrait and 5 square videos with different messages. YouTube is optimized for all video types. 70% of AdMob impressions are portrait. -
Proprietary + Confidential Google and Unity partner to boost mobile
game advertising August 2018 -
Proprietary + Confidential This deal offers a significant expansion of
Google Rewarded Video Supply at scale. -
Proprietary + Confidential UAC Betas UAC Value Offers the possibility
to tap into value based optimization and to acquire new users who are likely to have a set return within their selected conversion window, either based on: • Dynamic transactions • Or user centric dynamic pLTV via s2s -
Proprietary + Confidential UAC Betas UAC Value Offers the possibility
to tap into value based optimization and to acquire new users who are likely to have a set return within their selected conversion window, either based on: Similar Audience Device IDs as hints. ‘Feed’ UAC with additional ‘user’ information it currently does not have, but advertisers can provide. • Dynamic transactions • Or user centric dynamic pLTV via s2s -
Proprietary + Confidential UAC Betas UAC Value Offers the possibility
to tap into value based optimization and to acquire new users who are likely to have a set return within their selected conversion window, either based on: Similar Audience Device IDs as hints. ‘Feed’ UAC with additional ‘user’ information it currently does not have, but advertisers can provide. UACe UAC for Engagement is a dedicated, automated, cross-channel (Search, Play, Display, YouTube) campaign driving users back into the app to complete a specific action/conversion. • Dynamic transactions • Or user centric dynamic pLTV via s2s -
Proprietary + Confidential All Leads Clicks, app installs Buyers Top
LTV users Low v i n ra . (Aut -bi w b o ) Lon n si de . (on f t ~7 da c ) Hin : Ret e s se ‘Pro t Ca t’ in d -
Proprietary + Confidential All Leads Clicks, app installs We s
n of n o v on to A W or s at w si s n e t n u r Potential buyers Actual buyers -
Proprietary + Confidential Queries Placements Demo Targeting Device Targeting Granular
Location App Usage Value Session duration Usage App Settings Retention Referrals -
Proprietary + Confidential App installs pBuyers (potential) Buyers (actual) pBu
c er s a 10x hi r ve n a in tu $ t a n l pBu c er s e r to A W 3.5x so t a c a p h e pBu s ca r 80% of f u p ha s -
Proprietary + Confidential FEATURE TYPES: FEATURES: EXPLANATION: Required user_id Hashed
IDFA / Web_id of install / login_id as GA dimension age Number of days since install the data below is gathered for Numeric (0,1,2..) sessions Number of sessions by a certain day player_level in-game level reached (0+) battles Total In-game battles played messages Total in-app messages sent ... ... Binary (0 or 1) facebook_login Did the player use their facebook account for the app login? promo_used Did the player use any promo codes ... ... Categorical (A,B,C.) device User's device model, i.e. “Xiaomi” country User's location, i.e. “UK” ... ... -
Proprietary + Confidential Recommendations • Dataset and initial data-structure! •
Start with simple (e.g. linear regression) -
Proprietary + Confidential Recommendations • Dataset and initial data-structure! •
Start with simple (e.g. linear regression) • Collect features as much as you could. 10-20? 200-400-1000! ◦ Firebase, BigQuery, ClickHouse - some solutions to use -
Proprietary + Confidential Recommendations • Dataset and initial data-structure! •
Start with simple (e.g. linear regression) • Collect features as much as you could. 10-20? 200-400-1000! ◦ Firebase, BigQuery, ClickHouse - some solutions to use • Save historical data (no re-writings) -
Proprietary + Confidential Recommendations • Dataset and initial data-structure! •
Start with simple (e.g. linear regression) • Collect features as much as you could. 10-20? 200-400-1000! ◦ Firebase, BigQuery, ClickHouse - some solutions to use • Save historical data (no re-writings) • Define model quality metrics ◦ Precision / Recall ratio ◦ F1 score, Accuracy -
Proprietary + Confidential Recommendations • Dataset and initial data-structure! •
Start with simple (e.g. linear regression) • Collect features as much as you could. 10-20? 200-400-1000! ◦ Firebase, BigQuery, ClickHouse - some solutions to use • Save historical data (no re-writings) • Define model quality metrics ◦ Precision / Recall ratio ◦ F1 score, Accuracy • Revise your model regularly -
Proprietary + Confidential Thank you! Olesya Kozlovskaya | Analytical Consultant,
Gaming Arseniy Alikhanov | Analytical Consultant, Gaming