ืืชืื ืืืฆื ืื ืืงืืื ืขื ืืืคืืืงืฆืื Player FM !
๐๏ธ How to use econometric models to supplement SKAdNetwork for mobile measurement in iOS 14 - with Brian Krebs (MetricWorks), and Anthony Cross (formerly at Big Fish Games) โ๏ธ
Manage episode 278922760 series 2575608
Our guests today are Brian Krebs and Anthony Cross. Brian is the founder and CEO of MetricWorks, and Anthony formerly at Big Fish Games. In todayโs conversation they offer very interesting perspectives on why SKAdNetwork isnโt enough - and what it needs to be supplemented by in order for mobile measurement to truly reflect the value of marketing efforts.
In todayโs conversation, we reflect on why measurement in a world where marketing is influenced by multiple variables has no clear and easy answers - certainly none seemingly as simple as the solutions that deterministic measurement offered. The solution then is to embrace the multiple variables involved - and use an approach that is a mix of SKAdNetwork, econometric modeling and intelligent experimentation.
Key highlights:
๐ Why last touch attribution does not give the whole picture
๐๏ธโโ๏ธ Incrementality is THE criteria for marketing effectiveness
๐ค The limitations of SKAdNetwork signals for measurement
๐งฎ Top statistics-based approaches to measuring incrementality that work
๐ผ How statistical models rely on IDFAs to be useful
๐งจ No-IDFA is going to blow up the measurement landscape
๐ Big Fish used multiple sources of data in a custom dashboard to solve for attribution
๐ก How Big Fish used data science to get marketing insights
๐ฆ Using data science to comparing LTV curves through a gaming app launch cycle to amp up ROAS
โ๏ธ Reverse engineering LTV curves to inform ad spend decisions
๐บ๏ธ Mapping in-app events to monetisation and retention opportunities
๐ธ Spend needs to tie back to business, and LTV is the only way to do that
๐ Where to apply econometric modelling in marketing
๐จ How to use media mix models for marketing
๐๏ธ How to build models that calculate impact and outcomes
๐ฏ Micro and macro level factors to consider for accurate predictions
๐ How seasonality and trends impact lift
โ๏ธ Tweaking factors to generate insights about what-if scenarios
๐ The journey from deterministic to probabilistic has already begun
๐งช Testing successful campaigns with interrupted time series can give counterintuitive results
๐ช How the post-IDFA landscape is strengthening the case for incrementality testing
๐ป The opportunity cost for interrupted time series experiments
โ๏ธ The careful balance in ITS experiments
๐ Why models are necessary alongside experiments for real validation
๐งฑ How to construct an econometric model by choosing the right variables
๐
Why daily data is better than weekly data for training models in spite of the increased noise
๐ The value of seasonality and trends encoding in models
๐ฎ The key to good prediction for granularity in each event
๐ฌ Understanding counterfactual experiments and setting baselines
๐ฅ Why incrementality testing is the gold standard now but not post-IDFA
๐ฌ๐ป The two-fold team buy-ins for testing
๐ค Why transparent partners are very important to a UA team
๐ Why experiments are tied to a certain level of scale
๐ผ Why single apps donโt have to worry about incrementality till scale
๐ The tried and tested approach to learning how to model
Check out the show notes here: https://mobileuseracquisitionshow.com/episode/how-to-use-econometric-models-to-supplement-skadnetwork-for-mobile-measurement-in-ios-14-with-brian-krebs-and-anthony-cross/
**
Get more mobile user acquisition goodies here:
http://RocketShipHQ.com
http://RocketShipHQ.com/blog
**
Check out our podcast featuring inside stories of how technology evolves and grows:
http://HowThingsGrow.co
251 ืคืจืงืื
Manage episode 278922760 series 2575608
Our guests today are Brian Krebs and Anthony Cross. Brian is the founder and CEO of MetricWorks, and Anthony formerly at Big Fish Games. In todayโs conversation they offer very interesting perspectives on why SKAdNetwork isnโt enough - and what it needs to be supplemented by in order for mobile measurement to truly reflect the value of marketing efforts.
In todayโs conversation, we reflect on why measurement in a world where marketing is influenced by multiple variables has no clear and easy answers - certainly none seemingly as simple as the solutions that deterministic measurement offered. The solution then is to embrace the multiple variables involved - and use an approach that is a mix of SKAdNetwork, econometric modeling and intelligent experimentation.
Key highlights:
๐ Why last touch attribution does not give the whole picture
๐๏ธโโ๏ธ Incrementality is THE criteria for marketing effectiveness
๐ค The limitations of SKAdNetwork signals for measurement
๐งฎ Top statistics-based approaches to measuring incrementality that work
๐ผ How statistical models rely on IDFAs to be useful
๐งจ No-IDFA is going to blow up the measurement landscape
๐ Big Fish used multiple sources of data in a custom dashboard to solve for attribution
๐ก How Big Fish used data science to get marketing insights
๐ฆ Using data science to comparing LTV curves through a gaming app launch cycle to amp up ROAS
โ๏ธ Reverse engineering LTV curves to inform ad spend decisions
๐บ๏ธ Mapping in-app events to monetisation and retention opportunities
๐ธ Spend needs to tie back to business, and LTV is the only way to do that
๐ Where to apply econometric modelling in marketing
๐จ How to use media mix models for marketing
๐๏ธ How to build models that calculate impact and outcomes
๐ฏ Micro and macro level factors to consider for accurate predictions
๐ How seasonality and trends impact lift
โ๏ธ Tweaking factors to generate insights about what-if scenarios
๐ The journey from deterministic to probabilistic has already begun
๐งช Testing successful campaigns with interrupted time series can give counterintuitive results
๐ช How the post-IDFA landscape is strengthening the case for incrementality testing
๐ป The opportunity cost for interrupted time series experiments
โ๏ธ The careful balance in ITS experiments
๐ Why models are necessary alongside experiments for real validation
๐งฑ How to construct an econometric model by choosing the right variables
๐
Why daily data is better than weekly data for training models in spite of the increased noise
๐ The value of seasonality and trends encoding in models
๐ฎ The key to good prediction for granularity in each event
๐ฌ Understanding counterfactual experiments and setting baselines
๐ฅ Why incrementality testing is the gold standard now but not post-IDFA
๐ฌ๐ป The two-fold team buy-ins for testing
๐ค Why transparent partners are very important to a UA team
๐ Why experiments are tied to a certain level of scale
๐ผ Why single apps donโt have to worry about incrementality till scale
๐ The tried and tested approach to learning how to model
Check out the show notes here: https://mobileuseracquisitionshow.com/episode/how-to-use-econometric-models-to-supplement-skadnetwork-for-mobile-measurement-in-ios-14-with-brian-krebs-and-anthony-cross/
**
Get more mobile user acquisition goodies here:
http://RocketShipHQ.com
http://RocketShipHQ.com/blog
**
Check out our podcast featuring inside stories of how technology evolves and grows:
http://HowThingsGrow.co
251 ืคืจืงืื
ืื ืืคืจืงืื
×ืืจืืืื ืืืืื ืื Player FM!
Player FM ืกืืจืง ืืช ืืืื ืืจื ื ืขืืืจ ืคืืืงืืกืืื ืืืืืืช ืืืืื ืืฉืืืืื ืืื ืฉืชืื ื ืืื ืืจืืข. ืื ืืืฉืื ืืคืืืงืืกื ืืืื ืืืืชืจ ืืืื ืขืืื ืขื ืื ืืจืืืื, iPhone ืืืื ืืจื ื. ืืืจืฉืื ืืกื ืืจืื ืื ืืืื ืืืืฉืืจืื ืฉืื ืื.