Pau Quevedo (Head of Programmatic Trading at Goodgame - various free-to-play mobile games) talks with Peggy Anne Salz (Lead Analyst & Founder at Mobile Groove) about why you should be running DSP for mobile UA, the trials and tribulations of inhousing DSP, and what inventory works best for mobile gaming.
Leverage user-level data to develop and ID graph that allows them to follow users precisely. So they try to find users in video ad networks and follow them so they can acquire them elsewhere where CPMs are way lower.
The potential of in-housing DSP comes when you merge the BI data (all the post-install data) and the data from the auction.
Right now they are working with data based on the waterfall system. The minute that this changed they won't be able to apply their models so they need to be mindful of this.
For 2020: looking forward to header bidding and transparency to be able to reduce noise when it comes to the auction and be more independent. It will help develop all the other tools (like the user ID graph).
Header bidding is not going to work until publishers decide that they want to monetize through header bidding. Some publishers have seen products like Mopub and MAX bring a lower return, so they stop it.
Leverage user-level data to develop and ID graph that allows them to follow users precisely. So they try to find users in video ad networks and follow them so they can acquire them elsewhere where CPMs are way lower.
The potential of in-housing DSP comes when you merge the BI data (all the post-install data) and the data from the auction.
Right now they are working with data based on the waterfall system. The minute that this changed they won't be able to apply their models so they need to be mindful of this.
For 2020: looking forward to header bidding and transparency to be able to reduce noise when it comes to the auction and be more independent. It will help develop all the other tools (like the user ID graph).
Header bidding is not going to work until publishers decide that they want to monetize through header bidding. Some publishers have seen products like Mopub and MAX bring a lower return, so they stop it.
Leverage user-level data to develop and ID graph that allows them to follow users precisely. So they try to find users in video ad networks and follow them so they can acquire them elsewhere where CPMs are way lower.
The potential of in-housing DSP comes when you merge the BI data (all the post-install data) and the data from the auction.
Right now they are working with data based on the waterfall system. The minute that this changed they won't be able to apply their models so they need to be mindful of this.
For 2020: looking forward to header bidding and transparency to be able to reduce noise when it comes to the auction and be more independent. It will help develop all the other tools (like the user ID graph).
Header bidding is not going to work until publishers decide that they want to monetize through header bidding. Some publishers have seen products like Mopub and MAX bring a lower return, so they stop it.
Notes for this resource are currently being transferred and will be available soon.
Currently they are running various managed services and are trying to in-house mobile DSPs (already have one). The long term perspective is to build their own custom algorithm.
Evaluating DSP candidates:
With DSP, try to put everything under one roof. Because teams are way smaller than FB/Google, it is hard to measure precisely the impact or evaluate exactly when they start "cannibilizing" organic downloads.
First have to both teach the marketers to understand the data and get a marketing mindset into the data scientists.
There are also challenges with GDPR and the tech needed to model all that data. So you need to have a proof of concept to show management the potential.
DSPs do not have access to the same post-install data: they only get the events that are pushed. [💎@13:22] So the potential of in-housing DSP comes when you merge the BI data (all the post-install data) and the data from the auction.
This is why the DSPs working with user-level data turn to a SaaS-company to get more data from the advertisers to get more post-install data.
When launching campaigns, you are not just paying for the campaign but also to train the algorithm for you and other advertisers.
The more you the DSPs improve, the harder it will be to compete with them and in-house it
"we are feeding another beast"
They have some kind of Data Management Platform built.
"a gaming company is basically like a data company"
So far, haven't used the data to make the best out of it. Need to figure out how to structure the data and to actually gain (marketing insights) out of it.
[💎@20:49] For 2020: looking forward to header bidding and transparency to be able to reduce noise when it comes to the auction and be more independent. It will help develop all the other tools (like the user ID graph).
[💎@23:15] Header bidding is not going to work until publishers decide that they want to monetize through header bidding. Some publishers have seen products like Mopub and MAX bring a lower return, so they stop it.
Pau mentioned the UA Society event and a slack group focused on DSPs.
Follow-up comment by Pau (reach out to him on LinkedIn to get an invite): Slack group is one we've created for Advertisers who want to learn more on Mobile Programmatic, with a focus on UA. Most of the members are gaming and includes many famous companies like Kolibri, Wooga, Huuuge, Ubisoft, Innogames, Small Giant, Miniclip, etc... I can invite people over if they wish to join, doors are open. The Slack is called Programmatic Round Table. And we try to meet every now and then to discuss topics, the first meeting was in London, then some met again in Berlin, next ones will be Hamburg and London.
Currently they are running various managed services and are trying to in-house mobile DSPs (already have one). The long term perspective is to build their own custom algorithm.
Evaluating DSP candidates:
With DSP, try to put everything under one roof. Because teams are way smaller than FB/Google, it is hard to measure precisely the impact or evaluate exactly when they start "cannibilizing" organic downloads.
First have to both teach the marketers to understand the data and get a marketing mindset into the data scientists.
There are also challenges with GDPR and the tech needed to model all that data. So you need to have a proof of concept to show management the potential.
DSPs do not have access to the same post-install data: they only get the events that are pushed. [💎@13:22] So the potential of in-housing DSP comes when you merge the BI data (all the post-install data) and the data from the auction.
This is why the DSPs working with user-level data turn to a SaaS-company to get more data from the advertisers to get more post-install data.
When launching campaigns, you are not just paying for the campaign but also to train the algorithm for you and other advertisers.
The more you the DSPs improve, the harder it will be to compete with them and in-house it
"we are feeding another beast"
They have some kind of Data Management Platform built.
"a gaming company is basically like a data company"
So far, haven't used the data to make the best out of it. Need to figure out how to structure the data and to actually gain (marketing insights) out of it.
[💎@20:49] For 2020: looking forward to header bidding and transparency to be able to reduce noise when it comes to the auction and be more independent. It will help develop all the other tools (like the user ID graph).
[💎@23:15] Header bidding is not going to work until publishers decide that they want to monetize through header bidding. Some publishers have seen products like Mopub and MAX bring a lower return, so they stop it.
Pau mentioned the UA Society event and a slack group focused on DSPs.
Follow-up comment by Pau (reach out to him on LinkedIn to get an invite): Slack group is one we've created for Advertisers who want to learn more on Mobile Programmatic, with a focus on UA. Most of the members are gaming and includes many famous companies like Kolibri, Wooga, Huuuge, Ubisoft, Innogames, Small Giant, Miniclip, etc... I can invite people over if they wish to join, doors are open. The Slack is called Programmatic Round Table. And we try to meet every now and then to discuss topics, the first meeting was in London, then some met again in Berlin, next ones will be Hamburg and London.
Currently they are running various managed services and are trying to in-house mobile DSPs (already have one). The long term perspective is to build their own custom algorithm.
Evaluating DSP candidates:
With DSP, try to put everything under one roof. Because teams are way smaller than FB/Google, it is hard to measure precisely the impact or evaluate exactly when they start "cannibilizing" organic downloads.
First have to both teach the marketers to understand the data and get a marketing mindset into the data scientists.
There are also challenges with GDPR and the tech needed to model all that data. So you need to have a proof of concept to show management the potential.
DSPs do not have access to the same post-install data: they only get the events that are pushed. [💎@13:22] So the potential of in-housing DSP comes when you merge the BI data (all the post-install data) and the data from the auction.
This is why the DSPs working with user-level data turn to a SaaS-company to get more data from the advertisers to get more post-install data.
When launching campaigns, you are not just paying for the campaign but also to train the algorithm for you and other advertisers.
The more you the DSPs improve, the harder it will be to compete with them and in-house it
"we are feeding another beast"
They have some kind of Data Management Platform built.
"a gaming company is basically like a data company"
So far, haven't used the data to make the best out of it. Need to figure out how to structure the data and to actually gain (marketing insights) out of it.
[💎@20:49] For 2020: looking forward to header bidding and transparency to be able to reduce noise when it comes to the auction and be more independent. It will help develop all the other tools (like the user ID graph).
[💎@23:15] Header bidding is not going to work until publishers decide that they want to monetize through header bidding. Some publishers have seen products like Mopub and MAX bring a lower return, so they stop it.
Pau mentioned the UA Society event and a slack group focused on DSPs.
Follow-up comment by Pau (reach out to him on LinkedIn to get an invite): Slack group is one we've created for Advertisers who want to learn more on Mobile Programmatic, with a focus on UA. Most of the members are gaming and includes many famous companies like Kolibri, Wooga, Huuuge, Ubisoft, Innogames, Small Giant, Miniclip, etc... I can invite people over if they wish to join, doors are open. The Slack is called Programmatic Round Table. And we try to meet every now and then to discuss topics, the first meeting was in London, then some met again in Berlin, next ones will be Hamburg and London.