Lior Barak (Co-founder at Tale About Data - data consulting) talks about the common mistakes to avoid in order to leverage data for success.
Once you define your purpose as a team at the high-level, you can start asking which data you need to use to make your decisions and serve your purpose.
Having a data strategy means that you know your one true source of installs. You start with the different sources (example: MMP, unique devices in back-end, etc.) and decide what is your true source.
Split your budget: 85% of your budget decisions should be based on true data, and you can be creative with the rest of it.
Try to create smaller and granular campaigns with less budget because this will make you much more focused. The more metadata you can collect on your campaigns (country, language, age group) the more you will understand your data and make better decisions.
You need to define your creative metadata: the elements inside your creatives (e.g. showing the product, a person, etc.) so your designers can be more focused and so you can take better decisions.
You data engineer needs to have a quality insurance system that checks your data everyday and makes sure that it's correct. You should receive an email with "yes you can trust the data" or "there has been an issue".
Use only 3 KPIs on a daily basis to make decisions on increasing/decreasing budget, targeting or not users, etc. Use a weekly report with 10 KPIs to explain your decision during the week. Reduce the noise from the data. Do a monthly report with another 10 KPIs to communicate on how you invested the money.
Invite the engineer to sit with you, become a part of the team and see what you do so he can build what you need.
Enable yourself to consume the data you need when you need it in a very easy way. A lot of people are waiting for their analyst to build dashboards or reporting for them even though they could have done it themselves. Take a course and learn how to build your report.
The job of the analyst is to explain to you issues that you can not explain yourself: going into the data and finding answers for you. It's a different purpose than dashboard/reporting.
Don't start acquiring users before you understand the user path. It will help you bring better users and better understand how much you can actually invest in acquiring users.
Don't change your reports on a daily/weekly basis (KPIs, data visualization, etc.), change them on a quarterly basis or even every year. Try to work with them as much as you can to make your decisions (then make changes if needed).
Once you define your purpose as a team at the high-level, you can start asking which data you need to use to make your decisions and serve your purpose.
Having a data strategy means that you know your one true source of installs. You start with the different sources (example: MMP, unique devices in back-end, etc.) and decide what is your true source.
Split your budget: 85% of your budget decisions should be based on true data, and you can be creative with the rest of it.
Try to create smaller and granular campaigns with less budget because this will make you much more focused. The more metadata you can collect on your campaigns (country, language, age group) the more you will understand your data and make better decisions.
You need to define your creative metadata: the elements inside your creatives (e.g. showing the product, a person, etc.) so your designers can be more focused and so you can take better decisions.
You data engineer needs to have a quality insurance system that checks your data everyday and makes sure that it's correct. You should receive an email with "yes you can trust the data" or "there has been an issue".
Use only 3 KPIs on a daily basis to make decisions on increasing/decreasing budget, targeting or not users, etc. Use a weekly report with 10 KPIs to explain your decision during the week. Reduce the noise from the data. Do a monthly report with another 10 KPIs to communicate on how you invested the money.
Invite the engineer to sit with you, become a part of the team and see what you do so he can build what you need.
Enable yourself to consume the data you need when you need it in a very easy way. A lot of people are waiting for their analyst to build dashboards or reporting for them even though they could have done it themselves. Take a course and learn how to build your report.
The job of the analyst is to explain to you issues that you can not explain yourself: going into the data and finding answers for you. It's a different purpose than dashboard/reporting.
Don't start acquiring users before you understand the user path. It will help you bring better users and better understand how much you can actually invest in acquiring users.
Don't change your reports on a daily/weekly basis (KPIs, data visualization, etc.), change them on a quarterly basis or even every year. Try to work with them as much as you can to make your decisions (then make changes if needed).
Once you define your purpose as a team at the high-level, you can start asking which data you need to use to make your decisions and serve your purpose.
Having a data strategy means that you know your one true source of installs. You start with the different sources (example: MMP, unique devices in back-end, etc.) and decide what is your true source.
Split your budget: 85% of your budget decisions should be based on true data, and you can be creative with the rest of it.
Try to create smaller and granular campaigns with less budget because this will make you much more focused. The more metadata you can collect on your campaigns (country, language, age group) the more you will understand your data and make better decisions.
You need to define your creative metadata: the elements inside your creatives (e.g. showing the product, a person, etc.) so your designers can be more focused and so you can take better decisions.
You data engineer needs to have a quality insurance system that checks your data everyday and makes sure that it's correct. You should receive an email with "yes you can trust the data" or "there has been an issue".
Use only 3 KPIs on a daily basis to make decisions on increasing/decreasing budget, targeting or not users, etc. Use a weekly report with 10 KPIs to explain your decision during the week. Reduce the noise from the data. Do a monthly report with another 10 KPIs to communicate on how you invested the money.
Invite the engineer to sit with you, become a part of the team and see what you do so he can build what you need.
Enable yourself to consume the data you need when you need it in a very easy way. A lot of people are waiting for their analyst to build dashboards or reporting for them even though they could have done it themselves. Take a course and learn how to build your report.
The job of the analyst is to explain to you issues that you can not explain yourself: going into the data and finding answers for you. It's a different purpose than dashboard/reporting.
Don't start acquiring users before you understand the user path. It will help you bring better users and better understand how much you can actually invest in acquiring users.
Don't change your reports on a daily/weekly basis (KPIs, data visualization, etc.), change them on a quarterly basis or even every year. Try to work with them as much as you can to make your decisions (then make changes if needed).
Notes for this resource are currently being transferred and will be available soon.
You need to ask yourself: what is the purpose of me (performance, CRM, etc.) being here? What are the users you are trying to bring on and the purpose of the money you're spending?
[💎@01:33] Once you define your purpose as a team at the high-level, you can start asking which data you need to use to make your decisions and serve your purpose.
Read Lior's book ;)
[💎@02:48] Having a data strategy means that you know your one true source of installs. You start with the different sources (example: MMP, unique devices in back-end, etc.) and decide what is your true source.
The problem is about trusting the data, not about the data itself.
[💎@04:33] Split your budget: 85% of your budget decisions should be based on true data, and you can be creative with the rest of it.
Once you have a fixed budget for experiments you become much more sensitive on each $ you spend.
A lot of organizations set up campaigns in the entire Europe or the entire US or even worldwide.
Metadata. Metadata. Metadata.
Make sure you know when things go sideways. You should not be the one comparing your FB numbers and what your data engineering team is putting together.
[💎@07:40] You data engineer needs to have a quality insurance system that checks your data everyday and makes sure that it's correct. You should receive an email with "yes you can trust the data" or "there has been an issue".
[💎@08:23] Use only 3 KPIs on a daily basis to make decisions on increasing/decreasing budget, targeting or not users, etc. Use a weekly report with 10 KPIs to explain your decision during the week. Reduce the noise from the data. Do a monthly report with another 10 KPIs to communicate on how you invested the money.
[💎@09:54] Invite the engineer to sit with you, become a part of the team and see what you do so he can build what you need.
Enable yourself to consume the data you need when you need it in a very easy way
[💎@10:52] Enable yourself to consume the data you need when you need it in a very easy way. A lot of people are waiting for their analyst to build dashboards or reporting for them even though they could have done it themselves. Take a course and learn how to build your report.
[💎@11:30] The job of the analyst is to explain to you issues that you can not explain yourself: going into the data and finding answers for you. It's a different purpose than dashboard/reporting.
[💎@11:55] Don't start acquiring users before you understand the user path. It will help you bring better users and better understand how much you can actually invest in acquiring users.
Once you acquire the user, what are the points that he needs to do. Browse the products? Just order?
[💎@13:55] Don't change your reports on a daily/weekly basis (KPIs, data visualization, etc.), change them on a quarterly basis or even every year. Try to work with them as much as you can to make your decisions (then make changes if needed).
You need to ask yourself: what is the purpose of me (performance, CRM, etc.) being here? What are the users you are trying to bring on and the purpose of the money you're spending?
[💎@01:33] Once you define your purpose as a team at the high-level, you can start asking which data you need to use to make your decisions and serve your purpose.
Read Lior's book ;)
[💎@02:48] Having a data strategy means that you know your one true source of installs. You start with the different sources (example: MMP, unique devices in back-end, etc.) and decide what is your true source.
The problem is about trusting the data, not about the data itself.
[💎@04:33] Split your budget: 85% of your budget decisions should be based on true data, and you can be creative with the rest of it.
Once you have a fixed budget for experiments you become much more sensitive on each $ you spend.
A lot of organizations set up campaigns in the entire Europe or the entire US or even worldwide.
Metadata. Metadata. Metadata.
Make sure you know when things go sideways. You should not be the one comparing your FB numbers and what your data engineering team is putting together.
[💎@07:40] You data engineer needs to have a quality insurance system that checks your data everyday and makes sure that it's correct. You should receive an email with "yes you can trust the data" or "there has been an issue".
[💎@08:23] Use only 3 KPIs on a daily basis to make decisions on increasing/decreasing budget, targeting or not users, etc. Use a weekly report with 10 KPIs to explain your decision during the week. Reduce the noise from the data. Do a monthly report with another 10 KPIs to communicate on how you invested the money.
[💎@09:54] Invite the engineer to sit with you, become a part of the team and see what you do so he can build what you need.
Enable yourself to consume the data you need when you need it in a very easy way
[💎@10:52] Enable yourself to consume the data you need when you need it in a very easy way. A lot of people are waiting for their analyst to build dashboards or reporting for them even though they could have done it themselves. Take a course and learn how to build your report.
[💎@11:30] The job of the analyst is to explain to you issues that you can not explain yourself: going into the data and finding answers for you. It's a different purpose than dashboard/reporting.
[💎@11:55] Don't start acquiring users before you understand the user path. It will help you bring better users and better understand how much you can actually invest in acquiring users.
Once you acquire the user, what are the points that he needs to do. Browse the products? Just order?
[💎@13:55] Don't change your reports on a daily/weekly basis (KPIs, data visualization, etc.), change them on a quarterly basis or even every year. Try to work with them as much as you can to make your decisions (then make changes if needed).
You need to ask yourself: what is the purpose of me (performance, CRM, etc.) being here? What are the users you are trying to bring on and the purpose of the money you're spending?
[💎@01:33] Once you define your purpose as a team at the high-level, you can start asking which data you need to use to make your decisions and serve your purpose.
Read Lior's book ;)
[💎@02:48] Having a data strategy means that you know your one true source of installs. You start with the different sources (example: MMP, unique devices in back-end, etc.) and decide what is your true source.
The problem is about trusting the data, not about the data itself.
[💎@04:33] Split your budget: 85% of your budget decisions should be based on true data, and you can be creative with the rest of it.
Once you have a fixed budget for experiments you become much more sensitive on each $ you spend.
A lot of organizations set up campaigns in the entire Europe or the entire US or even worldwide.
Metadata. Metadata. Metadata.
Make sure you know when things go sideways. You should not be the one comparing your FB numbers and what your data engineering team is putting together.
[💎@07:40] You data engineer needs to have a quality insurance system that checks your data everyday and makes sure that it's correct. You should receive an email with "yes you can trust the data" or "there has been an issue".
[💎@08:23] Use only 3 KPIs on a daily basis to make decisions on increasing/decreasing budget, targeting or not users, etc. Use a weekly report with 10 KPIs to explain your decision during the week. Reduce the noise from the data. Do a monthly report with another 10 KPIs to communicate on how you invested the money.
[💎@09:54] Invite the engineer to sit with you, become a part of the team and see what you do so he can build what you need.
Enable yourself to consume the data you need when you need it in a very easy way
[💎@10:52] Enable yourself to consume the data you need when you need it in a very easy way. A lot of people are waiting for their analyst to build dashboards or reporting for them even though they could have done it themselves. Take a course and learn how to build your report.
[💎@11:30] The job of the analyst is to explain to you issues that you can not explain yourself: going into the data and finding answers for you. It's a different purpose than dashboard/reporting.
[💎@11:55] Don't start acquiring users before you understand the user path. It will help you bring better users and better understand how much you can actually invest in acquiring users.
Once you acquire the user, what are the points that he needs to do. Browse the products? Just order?
[💎@13:55] Don't change your reports on a daily/weekly basis (KPIs, data visualization, etc.), change them on a quarterly basis or even every year. Try to work with them as much as you can to make your decisions (then make changes if needed).
