Warren Woodward (CGO at Upptic) has extensive experience in automating UA processes and shares how to get started and what he learned.
An intermediary step before going with full automation can be to download the rules of a redundant task as a .csv, run a Python script and get the output that you can upload to the ad network. It can already save up 80% of the time.
Getting started with automation:
1. Identify a need and what's eating up your time,
2. Identify the infrastructure you need (data format you need),
3. Figure out if it's buy or build
If you are currently doing something day-to-day in a spreadsheet, it is probably a good candidate for automation.
Be ok with a "working solution" if it already helps you save half of your time. It doesn't have to be perfect and you can refine later.
When it comes to creatives, machines might not create ideas but can be amazing loops for feedback and iteration.
An intermediary step before going with full automation can be to download the rules of a redundant task as a .csv, run a Python script and get the output that you can upload to the ad network. It can already save up 80% of the time.
Getting started with automation:
1. Identify a need and what's eating up your time,
2. Identify the infrastructure you need (data format you need),
3. Figure out if it's buy or build
If you are currently doing something day-to-day in a spreadsheet, it is probably a good candidate for automation.
Be ok with a "working solution" if it already helps you save half of your time. It doesn't have to be perfect and you can refine later.
When it comes to creatives, machines might not create ideas but can be amazing loops for feedback and iteration.
An intermediary step before going with full automation can be to download the rules of a redundant task as a .csv, run a Python script and get the output that you can upload to the ad network. It can already save up 80% of the time.
Getting started with automation:
1. Identify a need and what's eating up your time,
2. Identify the infrastructure you need (data format you need),
3. Figure out if it's buy or build
If you are currently doing something day-to-day in a spreadsheet, it is probably a good candidate for automation.
Be ok with a "working solution" if it already helps you save half of your time. It doesn't have to be perfect and you can refine later.
When it comes to creatives, machines might not create ideas but can be amazing loops for feedback and iteration.
Notes for this resource are currently being transferred and will be available soon.
Start by thinking of "what's really stupid?" that you do on a day-to-day basis: something that time consuming.
Automation doesn't have to be a "blackbox". It can be as simple as automated reports, or simple UA rules for an ad network.
Some partners/ad networks offer automation rules. But the goal of that partner is most likely to sell its inventory.
Building in-house gives you ownership of the system.
Have the UA manager write what he does to automate campaigns:
[💎@07:54] Intermediary step before full automation: download the rules as a .csv, run a Python script and get the output that you can upload to the ad network. Can already save up 80% of the time.
The most important is to identify the need, then you can identify a provider you can work with in the space.
Identify the goal, then figure out if it's buy or build.
Before automation things were much more stressful because UA never sleeps. Automation allows you to refocus your team to work more closely with other teams in your company.
[💎@11:45]
[💎@13:15] If you are currently doing something day-to-day in a spreadsheet, it is probably a good candidate for automation.
For UA, automation is typically used to automate the most redundant task on:
It starts with verifying that the data is clean.
An area often overlooked: are you comparing what the App Store paid you with what you actually got paid.
You need a "first line of defense" of basic data governance (someone checking the data).
[💎@16:57] Start with clean data then figure out how to implement the system.
You do not need an engineering team to get started: "what's the most simple version of this".
[💎@18:15] Be ok with a "working solution" if it already helps you save half of your time. It doesn't have to be perfect and you can refine later.
Not many valid solutions to automate App Store Optimization (what Upptic does): split-testing process, combining elements, etc.
[💎@19:35] When it comes to creatives, machines might not create ideas but can be amazing loops for feedback and iteration.
Start by thinking of "what's really stupid?" that you do on a day-to-day basis: something that time consuming.
Automation doesn't have to be a "blackbox". It can be as simple as automated reports, or simple UA rules for an ad network.
Some partners/ad networks offer automation rules. But the goal of that partner is most likely to sell its inventory.
Building in-house gives you ownership of the system.
Have the UA manager write what he does to automate campaigns:
[💎@07:54] Intermediary step before full automation: download the rules as a .csv, run a Python script and get the output that you can upload to the ad network. Can already save up 80% of the time.
The most important is to identify the need, then you can identify a provider you can work with in the space.
Identify the goal, then figure out if it's buy or build.
Before automation things were much more stressful because UA never sleeps. Automation allows you to refocus your team to work more closely with other teams in your company.
[💎@11:45]
[💎@13:15] If you are currently doing something day-to-day in a spreadsheet, it is probably a good candidate for automation.
For UA, automation is typically used to automate the most redundant task on:
It starts with verifying that the data is clean.
An area often overlooked: are you comparing what the App Store paid you with what you actually got paid.
You need a "first line of defense" of basic data governance (someone checking the data).
[💎@16:57] Start with clean data then figure out how to implement the system.
You do not need an engineering team to get started: "what's the most simple version of this".
[💎@18:15] Be ok with a "working solution" if it already helps you save half of your time. It doesn't have to be perfect and you can refine later.
Not many valid solutions to automate App Store Optimization (what Upptic does): split-testing process, combining elements, etc.
[💎@19:35] When it comes to creatives, machines might not create ideas but can be amazing loops for feedback and iteration.
Start by thinking of "what's really stupid?" that you do on a day-to-day basis: something that time consuming.
Automation doesn't have to be a "blackbox". It can be as simple as automated reports, or simple UA rules for an ad network.
Some partners/ad networks offer automation rules. But the goal of that partner is most likely to sell its inventory.
Building in-house gives you ownership of the system.
Have the UA manager write what he does to automate campaigns:
[💎@07:54] Intermediary step before full automation: download the rules as a .csv, run a Python script and get the output that you can upload to the ad network. Can already save up 80% of the time.
The most important is to identify the need, then you can identify a provider you can work with in the space.
Identify the goal, then figure out if it's buy or build.
Before automation things were much more stressful because UA never sleeps. Automation allows you to refocus your team to work more closely with other teams in your company.
[💎@11:45]
[💎@13:15] If you are currently doing something day-to-day in a spreadsheet, it is probably a good candidate for automation.
For UA, automation is typically used to automate the most redundant task on:
It starts with verifying that the data is clean.
An area often overlooked: are you comparing what the App Store paid you with what you actually got paid.
You need a "first line of defense" of basic data governance (someone checking the data).
[💎@16:57] Start with clean data then figure out how to implement the system.
You do not need an engineering team to get started: "what's the most simple version of this".
[💎@18:15] Be ok with a "working solution" if it already helps you save half of your time. It doesn't have to be perfect and you can refine later.
Not many valid solutions to automate App Store Optimization (what Upptic does): split-testing process, combining elements, etc.
[💎@19:35] When it comes to creatives, machines might not create ideas but can be amazing loops for feedback and iteration.