Newbie to mobile marketing

[bryan]

Hi All, great to join affiliateFix, a place with a lot of generous people and asking for no subscription fee, this should give many thumb up!

My CPA journey started one year ago focused on desktop PPV, but as overall i making a loss like -50% ROI. Anyway i learned some coding for landing page, this is the plus side. Had subscribed paid forum, spy tools and whatever i consider should have but still have no fate.

I am ventured into mobile CPA around 3 months ago after reading Charles blog. I am still not giving up on CPA, haha. After reading around and getting admin stuff done, I had created 2 campaigns with one CPI offer (payout $0.30, i found it is very low compared to PPV offer previously) around 2 weeks+ ago. End up crying result -- spend ~$500, revenue $6, ....

Hopefully seniors here could give some advice, i willing to share every single details of campaign. BTW, first thing here is how to create a follow along thread.

 

[iMonetizeIt]

Hey Bryan, did you analyze, what was the reason of your unsuccessful results in the past? Did you get some insights?

 

[bryan]

Hey Bryan, did you analyze, what was the reason of your unsuccessful results in the past? Did you get some insights?

Hi iMonetizeIt, Thanks for your first reply in my post. I think my biggest problem maybe is how to analyzing the data gathered and optimize it accordingly, always i am in dilemma situation if i need to kill the target or keep run for more data. End up i am spending much of not performing target and killed potential target. I have the same problem here at my 2 mobile campaigns, i don't know if i need to stop that carrier already after spending 200X without conversion, or because my banner no good enough causing no conversion. Or maybe placement are not relevant. so and so.

 

[iMonetizeIt]

Bryan, all concerns you've mentioned can take place. To make sure which one is correct, you only need to test different variants. This is actually the only way to find the right reason. It won't work, if you have equally setup campaigns, try different combinations of targeting, creatives and other settings, it may help to find the right way. I wish you to get your money back soon and grow in profit!

 

[bryan]

Bryan, all concerns you've mentioned can take place. To make sure which one is correct, you only need to test different variants. This is actually the only way to find the right reason. It won't work, if you have equally setup campaigns, try different combinations of targeting, creatives and other settings, it may help to find the right way. I wish you to get your money back soon and grow in profit!

iMonitizeIt, what do you mean by different combinations of targeting, creative and settings? I am running display ads (banner) and it is RON, unless i need to blacklist the variants (carrier/OS/target) at the start. I learn that it always better to run broad at start, then gather data and blacklist those non-performing one (which i am bad at deciding this).....

 

[comegetbravo]

Split test offers from different affiliate networks.

When a placement has spent the offer payout without a conversion, blacklist it.

Focus on the placements in the beginning, OS version, Model, Carrier can come later. One thing you should split test though is OS (android - ios), tablet vs mobile, wifi vs 3/4g.

If you're running display traffic (banners) you should split test A LOT of banners, simple things as changing the font, color, image can show massive difference.

 

[T J Tutor]

Hi All, great to join affiliateFix, a place with a lot of generous people and asking for no subscription fee, this should give many thumb up!

My CPA journey started one year ago focused on desktop PPV, but as overall i making a loss like -50% ROI. Anyway i learned some coding for landing page, this is the plus side. Had subscribed paid forum, spy tools and whatever i consider should have but still have no fate.

I am ventured into mobile CPA around 3 months ago after reading Charles blog. I am still not giving up on CPA, haha. After reading around and getting admin stuff done, I had created 2 campaigns with one CPI offer (payout $0.30, i found it is very low compared to PPV offer previously) around 2 weeks+ ago. End up crying result -- spend ~$500, revenue $6, ....

Hopefully seniors here could give some advice, i willing to share every single details of campaign. BTW, first thing here is how to create a follow along thread.

To create a follow along, go to the Follow Alongs and Online Journeys forum and start a new thread. Start by listing the details of the campaigns you plan to run and ask for the answers you seek to get them going. As you progress, ask for help in the follow along to solve issues you encounter. Be sure to read through other sub forums in the community to see if someone has already answered similar questions by others. Our business is a data driven business and therefore research and studying collected data from your tracker will be paramount.

We have loads of great contributors posting actionable information every day.

I'm sure you will find we are a friendly and engaging group of marketers and easy to forge relationships with.

Looking forward to your posts and threads.

Check out the DOJO as it is full of great tools and contributions for successful marketers. Click the banner below to learn more.



See you around the forums.

T J

 

[bryan]

Split test offers from different affiliate networks.

When a placement has spent the offer payout without a conversion, blacklist it.

Focus on the placements in the beginning, OS version, Model, Carrier can come later. One thing you should split test though is OS (android - ios), tablet vs mobile, wifi vs 3/4g.

If you're running display traffic (banners) you should split test A LOT of banners, simple things as changing the font, color, image can show massive difference.

Hi comegetbravo, thanks for your kindly advice. Please allow me to ask for below:

• For low payout like $0.30 (CPI) also use 1x payout to blacklist the placement?
• So your hierarchy is, placement --> OS version --> model --> carrier?
• For placement, let say i have 1 conversion before spend the payout, what are you recommend for next? Every 1x payout should have 1 conversion or else kill it? I have many placements in this condition, got 1 conversion before 3x payout, then i keep running this placement, but no conversion fire for the next 3x payout. Maybe until 10x payout another conversion. i don't know if i need to continue this placement and blacklist its next level variants (like carrier / OS version / model) or just blacklist this placement straightaway.
• I killed my placement if it's CTR < average CTR, is this correct?
• I keep my banner if it is > 1% CTR, is this threshold appropriate?

Many thanks in advance for your generous and kindness help.

 

[bryan]

OK, Thanks TJ.

 

[comegetbravo]

Hi comegetbravo, thanks for your kindly advice. Please allow me to ask for below:

• For low payout like $0.30 (CPI) also use 1x payout to blacklist the placement?
• So your hierarchy is, placement --> OS version --> model --> carrier?
• For placement, let say i have 1 conversion before spend the payout, what are you recommend for next? Every 1x payout should have 1 conversion or else kill it? I have many placements in this condition, got 1 conversion before 3x payout, then i keep running this placement, but no conversion fire for the next 3x payout. Maybe until 10x payout another conversion. i don't know if i need to continue this placement and blacklist its next level variants (like carrier / OS version / model) or just blacklist this placement straightaway.
• I killed my placement if it's CTR < average CTR, is this correct?
• I keep my banner if it is > 1% CTR, is this threshold appropriate?

Many thanks in advance for your generous and kindness help.

Well, this is how I do it. It depends on campaign to campaign.
I usually spend the offer payout on a placement, if it doesn't convert then I'll block it (bear in mind that different placements perform differently based on which hour and which day, so best would be to run it for an entire week and see if it performs on any day at all, but that depends on your budet).

My hierarchy is testing angles, landing pages, banners, offers. When I've tested 3-4 different angles, I'll test different ad-formats on that angle (color, font, size, etc).

When I've got some decent results (conversions) I'll optimize after what is converting, first I usually see if WIFI traffic is converting on the placements that're converting, if they're not, I'll pause them. (if it's not profitable).

The last thing I do is optimize after OS version (android 6.1.1) and generally the model. Since a lot of traffic sources don't have options for blocking certain OS versions and models etc I'll keep optimizing those traffic sources on other criteria.

After you've gotten som good data, you shouldn't be looking at CTR, instead look at the CR, you want what's making you the most money, not most clicks.

On banners you should be looking for a 1% CTR, that's fine.

If a placement has converted, you should still let it run for a while, to make sure it wasn't just a random event.. That can be one of those placements that're performing better on weekends rather than weekdays.

 

[bryan]

Well, this is how I do it. It depends on campaign to campaign.
I usually spend the offer payout on a placement, if it doesn't convert then I'll block it (bear in mind that different placements perform differently based on which hour and which day, so best would be to run it for an entire week and see if it performs on any day at all, but that depends on your budet).

My hierarchy is testing angles, landing pages, banners, offers. When I've tested 3-4 different angles, I'll test different ad-formats on that angle (color, font, size, etc).

When I've got some decent results (conversions) I'll optimize after what is converting, first I usually see if WIFI traffic is converting on the placements that're converting, if they're not, I'll pause them. (if it's not profitable).

The last thing I do is optimize after OS version (android 6.1.1) and generally the model. Since a lot of traffic sources don't have options for blocking certain OS versions and models etc I'll keep optimizing those traffic sources on other criteria.

After you've gotten som good data, you shouldn't be looking at CTR, instead look at the CR, you want what's making you the most money, not most clicks.

On banners you should be looking for a 1% CTR, that's fine.

If a placement has converted, you should still let it run for a while, to make sure it wasn't just a random event.. That can be one of those placements that're performing better on weekends rather than weekdays.

Hi comegetbravo, this is great help.

Sorry, i am still in doubt. Mean that you only start blacklist your placement after you tested angles, landing pages, banners, offers? or you do it concurrently? Mean you will block placement once it hit 1x payout, even when you are on your first day testing 1st angle/LP?

How you decide your angle/LP/banner/offer is ok? Only by CR (what is the benchmark)?

BTW, you mentioned "when you've decent result", do you have any threshold for this, like how many conversion / CR / ROI?

I am sorry for multiple questions. Your generous are much appreciated.

 

[comegetbravo]

Hi comegetbravo, this is great help.

Sorry, i am still in doubt. Mean that you only start blacklist your placement after you tested angles, landing pages, banners, offers? or you do it concurrently? Mean you will block placement once it hit 1x payout, even when you are on your first day testing 1st angle/LP?

How you decide your angle/LP/banner/offer is ok? Only by CR (what is the benchmark)?

BTW, you mentioned "when you've decent result", do you have any threshold for this, like how many conversion / CR / ROI?

I am sorry for multiple questions. Your generous are much appreciated.

No, I first have to get some data for which landing pages/banners/angles are converting, before banning.

Read up on statistical significance as it varies from campaign to campaign. But a mark should be that you can predict what's going to happen in a day, how many conversions (roughly). Statistical significance - Wikipedia, the free encyclopedia

 

[azgold]

@bryan

You're already getting some great advice, so I'll just say hello and welcome!

 

[bryan]

No, I first have to get some data for which landing pages/banners/angles are converting, before banning.

Read up on statistical significance as it varies from campaign to campaign. But a mark should be that you can predict what's going to happen in a day, how many conversions (roughly).

Thanks comegetbravo, this is the thing I done differently so far, I always blacklist target first as for small budget. Your advice sound better but need burn a lot of budget!
Thanks again u r really helpful...

 

[bryan]

@bryan

You're already getting some great advice, so I'll just say hello and welcome!

Hi azgold, thanks for ur greeting.

 

[iMonetizeIt]

iMonitizeIt, what do you mean by different combinations of targeting, creative and settings? I am running display ads (banner) and it is RON, unless i need to blacklist the variants (carrier/OS/target) at the start. I learn that it always better to run broad at start, then gather data and blacklist those non-performing one (which i am bad at deciding this).....

Yep, just make the traffic range narrow relying on reports. It's also important to have appealing creatives, the banners must be attractive for viewers, otherwise it won't work. I know many cases when people seriously improved campaign performance just changing the images. There are different 'trends' in adult niche for each GEO.

 

[tyoussef]

U don't have to spend that kind of money Only for test
What platform are u using To advertise ?

 

[iMonetizeIt]

tests may not take much money, it's only testing. We always do split testing between different offers on the platform, and it really helps to choose the right offer for maximizing revenue.

And spending money like that almost always bring better return than loosing money on unprofitable single source.

 

[bryan]

U don't have to spend that kind of money Only for test
What platform are u using To advertise ?

Hi tyoussef, i am advertising on Go2Mobi - display banner.
Normally how much you use to test an offer, and what is the benchmark you use to decide whether to move forward with it or kill it?

 

[bryan]

tests may not take much money, it's only testing. We always do split testing between different offers on the platform, and it really helps to choose the right offer for maximizing revenue.

And spending money like that almost always bring better return than loosing money on unprofitable single source.

Hi iMonetizeIt, i asked the same question to @tyoussef , how much spending you use to test an offer and what is the criteria for you to decide continue or pause it.

Anyway i had post a follow along on my first campaign, with all the details. It would be appreciate if you could give some kind advice over there. Thanks in advance.