Home > Mobile > Follow-along Campaigns

A Newbie is learning how to optimize a mobile campaign (28)

07-12-2014 04:08 PM #1 bipedjd711 (Member)

Great job Harrison jumping in there and getting your feet wet right away! This is new to me as well. I just joined the forum myself! Best of luck to you!!

07-12-2014 10:10 PM #2 onleads (Member)

Quote Originally Posted by bipedjd711 View Post
Great job Harrison jumping in there and getting your feet wet right away! This is new to me as well. I just joined the forum myself! Best of luck to you!!
Hey, mate, thanks for the encouragement. I new here too. Let's learn together.

07-12-2014 10:20 PM #3 onleads (Member)

Second Day Results

Since there are no reply for my questions on the first thread. I decided to take actions like this:

1. I stopped 3 text ads that cost me more than my payout and generated 0 conversion.
2. I added in 3 more text ads to test.

Second day Results:
Spent: $10
Conversion: 0
Total P/L: -$20

My questions:

Should I keep adding text ads?
I spent $2.5 on a few phone models without any conversions. Should I stop them?

07-13-2014 06:59 AM #4 dennisp (Member)

Hello onleads. Probably you can expect to get feedback from the experts on Monday. In the meantime i would suggest to make a post with all the data of what is going on. (Ctr, cpc, cpm, cr,...). Then you'll get the best feedback.

I would suggest to focus on ctr first, to know what people click on. So don't remove ads with high ctr and no conversions just yet. Maybe the problem lays somewhere else. Are you direct linking or using a lander?

07-13-2014 07:05 AM #5 Mr Green (Administrator)

Dude, if your AM recommends you using a lander then you probably should!

07-13-2014 11:03 AM #6 onleads (Member)

Quote Originally Posted by dennisp View Post
Hello onleads. Probably you can expect to get feedback from the experts on Monday. In the meantime i would suggest to make a post with all the data of what is going on. (Ctr, cpc, cpm, cr,...). Then you'll get the best feedback.

I would suggest to focus on ctr first, to know what people click on. So don't remove ads with high ctr and no conversions just yet. Maybe the problem lays somewhere else. Are you direct linking or using a lander?
Thanks for the advice. I am direct linking

07-13-2014 11:07 AM #7 onleads (Member)

Quote Originally Posted by Mr Green View Post
Dude, if your AM recommends you using a lander then you probably should!
Hi Mr Green, finally I have some experts give me some advice.

I created a lander. However, I don't know whether the lander is converting or not. So I will not be able to test the devices, and OS as all the variables are changing.

So should I just focus and create the ads and a bunch of landers to find out which combination is working at this moment? I only test the devices and OS once I found the working combination?

07-13-2014 11:39 AM #8 dennisp (Member)

Hi Onleads,

What helped me a lots to create a positive ROI campaign is to setup tracking (I am no expert by any means). I am using Voluum, in the beginning it will be hard to setup but after a while you'll be creating new campaigns in minutes.

Voluum (or any other tracking service) will enable you to tell you exactely what is converting and how your landers are performing.

A 2nd advantage, if you setup a tracking system you'll get much better advice here because you can show some data.

07-13-2014 11:54 AM #9 onleads (Member)

Quote Originally Posted by dennisp View Post
Hi Onleads,

What helped me a lots to create a positive ROI campaign is to setup tracking (I am no expert by any means). I am using Voluum, in the beginning it will be hard to setup but after a while you'll be creating new campaigns in minutes.

Voluum (or any other tracking service) will enable you to tell you exactely what is converting and how your landers are performing.

A 2nd advantage, if you setup a tracking system you'll get much better advice here because you can show some data.
Hi Dennis,

I am using voluum too. I have the data but I don't know from which way I should analyze it. Maybe my mind is too complicated, I just feel that adding a lander, I will not know the converting combination of devices and OS. But without a lander, I will not be able to see the conversion. I really don't know which way I should start as there are so many variables are moving in my campaigns and I don't know how to touch them.

Harrison

07-13-2014 12:23 PM #10 dennisp (Member)

If you setup it correctly you can get detailed reports per landing page. Here is the on you where talking about for me:



This is a report for 1 specific landing page, I used the 2 drop downs above to make the exact report you asked. The clicks are the click throughs on the lander.

(I know I have a problem with tablets, it is an issue with the traffic source I am solving).

What helped me a lots is the following setup per campaign:

- Looots of banners
- 2 landers

Focus on CTR's of the banners first, meanwhile get the data on the landers and devices. (with a low budget it goes much faster to optimize your ctr than to get actionable data on landers and devices). After few days you will start to be able to make correct decisions.

And again, show some data in this thread!

07-13-2014 01:07 PM #11 onleads (Member)

Hi Dennis,

Thanks for the direction. Sorry, I wish to ask a few questions:

1. You said we should focus on CTR of the banner, what CTR you expect to get?
2. Everyday, besides adding banners, do you modify your lander based on the data of previous day? or I should focus and keep adding banners, without changing anything on the landers until I know which ads are converting?
3. Do you target all the phone devices at this stage?

Honestly, I wish I have the results that you have now. I have 632 clicks but conversion is 0.

07-13-2014 02:55 PM #12 dennisp (Member)

No problem asking questions, I'll try to answer them:


1. I really don't know, I haven't tried a similar offer. Probably some of the more experienced guys can answer this soon.
2. I don't modify my lander so much, the reason is that it takes long to get the data. I try to be as statistically correct as possible. There are some good calculators out there to see if the difference is statistically significant or not. I do try 2 completely landers to see faster results.
3. I make distinction between Tablet & phone for targeting on the traffic network. The data out of Voluum will tell me the rest. I can believe that it is smart to differentiate between ios & android also, but I am not doing this myself at this point.

Can you show us some Voluum and banner data?

grts!

07-13-2014 04:47 PM #13 Mr Green (Administrator)

Quote Originally Posted by onleads View Post
Hi Mr Green, finally I have some experts give me some advice.

I created a lander. However, I don't know whether the lander is converting or not. So I will not be able to test the devices, and OS as all the variables are changing.

So should I just focus and create the ads and a bunch of landers to find out which combination is working at this moment? I only test the devices and OS once I found the working combination?
PMed ya!

07-13-2014 10:42 PM #14 onleads (Member)

Quote Originally Posted by dennisp View Post
No problem asking questions, I'll try to answer them:


1. I really don't know, I haven't tried a similar offer. Probably some of the more experienced guys can answer this soon.
2. I don't modify my lander so much, the reason is that it takes long to get the data. I try to be as statistically correct as possible. There are some good calculators out there to see if the difference is statistically significant or not. I do try 2 completely landers to see faster results.
3. I make distinction between Tablet & phone for targeting on the traffic network. The data out of Voluum will tell me the rest. I can believe that it is smart to differentiate between ios & android also, but I am not doing this myself at this point.

Can you show us some Voluum and banner data?

grts!
Hi Dennis, I would like to share the data too. But I am too confused and I try and modify a lot of things until all the data is a mess. Now I am restarting process again and hopefully will show you some data tomorrow. Anyway, thanks for the info. Will follow your advice.

Harrison

07-13-2014 10:43 PM #15 onleads (Member)

Quote Originally Posted by Mr Green View Post
PMed ya!
Hi Mr. Green, PM replied

07-15-2014 06:55 AM #16 onleads (Member)

Ok guys, finally I decided my plan to test this Samsung offer and now it is time for me to start collecting data again.

Campaign is recreated again according to my plan. Here are the settings for the campaign:

Targeted Country: MY
Targeted Carrier: Digi
Targeted Connections: Cellular
Targeted Platform: All Android Versions
Targeted Phone: All Phone except tablet
Targeted Category: All
CPC Bids: $0.05 (minimum Bids)
Daily Budget: $15
Landing page: LP from AM
Ads: 5 Text ads

I will update the data here again tomorrow and get advice from you guys. Thanks

07-15-2014 12:03 PM #17 caurmen (Administrator)

Looking forward to seeing your results!

07-18-2014 08:17 AM #18 dennisp (Member)

- I would say that 17.8% ctr a your lander is a bit low. Try to optimize more, try something completely different.
- I think it is strange that you only get Samsung traffic, maybe it is a good idea to target budget phones (huawai and nokia for example) and exclude Samsung.


Your ads have not enough impressions yet to make a statistically confident conclusion.

07-18-2014 10:02 AM #19 caurmen (Administrator)

Exactly so. Let the ads get to 2x the payout at least before you touch them.

Are you culling placements (app or site)? You should definitely be looking out for any placement that's eating your spend without producing good results.

07-18-2014 11:08 AM #20 onleads (Member)

Quote Originally Posted by dennisp View Post
- I would say that 17.8% ctr a your lander is a bit low. Try to optimize more, try something completely different.
- I think it is strange that you only get Samsung traffic, maybe it is a good idea to target budget phones (huawai and nokia for example) and exclude Samsung.


Your ads have not enough impressions yet to make a statistically confident conclusion.
Hi Dennis, thanks for your advice. I will try my best to get a higher CTR on the lander. Currently I am targeting all phone models but I don't know why only samsung is getting the traffic. Ok I will let me ads run and get more data.

07-18-2014 11:10 AM #21 onleads (Member)

Quote Originally Posted by caurmen View Post
Exactly so. Let the ads get to 2x the payout at least before you touch them.

Are you culling placements (app or site)? You should definitely be looking out for any placement that's eating your spend without producing good results.
Ok, change my target to 2x of payout for all the variables.

I am using leadbolt and I don't think there are any options for me to select between apps and site, do they?

07-18-2014 11:34 AM #22 caurmen (Administrator)

Ah, right. No, I believe most Leadbolt advertisers don't have that option.

07-22-2014 02:21 PM #23 onleads (Member)

Hi everybody,

I am back again with the new data. This time I found a way to track the publisher/site in leadbolt and hopefully this will help me to optimize this campaign.

Without further ado, here is the result:

My Ads Status

http://snag.gy/yIiO2.jpg

My Voluum LP Report

http://snag.gy/50quK.jpg

I did some minor changes on my LP, however the CTR is still the same. Not much different.
Questions: So should I test the landing page with a totally different design? By the way, the current landing page I am using now is from my AM.

My Voluum Device Report

http://snag.gy/yzNwG.jpg

My Voluum Publisher Report

http://snag.gy/CYy9D.jpg

Results:
Spent: $30
conversion: 0

I hope to get some feedback from you guys here. Thanks

Harrison

07-22-2014 04:51 PM #24 caurmen (Administrator)

So should I test the landing page with a totally different design?
Definitely, DEFINITELY do this. Do really big, dramatic tests first.

07-23-2014 08:37 AM #25 onleads (Member)

Quote Originally Posted by caurmen View Post
Definitely, DEFINITELY do this. Do really big, dramatic tests first.
Should I take note that landing page is the first variable that I should test in the optimization process?

07-23-2014 10:36 AM #26 caurmen (Administrator)

I'd test it in parallel with your ads, placements, etc. There are some arguments for being really rigorous and only testing one variable at once, but generally you're OK to test the lander and the ads at the same time provided you're sticking to roughly the same angle.

07-23-2014 02:04 PM #27 onleads (Member)

Hi Caurmen, I went and read your post in http://stmforum.com/forum/showthread...TION-AND-INDEX and start learning a lot this time. maybe I have some experience and Data to understand what you are talking about inside the guide.

So can I do conclude something like this:

First, we must set a goal on the ROI that we wish to get from a campaign. By doing so, we will have the Minimum Viable Conversion Rate. By having this conversion rate, I can use it to determine which ads to use. Besides that I also use this conversion rate and my ROI goal to calculate the CTR of my lander that I must working to.

Is my mindset on the right track by thinking in this way?

07-23-2014 07:02 PM #28 caurmen (Administrator)

Read http://stmforum.com/forum/showthread...y-Don-t-Matter to get some clarity on the CTR question - it's a great piece.

Minimum Viable Conversion Rate: yes, but be aware in mobile there are a lot of interrelating factors, and you can't just isolate one (ads) on its own. So using a Min. Viable Conv. Rate is good, but be sure to account for possible optimisations elsewhere too.

Home > Mobile > Follow-along Campaigns