Home > Paid Traffic Sources > Mobile

---

The Green Mile of Mobile (47)

---

If the famous movie headline captured your attention, I welcome you to my mobile follow along

Long story short: got into AM, found Ngo`s blog, joined STM and launched a mobile campaign, mainly following The Main Course of the Cookbook.

I will focus on one traffic source (Decisive) and one vertical (utilities). I was going to try out services (dating, chats etc.) as well, but decided to stick only with utilities in order to avoid focus spread.

No “$xx,xxx/day in 6 months” goals set – my only objective is to learn and master the process of turning an unprofitable campaign into a profitable one. Speaking metaphorically, this is my “green mile” – the walk from red ROI's to green ones

Let's get down to business:

Offer: DU Battery Saver
Payout: $0,22
GEO: Ukraine
Traffic source: Decisive

Angles: 1) Stating the benefit. 2) Statistics

Banner examples:





Landers:



Headline examples (since stating benefits is rather obvious, I`ll include headlines from my 2nd angle):

“Only 7% of people know about this. How to double your battery life in 3 minutes”
“This App Triples Battery Life for 8 out of 10 phones”
“71% percent of Androids die in 3 years! Save yours now!”

CTAs:

“Click here”, “Improve Battery” and Google Play button.

Lander text:

Basic benefit statements and instructions on how install and open an app.

In total I have 8 campaigns (Wi-Fi/Mobile – App/Site for each angle) with 27 banners each (3 CTA x 3 Headlines x 3 Designs/Angle).

Introduction ends here. Updates to come

Thank you for attention!

---

Update 1

Info on campaign so far:

Spent: $13.07
Made: $7.92
ROI: -39.4%

Optimization:

1. Were not getting traffic/conversions from sites, so stopped them.

2. Cut a few placements.

Further actions:

1. Gather data.

2. Analyze banners to map the most effective combinations and create new banners based on those.

3. Analyze landing pages and define the winning combination.

4. Create and test a new angle.

5. Create 3 unique landing pages and test them.

I will cover my process of analyzing banners when they will be live in campaign and ideas for landers/new angle in the next update, when I will have some data.

I would really appreciate advice and suggestions! Thank you for your attention

---

Pretty good ROI to start off with. Keep optimizing!

---

Agree 100%! Very sweet ROI for something that isn't even optimized yet! And I love your approach - looking good!

The only thing I could suggest is that you may want to split-test your lander with direct-linking (from banner to playstore). Sometimes for free mobile apps, direct-linking works better.

Looking forward to your updates!


Amy

---

Suprised Decisive let the Android image ride. They denied all of my banners with that on them and just out of curiosity how many times do your gifs loop? Great job man.

---

That's a promising first round of data -- especially if you're starting on Decisive.

Good choice of GEO, simple landing pages, proven ads...

Nice start!

---

Update 2

Update coming a bit late, but I guess I`ll do 2 today.

May 21st:

Spent: $5,8
Earned: $4,62
ROI: -20.34%

Optimization

Banners

After gathering enough data (18-20K views/banner) I worked on cutting and making new creatives, analyzing their effectiveness by 2 metrics – CTR and CR. At first, I analyzed the banners with the highest CTR and figured out patterns between them – top 5 banners all had the same CTA, so it was clear enough. After that, I studied the ones with the highest CR and using the method Amy have posted (Thank you! ) defined clear winners.

In the end, I had mapped out the elements which produced the highest CTR and 2 angles (more of a sub-angles, actually) with the highest CR. Also, as a rule of thumb – if the banner did not bring any conversions, it gets cut (after CTR analysis).

– For the angle with the highest CR, I created a set of banners based on the data from the creatives with highest CTR:



– For the second best angle, I made a set of banners with very different style to see if it can be made effective with a different approach:



In about 8 hours I will be able to see what worked and what didnt.

Landers

I made a few mistakes with landers.

1. In the beginning, I created 2 landers with different designs and style of sales message delivery, but rather same angle. So instead of testing 2 angles, I tested 1.

2. I did not write down the key elements of the lander, so mapping out the combination for a new one is guessing game.

3. Did not optimize all the pictures used in landers. Obviously, 33kb to 3kb is a huge size reduction. Even though my VPS is close to the country I run campaigns in, so the loading speed is fast, maximum possible optimization is a non-breakable rule.

I created 1 lander, which is a mix of the ones I had from the start (they both convert at about the same rate), and 2 unique ones. Will report on them after testing.

Placements

This is rather simple – if it spent about 3x the payout and brings no conversions, it gets cut. I blacklisted a few huge non-converting money eaters.

Key findings from May 21st:

1. The drive to avoid pain/get pleasure sells. Being the key motivators of our behavior, they are the highest converting sub-angles for both campaigns I am testing. Lesson for me – make sure that headlines and LP content are not only creative, but are compliant with marketing/psychological principles. Ca$hvertising is not called the Bible of AM for nothing, right?

2. Know exactly what you are doing. … and dont create LPs just to create LPs. I have to figure out exactly what I am trying to achieve with landers, how they are different from one another, how their elements can influence traffic behavior and etc. As an experienced AM told me: “Let the data guide you”. So I have make sure that I am actually getting clear information, and not playing a guessing game.

3. Analysis is the key. The book “Problem Solving 101”, mind-map tool XMind, split-testing websites and other spreadsheets for calculation were my best buds when it came to analysis.

That is it for my second update – next will be posted after getting data on new banners/landers.

Thank you for attention!

---

@ssmarketers -- Working on it, thanks .


@vortex -- Yep, did that. LPs outperformed DL 4:1. But I think the inconsistency between banner and content on the store played a big role on DL performance: all my banners are basically based on eliciting curiosity and emotional response, while the page of an app on Play Store is all benefits oriented. There is a connection, but not really relevant. I actually want to create a set of banners which are consistent with the store content and see how DL will do, so I know for sure that my theory is right. Thank you for suggestion!


@yussef -- Thanks, mate . 3 loops. At first I submitted gifs with infinity animation, but, obviously, all got rejected. And btw, all the banners with android did not make from the CTR/CR point of view.


@Finch -- Thank you!

---

Following since your in about the same stage as me :-) Keep up the good work!

Cheers

---

Awesome start, i got a feeling you'll be seeing green in no time!

---

Update 3

This one is going to be short -- AM asked to stop the campaigns, because the advertiser requested to use cretives from the Play Store only. Obviously, a big slap on the wrist .

The other network has very strict guidelines on creatives for DU Booster, which cross out all of my banners.

Summing everything up, I have to create new angles.

Today I also have gathered enough data to analyze the performance of new banners, and they did pretty well -- ~0.8% CTR increase. The new LP seemed to be performing better too, but not enough data to tell for sure.


That`s what I have for today, thank you for attention.

---

Thanks! Do you have any success with your campaigns yet?

Thank you .

---

Wow! Way to make the best of a bleak situation!

The only important factor that separates winners from losers is mindset - knowledge can be learned, and experience can be gained - but mindset you can't change so readily even with time. You have what it takes to succeed. I can usually tell about people.

Once you're confident you know how to properly launch and tweak a campaign, feel free to launch multiple offers at a time. If budget's a problem you can always bid low at first and set a low daily budget for each camp. When you have eggs in multiple baskets, you're not holding your breath for any one offer to be a winner, which takes a lot of the pressure off the mind (and, if you believe in the law of attraction, it eliminate lot of resistant energy that results from being afraid that your camp will fail). I know I've referenced stackman's thread many times before but it's one of my all-time favorite threads on STM:

http://stmforum.com/forum/showthread...obile-Campaign

With all his experience, he launched 30 camps to find 3 that had potential.

While you're still learning though, it would be wise to stick with one offer and use it to learn all your basics. Once you have the basics down pat, you could then cast a wide net, then focus on the promising areas/offers.

So yeah - no need to feel bad about the "slap on the wrist" from a single offer. Not that you're letting that discourage you anyways.

Looking forward to future updates...

Amy

---

Update 4

Today I created campaigns for a new offer. As I said in my previous post, there were some issues with creatives, so I wanted to run this campaign for a different network -- turned out it requires approval and my AM doesn`t work on weekends. I got approved in another network though, but DU Saver is not on their offer list.

I decided not to run the campaign just for the sake of running it and work on a new offer until I will be allowed to use my creatives or get approved by network which has DU Saver.

New campaign

Offer: LEO Privacy Guard
Geo: Russia (only)
Traffic Source: Decisive

This app has a feature which helps to save phone battery, what means that I can use most effective angle/creatives from the DU Saver campaign. Also I wasn`t ignorant this time and asked my AM which offers are converting well right now and are run on high volumes -- LEO was the recommendation.

Campaign details

Angle

Similar to the one I used before. I analyzed features of the app, mapped them with the biggest concerns of Android users which are listed in this case study from bbrock32 (invaluable info!) and came up with an angle, which is focused on the motivation to avoid pain. I surely want to try out some very different angles, but for the beginning I decided to stick with what is proven to work.

Banners

Again, similar to the ones I used before. My main objective here is to find out which sub-angle/headlines are the most effective -- I`ll focus on unique designs later. I have chosen the ones which worked best for previous campaign.

In total -- 10 banners/campaign.

I also used a different approach to creating banners: 5 headlines (same style, but delivering different value), 2 designs and 2 CTAS. This way I am testing the most important variable (headline) and getting the most accurate data.

Lander

No screwing up this time -- similar lander, similar style of headline, similar sales message approach and similar CTA for both angles. Basically, I use one lander with different content, which is tuned to an angle.

Summary and main objectives

1. Testing, testing, testing. My objective is to get as much precise data from this campaign as possible. Also I have a big list of angles/landers/banners I want to test.

2. Working on landers. This is my weakest point, that is why I am focusing on it. I have alot of ideas for landers, but coding skills are limiting me. With time and practice it will be solved, so I`m just working on improving them bit by bit.

3. 100 headlines/day. This is a technique of improving headline writing I learned on Jon Morrow`s blog. He said that he mastered headlines by writing 200 of them per day. But I`ll start with a 100 and go for more if time will let me.

As far as I can tell, most mobile landers have short copies, so getting good with headlines will partly improve copywriting skills as well. Of course, long copy sells too, but I dont want to spread my focus.

4. Read less. I figured out that I slightly overread, so as a rule of thumb: I only will read posts during work time which are related to the issues I am facing. Find a solution to a problem, implement it, close the damn tab . I have already limited my reading time to 30 minutes/day, but this new fix will save more time and boost productivity.

That is what I have for Saturday. Stay tuned for info on campaigns .

Thank you for attention!

---

Thank you so much for advice, Amy! Hopefully, soon I will be able to share something of a value with you too.

The situation with creatives/offer pausing is just a part of this business, I believe. So the only option is to make it work one way or another.

And yes, stackman`s post is fantastic! Adjusting famous Glengarry Glen Ross speech to affiliate marketing: "ABT: Always. Be. Testing". Also I implemented similar method of banner creation stackman used (creating generic banners which work fine for all offers), so I believe I`m doing something right.

About learning the basics -- I write down all the steps I take in order to create a well organized, rational structure for testing, optimization, analysis and etc. Practice makes it so much easier too -- today I`ve noticed a small option in Photoshop, which allowed to make banners 3-4 times faster .

Looking forward to your posts!

---

Sounds like a plan! I'm going to ask the same question again though: Are you planning on including direct-linking in your testing? If creating landers is your weak point, it may be easier to just concentrate on optimizing banners at first. With leo, I was able to take it to profitability without using a lander (and so did a lot of other people I spoke to). It was in another geo though.

And I think you chose a great geo from the looks of ranking trends in appannie:



Started trending upwards relatively recently and shows no signs of coming down.

This thread is just getting funner and funner!


Amy

---

Yes, all campaigns are split DL/lander. I too know that alot of affiliates run this profitably with DL in Russia, but from another traffic source.

While waiting for approval of my campaigns on Decisive, I`ll finally finish the lander I`ve been working on post it here for some constructive criticism . I want to have a few generic template ready, so it will be easy to adjust them to any offer and try different approaches.

I seem to understand banner otpimization and do it rather quickly, but would be thankful for your opinion on the process (will cover it in my next updates), so I can be sure I am not mistaken.

---

Great follow along, you are doing everything "by the book". You should see success soon.

Subscribed

---

Update 4

No data today, because creatives were not approved.

I used free time to work on landers -- finally finished the one I had in mind for a week, created 2 testing LP concepts, and coded a few pages for a 2 page lander. Here is one of those:



Also I am working on a few other LPs which were kindly shared in our mastermind group by a few affiliates.

Talking about mastermind, we had a very insightful group call today -- analyzed our progress, shared some insights, reported on current situation, figured out goals for the week and etc.

For the rest of the Sunday (21:31 for me) I will write headlines as I mentioned in previous update and dig for some insights in "Tested Advertising Methods."

Have a nice Sunday and thank you for attention.

---

Thank you! Working on understanding the basics and figuring out rational approach for offer/campaign testing/optimization (creating a minmap for that).

---

Looking great! Some brief suggestions regarding landing page:

I've never used landers for free app download offers, but I've seen enough to notice that short copy will ALMOST always work better than longer ones. This is most probably because for a free app download people don't need much convincing. So I would suggest split-testing the format you have, with a shorter one that has just a headline, image, a couple of benefits (like maybe 1-2 or even none) and a CTA. Make sure that's ALL above the fold. Stuff like testimonials and reviews etc. are nice to split-test as well, but I would suggest putting them after the CTA so that they start above the fold and continue below the fold.

Your banner optimization plan as you've outlined in a previous post, looks solid, so I'd suggest just sticking to it and refining as necessary. Also, don't be afraid to make mistakes! That's how we all learn.

Also - are you using Voluum? Something else you can play with, is calling out the user's information on the lander. For example, "Is your Samsung Galaxy Note getting laggy? Get this app and see the difference!" You can call out stuff like geo location (country/city/region), device model, carrier, and a bunch of other info. Caurmen's guide will tell you everything you need to know:

http://stmforum.com/forum/showthread...ing-Parameters

It sounds like you're getting a lot of practice writing headlines! You know where you can find the best ones to model after? Viral sites like ViralNova and UpWorthy! These sites will typically make their writers/editors write lots of headlines for each piece of content they write, before deciding on the final one to use (I think 30+ headlines for UpWorthy). They also have TONS of data on what kinds of headlines work best, which enables them to improve over time. Just this one spying technique has benefited me more than all the ebooks and course materials I've paid for in the past on copywriting. Hope you can get some benefit from that too!

(Those headlines tend to be longer than what we need for our banners and landers - would be best to keep them short and powerful. But we can use the same techniques nonetheless.)


Amy

---

Update 5

Spent: $31.51
Earned: $0.00
ROI: -100.00%

Obviously, testing Leo and UC Browser (another offer I`ve launched a few days ago) didn`t go as well as it was with campaigns for DU. Fortunately, it allowed me to realize that the only reason why I was close to breaking even on DU campaigns is because of very low CPM -- I was buying traffic from one placement, had low CR, low lander > offer CTR, and did not optimize LP to increase conversions. Therefore, I was buying a lot of cheap traffic and converting it with a below average CV.

Right now I have 20 active campaigns for privcacy guard (testing 5 angles) and 8 paused for UC, since I cannot properly analyze gathered data. I know that quite a few affiliates run UC succesfully and I know that Privacy Guard is working just fine for a lot of people right now, so there a few theories on my current unsuccess:

1. Non-converting placements. I am still trying to figure out the nature of how Decisive works -- it seems that after blocking ~10 placements (from which only 2-3 bring the most impressions/clicks and others are just being bid on while not bringing any traffic/impressions), I get another 10 placements which behave the same way as the previous ones.

Also, I am getting very low traffic, even though my bid is about 30-40% higher than the average for GEO, banner size, placements. My budget is unlimited and daily budget is set to $90, I am using flat CPM model, but nothing changes the traffic flow. I surely need to contact Decisive to understand how the system works and operate it properly.

2. Bad angles/creatives. I cannot tell about angles for Privacy Guard campaigns, since I do not have any data on them just yet due to low traffic flow, but UC campaigns doesnt seem to be working. 1 angle is "Speed up your internet by 40% percent" and the other one is "Watch your favourite movies and TV-shows from any website for free without ads". My bet was on the last one, because I know that other affiliates were having success with it, but there a few reasons why it may not work, which I cannot identify due to poissibility of bad data.

Talking about the lack of data, the only way I can solve this issue is by getting reasonable traffic flow - right now traffic is very low on all campaigns.

Quick summary

The biggest mistake I am close to making right now is calling campaigns a fail without significant data, so I will limit myself from any further analysis untill I will fix the issues with Decisive and figure out how system works -- then I will be able to tell if the issue is with traffic quality or creaives/lander/angle. I don`t care to lose money, because I am paying for data and to see my mistakes and master the process.

I still have a lot to learn about analyzying the data properly, because there are so many golden nuggets in the tracker stats that I simply do not know how to identify yet.

Also, I want to run another offer in Ukraine today to test some landers.

Thank you for attention. Updates to come.

---

Thanks, Amy! My landers for Privacy Guard have one sentence, CTA and an icon image, all above the fold. I did not use any social proof this time to keep everything short and keep the relevancy of banner/lander, but I will split-test it with another lander style on later stages for sure.

I`ve actually wrote an article about UpWorthy and how it became viral in a very short term. They surely have excellent headlines and few other techniques to engage users. For example, they use dialogue pop-up with a short message and 2 CTAs, something like: "Do you like cute puppies?" and CTAs "Of course I do" and "No, hate them". When a visitor agrees with a statement by clicking on relevant CTA, another pop-up appears with a subscription form and a message, that emphasizes on a previous statement: "We love animals too and write about them daily! Support the doggies by subscribing!". I believe that this technique is based on the consistency principle. Surely something worth testing.

I forgot about UpWorthy, so big thanks for reminding me about this goldmine .

I do use Voluum, yes, and surely want to test personalization -- bbrock32`s case study showed what impact calling out user's phone can have.

---

Thanks for talking about this! It has just given me an idea actually: Instead of having the typical 'alert' entry popup, I'll try having a "confirm" popup with something similar to the technique you're describing. I could ask a question that most people will agree with. Once a visitor clicks on a "Yes" they'll be psychologically more accepting of the rest of the lander which may increase the likelihood of their signing up to the offer at the end.

Gonna go test out this idea this week...will try to remember to share results here later!

Thanks again sylar!


Amy

---

there are other issues you might want to test out
1. offer is unsuitable for traffic source, switch to other offers, like apus hola etc
2. affiliate network server/scrubbing issues. i can see regular 50-100% differences in revenue, and sometimes it can go up to 2-3x or even 10x.
3. geo is unsuitable for offer, change geo (but ukraine should be fine?)

---

You can also try altercasting -- projecting a role onto your target audience, which is likely to boost their desire to convert. For example: "You are an intelligent woman who cares about her looks, so you surely understand why [paste fancy offer name here] is able to make wrinkles disappear for 9 out of 10 women". I think this tactic can be easily applied to popups too -- a short altercasting sentence and UpWorthy style CTAs might do a good job.

Jon Morrow (link to his blog) uses altercasting alot in first paragraphs of his posts to put audience in the right mood. And here is another nice article which covers altercasting a little bit.

I also noticed that alot of marketing blogs have been using 2 CTAs in pops lately: 1 which enlarges on benefits of an offer ("Yes, send me these free insights") and one which bleaks subscription rejection ("No, I will continue to make silly mistakes and lose money every day").

Hopefully you`ll find something worth testing, Amy .

---

Update 7

Spent: $16,93
Profit: $0.87
ROI: -94.86%

Testing new approaches on Decisive did not show much potential -- got 1 conversion on a targeted campaign (lander, angle and banners are adjusted for a specifid placements). I targeted 3 mid/low volume apps (social media clients for VK, russian social network) and it seems that there is a placement > banner > lander combo with some potential. I suppose some heavy otpimization could make this combo work.

I have also tested 2 angles for LEO on AdModa -- 0 convs (both from DL and landers). I believe banners had OK CTR (~1.2% -- ~1.6%) and there was only 1 obviously bad placement (109% CTR), but lander > offer CTR was extremely low. Banners and landers were compliant -- relevant style and messaging.

Besides what Xavier have suggested, I believe an angle can be an issue as well, since I am using something that was widely used for a long a time.

Btw, I am split-testing between different networks and get the same results, so I believe it`s an issue with an offer/angle/geo/traffic source, not with networks. Anyways, testing will tell.

So here are goals for next few days:

1. Launch Privacy Guard on pop. I have created 3 landers today, but couldn`t start campaign due to some issues with Paypal.

2. Launch on different geos. I will follow Xavier`s advice and try different geos. I am running only in Russia now (networks I take Privacy Guard from have it open only for RU, so I havent tested different geos yet).

3. Try completely new angle for Leo on Decisive. I`ve been told that DSPs are all about finding those gold mine placements, so that is what I`ll try to do with a different angle, which is unlikely to be overmilked.

4. Optimize UC on Decisive. Small campaign with narrow targeting will be my playground for testing -- I want to practice my approaches with banner otpimization, lander optimization, stat analysis and etc. to be sure that I nailed the basics and that I am working in the right direction. As long as I will see some progress, I will be satisfied.

5. Find network with DU Battery Saver. I am really looking forward to bringing this campaign to life again. As soon as I will find a network with this offer, I will start it again on Decisive and launch it on AdModa as well (I believe it should do good there).

6. Find some new offers to test. Always. Be. Testing.

I also made a few tweaks to my testing/optimization approach:

-- Instead of testing 10 banners, I will test 5. Generic banners have proven to bring good CTR so it all comes down to finding the right headline style -- when this is accomplished, I can test different CTA colors, backgrounds, styles, etc. Besides, this will give me an ability to gather more data and get higher statistical significance.

-- I am starting to look more closely at stats in Voluum. There are alot of golden nuggets hidden there which I was missing before.

To sum everything up -- I`m in testing mode. Test, optimize, get new info, make hypothesis -- rinse and repeat.

Updates are to come, and thank you for attention.

---

Think Du pulled the offer from all networks

---

Thanks SO much for this! I don't spend nearly enough time furthering my copywriting skills. This technique is going straight to my mental swipe file!

Way to go with mass testing, but make sure you have a systematic plan in place. Otherwise things could get messy real fast.

Looking forward to your next update...

Amy

---

Update 8

No, this follow along is not abandoned – I`ve been busy testing offers, traffic sources, geos and angles. But from now on I will post daily updates in order keep a better track of my progress. Besides, writing helps to clear the mind.

Quick report on offers:

AliExpress – ran on Decisive, did not work in UA and BY, showed better results in KZ (more convs and higher CTR).

Privacy Guard – showed better results on pops (PopAds performed x3 times better than PopCash). -87% ROI.

UC Browser – narrow targeted campaign did not work on Decisive. There a few other ways to run this offer, and I will consider testing them next week.

Battery Saver – found it on Matomy, started running on Decisive in Ukraine. Tested 2 new angles and an old one – 1 of new angles has better CR and an old one has higher banner and lander CTR. More data is required for a clear picture.

Talking about Battery Saver – in the beginning I had high volume placement (app) which was converting rather okay, but I got it by accident (this app does not allow promotions of utilities) and when it was excluded traffic flow and conversions dropped significantly.

Another high volume placement (which I already blacklisted) had very high banner CTR but low lander CTR and CR. I will create a separate campaign for it and test it after 2 rounds of banner/lp optimization. It may work with lower bid and higher CTR/CR.

I have a lot of testing an optimization to do:

1. Test one more angle.
2. Complete 2 full round of optimizations for current angles after getting enough data.
3. After defining winning angle, test it on different traffic source for UA and different geo on Decisive.

Btw, I dont cut placements in first 2 optimization rounds, unless they are absolutely ineffective (very low banner/lander ctr and no convs). The best way for me to analyze placement is by looking at tracker stats – by analyzing placement > banner > lander combos I am able to form proper optimization theories. Data on traffic sources is just a quick way to see if certain placement have gathered enough data or not.

How I increased banner and lander CTR

One of the new angles which worked OK for battery saver is "test" – user is offered to complete a test (answer 4 simple questions about his phone and the way he is using it) in order to find out how much longer his device can work. After completing a test, user goes to a page with scanner animation (which after a few secs of loading tells him that his phone can work longer for 3 days) and recommendations on how to improve battery life.

After I gathered enough data on this angle, I created another lander (scanner page, no test questions).

Testing showed which banner > lander combos worked best -- basically, the more relevant banner text was to lander content, the higher the CTR was. For example, text “Complete a test to find out how much longer your phone can work” worked best with test lander and text “Find out how much longer your phone can work” worked best with scanner lander.

To increase banner CTR, I created 5 new creative designs with borders, contrast text and etc. and tweaked headlines – changed wording, approaches, added word “short” (Complete a SHORT test...) and numbers (Answer 4 small questions to..) to ease anxiety.

To increase lander CTR, I added a company (named it Android Experts) logo so LP looks more credible, tweaked text a little and added JS timer, so <div> with recommendations on how to improve battery life (install Battery Saver, ofc) is shown after the scan animation is loaded – my theory was that this tweak would make recommendation seem more tailored to scan/test results, therefore more believable.

Results: ~0.2 increase in banner CTR and about x2 increase in lander CTR.

Next I would work on conversions by testing these theories:

1. Information about how much longer a phone can work is not believable, since 3 day increase may sound like a complete bs. Therefore, 2 versions of ladner should be tested – one with testimonials and the other with more believable info (“Your phone can work 1 day longer”).
2. Lander and offer page are not compliant – on offer page it says that phone will work 50% longer, while I use days as quality measure of improvement on ladner. Therefore, changing quality measure from days to percents on LP may increase CTR.
3. Testing urgency, exclusiveness and a few other marketing tactics.



I also want to share a few nice tools I use for optimization:

GS Stat Counter. I use it to check what mobile screen resolutions are most common in the geo I am targeting and optimize landers accordingly.



Optimzilla. This free image compression tool is a beast -- you can manually decrease image size from 22KB to 1.6K without sacrificing quality. We tested Optimzilla vs Kraken.io vs Compressor.io in our mastermind group and Optimzilla was a clear winner.



For testing how LPs look on mobile screens I use chrome JS console. I also want to create a few more lander templates and test them with Browser Stack.

That is it for today. Updates to come and thanks for attention.

---

Holy crap Sylar! (Where's that jaw-dropping emoticon when I need it?? Oh wait....I'm not on skype...)

You're a true marketer! Very good and solid thinking. Once again, I picked up quite a few tips from your observations.

It's so refreshing seeing someone like yourself who's willing to sit down and come up with ideas, and then perform thorough testing to see how well those ideas work. I've answered more than a few PMs/skype messages where people said "I'm running out of ideas on what to test - please help!" when they've only tested a few ripped landers. Although I understand and empathize with their frustration, seeing your post is like a breath of fresh air.

And those tools you recommended - both bookmarked! I always thought kraken was "da bomb" but you've convinced me to switch to Optimizilla!

Thanks...


Amy

---

Update 9

After gathering enough data on Battery Saver campaigns I was able to identify the best placement-banner-lander combo (which, by the way, significantly outperformed test/scanner campaigns). Tracker stats analysis very well explained why a particular banner had the highest CTR and CR – it all came down to how it looked in a specific app an how it`s users most probably interpreted it (or in which state they were in when this creative was displayed).

I created a new set of banners and 4 new landers for this specific placement, but campaigns were finally launched, I had 0 traffic. Turned out that this app was excluded from Decisive placements (particularly for Ukraine), so my theories, creatives and landers cannot be tested.

To solve this issue, I will look for this specific placement in all of the geos which Battery Saver can be run on and then pick 2-3 countries I want to test. Besides the placement targeted angle, I also want to test 2 other angles which worked rather well in Ukraine. To sum everything up – 3 angles in 3 countries.

I am not going to run campaigns tomorrow though, because I want to contact my AM to find out which geos convert well.

According to info from one of affiliate networks I work with the most, conversion rates for AliExpress and UC Browser in geos I ran them in dropped significantly in last few days – average CR is about 0.5-0.7. My yesterdays tests on Decisive confirmed it, so conclusions are obvious.

Tomorrow I want to read 2 books and look for some new offers to run – I already have a few ideas, but I want to make a list of 5-6 offers and get data on the performance from a few different networks.

That is it for today. Thank you for attention, have high conversions and a nice weekend.

P.S. And a quick snack for the ending – IconsDB is a resource with 3669 flat icons, which you can customize by making them whatever color and sieze you need. Useful for banner and LP creation.

---

Thanks, Amy! Glad to be helpful .

Btw, congrats on your recent moderator promotion! You are extemely helpful to alot of stackers, contributing alot and absolutely more than worthy to be a moderator!

---

Update 10

DU Battery Saver was paused on Matomy today. A few major networks have it private, so I will not be able to run for them. Also, this offer may be available (according to Offer Vault) on a few other networks, but I need more info about them before doing business.

Today I will gather info from my AMS on the top performing offers and decide what I will test. I will either run offers I have tested before on different traffic source or choose new ones to run on DSP.

Updates to come and thank you for attention.

---

i think there are some other offers that have something to do with battery but not exactly a battery saver. something like a pin/mobile content thing u can try?

somehow the battery image seems to make alot of difference when i tested it, but that was 6 months ago.

i heard many camps in ukraine for battery... so im guessing that caused decisive to tighten up their compliance.

i like the detail in your post

---

So many follow along posts threads and yet people never understand that it's impossible to run RON campaign and actually learn something from it with such low daily caps.

---

Really great follow along Sylar. Anticipating the day you're in the green.

Hey sahbak, I noticed you mentioning RON campaigns. I'm relatively new to the forum but just wanted to understand what that means?

After doing a quick search on Wikipedia I came up with this:

"Run of Network (RON) advertising is a form of internet marketing where an online advertising campaign is applied to a wide collection of websites without the ability to choose specific sites..."

Does this mean the traffic source is not targeting right people?

Also I'm guessing that a daily cap is the amount of conversions allowed per day. How would it be increased?

Apologies if I sound really nooby but I looked forward to your response.

---

Update 11

Worked on angles, banners and landers for new offers – Omni Swipe (old Lazy Swipe) and 360 Security Antivirus. Creatives and landers have to approved by AM/adv, so I suppose it may take longer than usual to launch them.

I am going to test 1 angle which is commonly used for this kind of apps and a unique one I havent seen before. My goal was to come with unique approach which would be focused on benefits of user/solving his problems.

By the way, on weekend I have finished a book “The Adweek Copywriting Handbook” by Joseph Sugarman. Solid, insightful read – author shares his approach, which is a bit different from what I have read before. Good book for getting fresh ideas and learning how write copy with good flow.

Updates to come and thank you for attention.

---

Thank you. RON means that your run your campaigns on all of the placements available for the geo you are targeting. I believe what sahbak means is that you have to get high volume in order to get significant data on RON campaigns. Daily caps here mean your daily budget.

---

How many impressions per placement do you consider significant?

---

I was considering battery pins, yes, but decided to test launcher as you recommended . As for 360 Antivirus, it was one of the offers I tested before joining STM and it worked rather fine (but I could not continue running it due to critical mistakes).

And yes, I have been told that battery icon can have a decent effect on CTR. Actually, this would be the last element I would test . I guess that icon which looks native to OS interface may have a higher CTR, but, ofc, it can be competely vice versa. I will surely split-test a few theories.

---

Per placement, anywhere between 2x-4x payout should give you a good indication. Cutting based on impressions isn't the best yard stick... CPM varies wildly by placement.

---

Update 12

Nothing much to report. I decided to re-do my campaigns, because felt like angles were not strong enough and lp/banner copy could be otpimized more. Worked a lot on wording -- used book "Words that sell" and some posts from Boost Blog Traffic.

Now the campaigns are ready and will be live tomorrow if approved by AM/advert and Decisive.

---

Update 13

Waiting for campaigns to be approved. Haven`t spent much time on AM today, because had to work on my diploma (my thesis defence is next week). But between work sessions (I work by 45/15 sessions, that is 45 minutes of work and 15 minutes of rest) I watched a few insightful videos on copywriting and have a few theories on how to improve/optimize my copy.

I decided to test one more offer, so tomorrow I will spend time researching on it and maybe creating angles/campaigns for it.

Have high conversions

---

Final update

Haven`t been updating lately because testing and launching campaigns takes all of my time. To give a short update – tested a few offers, some of which worked, and some of which didn’t. I spend 50% of time on current campaigns and 50% on new ones.

Test, optimize, learn, rinse and repeat. Simple.

Instead of posting here, I will later share some case studies or useful lander tricks, so I will see you stackers on other threads.

Thank you for attention and have high conversions.

---

2x-4x spend doesn’t always give you enough data. If you run in some T3 country with avg bid of 0.05 and spend .75/placement (assuming you have an app with .25 payout), you`ll get ~15K impressions, which can be considered enough to analyze. But if you run in some countries with avg .25 CPM, you`ll get only 3K impressions, and that is definitely not enough to make a statistically significant decision. And as you said, CPM of placements varies a lot, so you might not even get those 3K impressions.

So I believe that one has to spend as much as needed to get enough impressions/conversions for statistical significance.

My approach comes down to gathering some rather decent amount of data and analyzing a placement by drilldown reports. Sometimes it takes more spend to make a rational decision and sometimes it takes less.

---

Thanks again sylar for such a stellar follow-along! Learned a lot from it. Looking forward to seeing future threads by you!

Have fun and good luck!


Amy

---

Home > Paid Traffic Sources > Mobile