Home > Push Traffic >

---

Road to XXX/day - Follow the Pianist 🎹 on his Journey 🔥 (31)

---

Hello everyone ,

First of all, I wanna introduce myself.. I'm Ayoub from Morocco, 23 years old, just graduated as a Software Engineer, I'm a Piano Player and now a full time Affiliate Marketer.

I feel so lucky that I joined STM, The amount of information that I found here is priceless. So thanks to all members who provide us with those golden nuggets.

After going through Amy's 40 day guide (thanks a lot @vortex), I learned a lot of things about the process working with Pop, and now I wanna graduate to Push.

My goal is to hit XXX/day and make it as consistent as possible. Some of you may say that i shouldn't set Numbers as a target, I respect your opinion but this is a realistic achievable goal for me, so let's move to the action plan:

Here is my action plan for this follow along:

* Initial Budget: 4000 $

** Software Per Month :

- Hosting (FunnelFlux Managed LP) : 30$
- Tracking (Voluum Profit) : 138$
- Spy Tool (Anstrex Push) : 80$

*** People :

- Translations (Fiverr.com) 5$ x3 = 15$
- Cleaning/Optimizing Landers (lnnrt.com): 20$ x 3 = 60$

Note: People cost depends on number of landers, let's assume that I'm gonna use 3 types of landers.

**** Total Business cost : 323$

*****Traffic Cost : 3677$

- Vertical : Sweepstakes
- Type of offers : CPL (SOI/Email Submit)
- Geos : Germany, South Africa
- Working with one Affiliate Network at the moment : Clickdealer.
- Make a list of Hot offers �� by asking my great AM (BIG shoutout to Andrey ).


- I'm using Voluum as a tracker (Profit Plan)


- Go to Spy tools and see what others are running. I'm using Anstrex Push for spying.
- Follow the famous quote : Immitate then innovate.


- Wheel of fortune
- Gift Boxes
- Quiz

- PropellerAds
- ZeroPark

- Continually Test to Find the Best Performing Offers
- Continually Test to Find the Best Performing Landing Page
- Continually Test to Find the Best Performing Ad Creatives
- Ask for Offer Pay Bumps
- Scale to other sources
1- Preparation

2- Split-test Offers

3- Split-test Angles/Landers

4- Optimization

5- Scaling $$$



I would be more than happy if you subscribe to my thread and help me along the way. I will keep this thread updated. let's make this follow along full of golden nuggets.

Thank you all.

Ayoupianist.


PS: I'm using @erikgyepes follow along structure, thanks Eric!

And the Campaign Canvas is from @dr_ngo Article!

---

Hi Ayoub, sounds like a good plan so far.

Waiting for an update

---

I like how arabs are coming to the industry , finally
Welcome to the Club man
wish you the best with campaigns

---

hi
i am excited of results ..
Good luck



Sent from my iPhone using STM Forums mobile app

---

Good luck, hope you achieve your goals!

---

Thanks guys for following my journey.

So here is what I've done so far in Day 1:

- I got a list of good performing DE offers from my AM at Clickdealer and chose 5 to test.

- I spent 2 Hours on my Spy Tool trying to get an idea of what others are running then ripped some landers, and saved 20 creatives (icon/banner/title/description).

- I already have a lander that I tested on Pop in Germany, so i started with it to take action immediately (one big thing I learned after Amy's guide is that you should always start with what you have, don't try to make things perfect from the start, just take action and leave your overthinking for analyzing the data )

- I customized the lander to fit the chosen offer

- I added the back button script (Monetizer)

- I uploaded my lander to my Hosting (I host landers on FunnelFlux Managed LP)

- I tested the lander's speed with GTmetrix (I chose the closest server to Germany which is Uk), and here is the result:



1.1s not too bad right?

- I set up the campaign on Voluum (Adding 5 offers, 1 lander)

- I launched 2 campaigns for the 1st Offer : 1 on PropellerAds and 1 on ZeroPark with 1 Creative (which was running longest on spy tool for this specific offer)

Here are my results of Day 1:



Seeing green from the first day makes you very happy

First thing I noticed is that the CTR in ZeroPark is too low compared to PropellerAds.


Anyway, here's my first day summary:

Offer 1:

COST: 59$

REVENUE : 37$ (+2.5$ from Monetizer)

PROFIT: -19.5$

ROI : -35% 🍎


That's all for the first day.

---

Damn. Looks like you're ready to go already!

Those are really good numbers for a first push campaign, especially in DE

Propeller push is a lot easier to test and get some good numbers with. Is that with just Hi Activity? Or Hi+Med? Or Hi-Med-Low?

ZP Push is tough as you can't choose subscriber freshness (unless it's changed). But if you ever crack it, you'll be sittin' on a mountain of gold They also allow more aggressive creatives, so... feel free to experiment with that.

I personally would stick with Propeller first and see how much you can milk it before scaling to ZP.

But I'm no push expert

---

Thanks man for your advice

Yes, this is High Activity Only on PropellerAds.

ZP is so tough as you will see in my upcoming updates.

---

I hope you all doing good

Okey, for day 2 I have 7 campaigns launched:

Campaign 1:

* This is the one from Day 1.

Offer 1 / Voucher / DE / High Activity Only / PropellerAds

CTR: 35%

CR: 6%

COST: 36$

REVENUE : 51$

PROFIT: +15$

ROI : +43% 🍏

Campaign 2:

Offer 1 / Voucher / DE / Medium Activity Only / PropellerAds

CTR: 30%

CR: 0%

COST: 13$

REVENUE : 0$

PROFIT: -13$

ROI : -100% 🍎

Question 1: Not even a single conversion for this user activity, Is this normal?

Campaign 3:

Offer 1 / Voucher / DE / Low Activity Only / PropellerAds

CTR: 15%

CR: 9%

COST: 55$

REVENUE : 51$

PROFIT: -4$

ROI : -6% 🍎

Question 2: Low is doing better than Medium, I was expecting the opposite. What do you think?

Campaign 4:

* I don't really know what happened here. I reached out to my AM in ZeroPark and said that my bid was too low.. so i've been outbided for good placements. Also he recommended to start my camps with 100$ daily budget, and sent a list of top sources.

Offer 1 / Voucher / DE / ZeroPark

CTR: 2%

CR: 5%

COST: 74$

REVENUE : 10$

PROFIT: -64$

ROI : -86% 🍎

Question 3: Is it necessary to start campaigns with 100$ daily budget on ZeroPark?

Question 4: CTR is too low compared to PropellerAds, How do you explain that? (I know that CR is more important than CTR, but the CR is pretty much the same in both networks, this is why I compare the CTR).

Campaign 5:

* I'm split-testing another 2 offers here.

* Green from the first launch again!

Offer 2 & 3 / Voucher / DE / High Activity Only / PropellerAds

CTR: 34%

CR: 7.5%

COST: 24$

REVENUE : 32$

PROFIT: +8$

ROI : +35% 🍏

Note: I had issues with this campaign later being rejected due to use of copyrighted content on my lander.

Campaign 6:

Offer 4 / iPhone / ZA / High Activity Only / PropellerAds

CTR: 39%

CR: 1.43%

COST: 12$

REVENUE : 8$

PROFIT: -4$

ROI : -35% 🍎

Campaign 7:

Offer 4 / iPhone / ZA / ZeroPark

CTR: 15%

CR: 1.65%

COST: 21$

REVENUE : 8$

PROFIT: -13$

ROI : -62% 🍎

****************************

Notes:

* I am using 1 lander so far for Germany. (I should split-test more types of landers).

** I am split-testing 5 creatives. After enough data for statistical significance I disable the less performing ones.

( I use this tool : https://marketing.dynamicyield.com/bayesian-calculator)

*** At this phase, I haven't made any optimization yet (cutting bad placements, changing bids.. ).

****************************

So here's my Day 2 summary:

COST: 237$

REVENUE : 162$ (+11$ from Monetizer)

PROFIT: -63$

ROI : -31% 🍎

****************************

Voluum Stats :



I'm really looking for your opinions, tips, advices, answers .. Thank you all!

---

It´s exactly as jaybot said.

On ZP you basically run RON without any targeting for subscription age at all.

That´s great for scaling but not that good for testing.

I would also recommend to test new campaigns on Propeller and target high activity.

When it´s doing good also test it on high + medium activity.

When it´s doing good there as well and you need more volume to scale test it on ZP as well.

Hard to tell without more info.

But it´s not unusual, the $13 can be split across many different placements for example so that the single placements itself didn´t receive enough traffic to judge yet.

I also have campaigns that look bad in the beginning but are much better when they run for longer so that all placements have the chance to receive more traffic so that you can identify their performance.

The thing is that on medium activity you can get a nice traffic boost compared to high activity.
That means more traffic from more placements so that you need to run the campaign for longer to get a better impression about it.

It doesn´t help much when you then receive from 1000 placements with only 1 click on each placement.

Can happen, it´s all about the best balance between bids and conversions.

For example when a campaign is converting good but the bids in medium activity are too expensive it can be that it´s still converting on low activity and the lower bids then make it worth it.

But in most cases you are right that higher activity usually works better than low activity.

Nontheless the medium campaign probably has not received enough traffic to make real statements there.

It´s not necessary but it can help.

Propeller and ZP are two different platforms.

On Propeller you can test on lower budget when you target high activity.
That way you can run on lower volume but higher quality.

On ZP you can´t target that way, that means you need to spend more money because you also have much more volume for your campaigns.

Yes, for sweeps I test at least questionnaire, gift box and spinwheel.

Great that you use the SS calculator.

But also use common sense, often the trafficsource will send different volume to the different creatives anyway.
It´s based on the performance for the source.

Higher CTR = more clicks = more money for the trafficsource = more volume.

So it doesn´t mean that the creative that is doing best for you in terms of performance is also receiving most volume.

Often when you use 5 creatives you can see that after some time only 1 of them will receive the majority of the traffic.

That doesn´t look that bad

---

Thanks man! Makes sense now..

I see.. I will let it run longer!

That's my next step actually! I'm waiting for the other landers to be translated..

Totally right man!

Thank you so much for your tips! I really appreciate it!

---

Best of luck in your follow along will be watching

---

Hello guys, Sorry for not updating my fa yesterday..

So I decided to follow @twinaxe advice, which is : some sources are good for testing and others are good for scaling. So i will focus on PropellerAds at the moment.

Also I will focus on 1 Geo for now (Germany), and test multiple offers in it and try to scale the ones that show potential right off the bat.

I don't like the preparation part because it's so time consuming especially preparing the landers and translating them.. I guess it's either I pay with my time or with my money (using OHT for example is more expansive than fiverr, and for now every dollar counts to spend in ads) so I'll just use the hard way which is making an Excel Sheet with all lander content and giving it to someone on Fiverr, then replace the whole lander sentence by sentence. I could just rip a lander in the exact geo and test it out, but the translation maybe incorrect so I won't have the right decisions.

Question 1: How do you guys handle this part?

Also, I should make templates for landers to make them easy to adjust to offers.. this will save me a lot of time. ( I read that on @offshore follow along, thanks buddy )

so let's get to my results so far ( this is the last 7 days results):

Campaign 1:

Offer 1 / Voucher / DE / High Activity Only / PropellerAds

CTR: 29.84%

CR: 6.26%

COST: 459$

REVENUE : 500$

PROFIT: +41$

ROI : +8.96% 🍏

Voluum Stats for campaign 1 : https://prnt.sc/r0fclt

Question 2: I launched a WL campaign for this offer with the exact parameters as the RON campaign, but I had negative ROI. I'm wondering why? Is this not the right way to optimize? P.S: I blacklisted the WL placements in the RON campaign so the audience don't see my ads twice.

Voluum Stats for WL campaign : https://prnt.sc/r0fco3

Question 3: Is the data I gathered so far in the campaign 1 enough to make a global BL?

Campaign 3:

Offer 1 / Voucher / DE / Low Activity Only / PropellerAds

CTR: 13%

CR: 7%

COST: 315$

REVENUE : 186$

PROFIT: -129$

ROI : -42% 🍎

Voluum Stats for campaign 3 : https://prnt.sc/r0fd97

Question 4: What can I do to get this campaign to green? Should I test more creatives? Or reduce the bid?

So here's my Week 1 summary (PropellerAds Only) :

COST: 999$

REVENUE : 826$ (+95$ from Monetizer)

PROFIT: -78$

ROI : -8.46% 🍎

Voluum Stats : https://prnt.sc/r0ffo0

so my next steps are :

- Make templates for landers to simplify this task for me.
- Translate 2 more types of landers.
- Test other offers in Germany with my BL/WL.
- Try to scale my profitable campaign 1.

Any tips guys that could help me would be much appreciated.. Thank you all!!

---

Geez... no slowing you down at all!

I personally would focus on the profitable camp so far, Camp 1, as it seems lots of potential there. Get it working with Hi+Med, and once you can handle that volume with profit, scale to other sources, first pushground, advertiser, richpush and then if it's still looking good, flood it with traffic on megapush

---

Hi,folks! I want to advise you to use MegaPush, because our network is a full-fledged DSP platform with ALL push traffic from all around the world in ONE place! Thus, you can try the traffic of more than 100 partners / push traffic sources using one simple MegaPush account and one balance

---

Thanks man for your advice! will do that!

---

Exactly the same targeting, same bid and so on?

Well, how did you consider your placements to be worth to run in a WL?

Did you take placements that had at least 1 conversion or only placements with several more conversions?

Check your placements and then put non converters in a global BL.
I usually also start with global BL and when I have few hundred conversions for a single Geo I check placements for that Geo only to make more specific BL.

Many placements are interchangable across many Geos but finetuning can give you a little boost as well.

For creatives, mostly you will see that the trafficsource will send most volume to 1-2 creatives only anyway when the campaign ran a bit.

You run that campaign on low activity, this means you have much more volume.
Cut more aggressive and get rid of bad placements fast, you still should have enough volume left.

But the best way to run on low activity is to run WL campaigns with proven placements.

But to do so you need enough stats first.

This doesn´t look bad.

One more thing, compare the cost stats from your tracker with the stats from your trafficsource.
Real cost can be a bit higher because of clickloss.

Shouldn´t be much for Push traffic but happens nontheless.

---

yes I took placements that had at least 1 conversion.

what's a rule of thumb for cutting placements here? is it x2 payout with no conversions?

So far so good haha

Yes I've done that and found that the clickloss is around 3% of the cost shown on the tracker. (30$)

Thanks man for your tips.. will absolutely implement them!

---

Thanks man for your response.. I have a question though:

Should I run Hi+Med combined? or Should I separate them to 2 different campaigns?

---

Don´t do it that way.

1 conversion can only be a lottery conversion.

In a WL you should have only placements that convert more often so that you can see that they show stable performance and not just a lucky hit.

When there is enough volume I am even stricter, with a good converting funnel I blacklist placements with more than 1x payout in loss.

When there´s not that much volume keep it running till 2x payout so that you don´t risk to be left with next to no volume when you cut too aggressive.

I run them both together.

---

Alright thanks man..

I have more questions please:

1- Do you blacklist by conversions or profit? for example if a placement makes conversions but still in loss do you blacklist it or let it run longer?

2- Isn't it more logical to separate Hi and Med to see the performance of each user activity?

3- Looking at my stats of Low campaign, I see that only one placement made 19 conversions (183 clicks) and the others made 1 conversion each. So I should go with Blacklisting here right?

---

It depends on how much in loss these placements are.

When a placement converts I would keep it running till 2x payout in loss, especially when it showed good performance before it can be good to let it run for a bit longer.

When it's more than 2x payout in loss I blacklist it.

But when I know the campaigns performance good and it's converting very well then I would also kill placements with 1x payout in loss.

Sure it's better when you want to compare the two activities directly.

But when I start the campaign and test on high activity then I only target medium when I have a working funnel already.
Then I want more volume.

So I just target high and medium together for more volume and also some cheaper clicks from high activity.

And when you use the user activity token you still can compare the two activities in your tracker.

When the funnel is converting good I would kill all placements with 1x payout and no conversions.

But also keep an eye on your volume, if there is enough left then so that you don't cut too aggressive and blacklist too many placements.

---

Hey Guys

In the last 2 days I tried to take notes of all your valuable tips, a special thanks to @twinaxe and @jaybot and all the other members that are following my journey.

I decided to focus on the first campaign since it has potential. So i analyzed the data for the past days and here are my results:

1- Browsers:

Most of the traffic comes from 2 main browsers (the others don't make conversions and have very little amount of traffic) so I included just the main 2 in my campaign.

2- Hours of day:

There are hours when I don't get any conversions at all, and hours when the CR goes down a little bit (I break even here) and hours when the CR is very good.

So I excluded the hours when I don't any conversions and left the other two.

3- Languages:

I included German only ( I read somewhere in the forum that this is good for quality leads, because there are a lot of immigrants in Germany who speak other languages )

4- OS & OS Versions:

I found 2 OS Versions that are losing me money, so I excluded them. They represent around 16% of the available traffic in the estimation chart.

Voluum Stats (last 7 days) : https://prnt.sc/r1clls

5- Bid

So here I made a HUGE mistake, I raised the bid by 0.02 because I thought that I will get more traffic on the High activity. However, my campaign went to -16% ROI.

Then, I set it back to the old bid and then it got even worse to -32% ROI. I guess when you play with the bid like that the campaign gets weird.

Question 1 : What's the best approach here? Isn't it that raising bid will get me more traffic on the High activity?

Voluum Stats (last 7 days by day partying) : https://prnt.sc/r1cmnf

6- Zones:

I followed @twinaxe advice of cutting bad placements. Here's a recap :

IF ( good converting funnel AND there is enough volume AND no conversions ) THEN (blacklist placements with more than 1x payout in loss )

IF ( placement converts OR there is not enough volume ) THEN (keep it running till 2x payout )

I hope that this is correct @twinaxe ?

So based on the rules above i started to build my global BL.


Okey, so that was my analysis of the first campaign (High only) which I stopped since it showed some weird behaviour after raising the bid by 0.02 and then set it back.

Today, I launched 5 High+Med campaigns with 5 different creatives ( I fixed the headline and changed the icons and banners), because @twinaxe said that after a while the traffic source will send most of the traffic of 1 or 2 creatives. So my goal here is to separate them to find the best creative ( Highest CTR ).

Action : Split-test more creatives > increase CTR > more clicks > (hopefully) more conversions

So after a few hours, here is my first High+Med campaign (creative 1) results :

High+Med (Creative 1):

CTR: 44.3%

CR: 11.43%

COST: 9.3$

REVENUE : 27.6$

PROFIT: +18.3$

ROI : +196% 🍏

Voluum Stats : https://prnt.sc/r1cl9s

Note: I lowered my bid in the campaign above bellow recommended to see what will happen.

The other campaigns don't have enough traffic yet. (same bid but different creatives)

Will update results when I have enough data.

That's it for now, if you have any advice please don't hesitate to share it with me.. I still have a lot of things to learn..

---

I wouldn´t do it that way.

When the non converting browsers only have very low volume just keep them running.
Important is that the browsers with most volume are profitable.

When you run only the two main browsers you risk to lose volume on the whole campaign for targeting too tight.

For how long do you run the campaign already and how much stasts do yiu have?

Dayparting only makes sense when you have enough longterm stats so that you can see patterns.

Yes, targeting the countries native language can be good for quality.

But it can happen that you lose too much volume when you only target one language, then you can try to add english as well.

In a good campaign OS versions don´t make a big difference.

Try to run your campaigns without such micro targeting if possible.

Yes, to get more traffic on Push at the same activity you ened to bid higher and find the sweetspot of bid, volume and performance.

But I recpmmend against changing bids in running campaigns because as you can see it can destroy your campaigns.

When you want to test the campaign on a new bid create a new campiagn for it.

Then you can also compare the different bids much better.

Yes, that´s correct.

But keep in your mind that a global BL is a good start but it´s not as accurate as a country specific BL.

Many placements are interchangeable across different Geos but nonetheless there are few differences.

So when you have collected enough stats try to build more country specific lists.

But to start a global BL is perfectly fine

WHY?

Why did you do that?

You test a campaign, it´s working right from the start and instead of trying to earn more money from the campaign you lower the bid and risk to make less.

---

I get your point.. I will include all browsers then.

It's been more than 1 week now.. 157 conversions..

Is this enough to make statistical significant based decisions? or you use another metric to do that like visits or clicks or maybe duration?

You are right! Other speaking languages represent almost half of the volume..

Alright man, will try that.

Will do that from now on.

No, my bid in the first campaign was above recommended, then when I played with it too much the campaign's behavior changed badly.

I was talking here about a new campaign that I launched with a bid bellow recommended to test bids starting from the bottom and raising them up to find that sweet spot.

As you see the campaign was at +196 ROI with that bid (bellow recommended).

I launched other campaigns with higher bids to compare.

Is this the right way to do it?

My question here : will the audience see my ads many times ?

---

Following. all the best buddy

---

Thanks man.. I really appreciate it!

---

You can check about time patterns but also try to see the bigger picture.

It´s normal that there are hours a day that have better performance and hours that are not that good.

For example, usually it´s safe to say that 2-6am are not that good hours.

But then also check the other parameters, mostly they are connected pretty good.

These low CR hours usually also have low volume and low spend = not that much loss.

On the other hand the high volume hours also have higher spend but therefore also higher CR.

So normally I would just keep it running 24/7 when it´s like above, then it´s not really necessary to daypart.

I would only daypart when the volume and adspend is still (more or less) the same in the bad CR hours.
Then it can make a difference in performance.

But when the volume and adspend is low in the low CR hours anyway I just keep it running.

For me it also depends a bit on the trafficsource how I work with language targeting.

When I can target multiple languages I would stick with native language and english, when I can target only one language or all I would often run all.
But this also depends a bit on the Geo.
For example in SG they speak English, Chinese and Malay.
When I can target multiple languages I would target these, when I can only target one or all on the trafficsource I would keep all languages running.

Alright, so you didn´t lower the bid on the successful campaign but you created a new campaign with a lower bid.

Then it´s all fine

Yes, that´s right.

I like to keep campaigns with different bids separated for two reasons:

First reason is that cganing the bid in a successful campaign can destroy the campaign.

Second reason is that you can compare the performance of the different bids much better when you keep them separated.

To be honest, I don´t know.

But I could imagine that it´s possbile.

But you are running the campaigns on CPC so you only pay when a user clicks your ad.
Then it shouldn´t be a problem when he sees it more often.

---

Hey guys.. sorry for not updating my FA since 3 days.. Well, I've been losing some money

Here is a screenshot of the last 3 days : https://prnt.sc/r3ocdr

(Still focusing on 1 offer and trying to make it profitable)

CTR: 23%

CR: 6.5%

COST: 397$

REVENUE : 276$

PROFIT: -121$

ROI : -30% 🍎


I launched different campaigns for different bids, i started with bellow recommended then added 0.01 each time.. but the problem is the highest the bids the more money I lose.. so I stopped the campaigns that were losing money.

Then I launched new campaigns today with different activity level (High+Med/Low) and different platforms (Mobile/Desktop) and I'm testing the CPA Goal Optimization provided by PropellerAds.

I also picked the two best performing creatives from the previous campaigns (icon+banner) based on statistical significance and used them on my new campaigns.

Now, i'm splittesting the headlines, I added an emoji at the end of my headline and the CTR went from around 2% to 3% ( I read that on the DO's and DONT's of Push thread)

So now i'll be running the new campaigns for sometime then update my FA with the results.

I also signed up to some other sources following the great @jaybot advice : Pushground, RichPush, Adsterra

I'm testing them out too.. waiting for some data..

If there is any advice don't hesitate to share it with me! I know I can improve faster with your tips

---

Well i guess you all excited about the results of the campaigns I launched yesterday

So, 2 campaigns did really well ( i'm surprised actually ) check this out : https://prnt.sc/r42738

Those are High and High+Med activities, with CPA Goal and WL.

Well, @twinaxe said that I should run WL only on proven placements that have multiple conversions.. But I still wanted to test out those placements who have only 1 or 2 conversions which I have a good list of.. so these 2 campaigns are running on those placements and are doing well till now.. Maybe when they receive enough traffic they well have more conversions and then i'll move them to my first list which has only proven placements.

That's the bright side of my campaigns, let's move now to the dark side lol

Here is how my everyday looks like : https://prnt.sc/r42bjn

These are my yesterday's overall results on PropellerAds.. with different campaigns..

My biggest loss was with Desktop RON campaigns..

Question 1: Do Mobile and Desktop share the same placements? So I can use my Mobile WL on Desktop?

Question 2: How much time should I let a CPA Goal campaign run till it can optimize well?

That's it for today!

PS: if you have any advice, don't hesitate to share it with me.. I started this FA to learn from you guys Thank you all!

---

cool, which offer types and which traffic ? thanks.

---

Home > Push Traffic >