Home > >

---

Push Traffic 101 - Trafficsources (39)

---

Trafficsources

To run your Push campaigns of course you need traffic and choosing the right trafficsources that work good for you can make a big difference if you succeed or not.

As you probably know there are many trafficsources out there who offer Push traffic.
This can make it little bit complicated (especially in the beginning) to choose a good trafficsource to start with.

Some sources are too expensive for testing, others have low volume, some sources allow more aggressive promotion and then there are sources that only have good traffic for specific geos.
The best combination is when the traffic is not too expensive, when you can target higher quality traffic and when there is enough volume for the campaigns you want to run.

In this post I will tell you a bit about different trafficsources, how they work and how you can see which one is good to use and which not.

I won´t tell you that source ABC is the best one to start with because from my experience there is not THE best trafficsource for Push traffic.
When I read threads here on STM there were users who had big success on sources that didn´t work for me and vice versa.

We all have different needs because we run different campaigns so it´s hard to name a source that works best for all of us.

But don´t get me wrong, this is not bad at all.

Different trafficsources have different advantages and disadvantages.
When you understand what to look for it won´t be hard for you to find good trafficsources for your campaigns.

Also I want to prevent that all users start running same or similar stuff on the same source.
This would only increase competition and would make it harder for all of you to succeed.

Important is that you don´t jump from one source to the next.
Try to find 2-3 sources that are working good for you and then stick with them.

When they are big enough you can run bigger campaigns on so few sources already.
Check my campaign HERE, there the most volume was from only two sources.

The better you know a trafficsource the easier and cheaper it will become to test and run stuff there.
When you focus on only 2-3 sources you have a better understanding about what verticals are doing good there and which geos have good volume with not too hard competition.
And when you run lots of traffic on a source and have enough stats for high quality blacklists and whitelists then the testing and running of new campaigns will also become much cheaper.

Sources for testing and sources for scaling

I make a huge difference between trafficsources for testing and trafficsources for scaling.

There are trafficsources that are good for testing and then there are sources that are very good for scaling but not that good to test new campaigns.
The big difference is what targeting options a trafficsource has to offer.

I never test new campaigns on a trafficsource that doesn´t support targeting by subscription age or user activity.
Without such targeting you would receive maximum volume but very mixed quality and also high volume from rather the lower end so you need to spend much more money to weed out the bad placements.

And before you are ready to scale to higher volume you first need to have a working funnel, otherwise you can lose lots of money.

So first thing you should check on a trafficsource is if they support targeting by user activity or subscription age.

For testing new campaigns is important: Quality over quantity.

Better test on lower volume but higher quality than on high volume with unstable quality.

Here are some examples to see how user activity/subscription age targeting can look:

Evadav



Propeller Ads



Adsterra



Pushground



Advertizer



Clickadu



It´s always good to start with very good quality.
Then you will have lower volume and mostly you need to bid higher but the difference in quality is worth it.

Some sources also allow targeting down to 1 hour subscription age.
These are the freshest users you can get but volume there is very low.
Usually it´s good enough to target 1 day freshness, that way you already get good quality but also enough volume to be worth it.

Your first goal is to find a working funnel and this can be archieved fastest and most cost-effective with higher quality traffic.

When you test your campaigns on the high quality traffic and you have non converting placements there you can be sure that these placements also won´t convert on lower quality traffic.
Better get rid of them before you start to scale to higher volume.

Again, for testing campaigns choose traffic sources where you can target by user activity or subscription age so that you can get fresh and active users for your tests.
On these trafficsources then start with high quality traffic and work your way down the quality levels when you have a working funnel and need more volume.

Then there are also trafficsources where you can´t target by user activity or subscription base.

That doesn´t mean that these sources are not good and that you shouldn´t use them.
Just because you don´t have these targeting options it doesn´t say anything about the quality.

It just says that these sources are better for scaling and not for testing.
Few example sources would be Hilltop Ads, Zeropark, Adcash, Admaven.

On all of them you can run very good campaigns.
But I first test on other sources and when I have a working funnel I use these sources to scale.

The difference is that on sources where you can´t target by activity you always buy all traffic with all different quality levels.
Fresh users who just subscribed, old users who subscribed a month ago, very active users, users who don´t click much.
Just all traffic that is available and that can make a really huge difference in performance.

On the other hand on these sources often have huge volume available and bids can also be lower than running only on expensive premium traffic.
So when you have a good converting funnel these sources can help alot to scale your campaigns to much more volume.

But again, to test new campaigns such high volume sources can be too expensive because you have to spend much more money to weed out the bad placements and find the good ones.

The bids

Now let´s talk about finding the right bids.

I recommend to always use CPC bids.
CPM bids can work as well but then you need to be 100% sure that your creatives CTR is good enough for it but I personally only use CPC bids.

For several reasons I like to to start with lower bids than the recommended bids from the trafficsources.

The first reason why I like to start with lower bids is that I can identify and block high volume/low performance placements for less money compared to starting with a higher bid from the beginning.
In many campaigns you will find placements that send high volume and don´t convert.
There´s no need to overpay for such traffic.
This alone can give you a small performance or quality boost already when you then increase your bids and don´t have to pay for non converting high volume placements.

The next reason is that it´s never to late to increase a bid but once you spent the money you can´t get it back.
When it comes to the bids you and the trafficsources work at opposing interests.

The trafficsources of course want to sell their traffic for as much money as possible, you want to pay as less as possible.
There is abolutely nothing bad or objectionable about it, that´s just how business goes.

But the trafficsources can´t give customized bid recommendations for everyone of us so the recommended bids are for all campaigns across the platform who run your targeting.
They are basically just an average bid that people run with your targeting but they don´t tell if they are a good match for your very own campaigns as well.
When you run low payout offers and many other campaigns for your targeting run good converting high payout offers then the bid recommendations could work very good for the high payout campaigns but would be too high for your campaigns.

We are performance marketers so our job is to run our campaigns based on our own stats and not based on general recommendations.
When you run the recommended bids from the very beginning where you don´t even know if your campaign is working or not it often becomes too expensive.
It´s much better to start lower but safe and when you see that your campaign is working you can still increase the bid.

Maybe you don´t get access to the best placements then right away and maybe you have lower volume then but this is absolutely ok because in the beginning it´s only about finding a working funnel and for this task a lower volume on higher quality traffic is absolutely enough.

Blacklists and Whitelists

On Push traffic the importance of placements is a little bit different compared to other traffic types.

When you run native ads or banners for example the user sees your creative directly on a website so there is a direct connection between the creative and the placements.
That means that often websites work better for related verticals when they are in a specific niche.

On pop traffic the pops get triggered when the user clicks on a link on a website so there is also a rather direct connection between website and ad although it´s not that connected as on native or banner ads..

With Push traffic it´s a bit different, there the user accepts to receive Push notifications on a website but the notifications he receives are not shown on the website itself.
The user receives the notifications directly on his phone and doesn´t even see the website that´s sending the push notification.
There is an exception, In-Page Push, but this guide is mostly about Web-Push and there many websites to collect push subscribers are not bound to one niche or vertical.

It can happen that the last time the user visited the website where he agreed to receive notifications is already a week ago and he still receives notifications from that website.
In this case there is no direct connection between the creative and the placement, the user probably doesn´t even remember where exactly he agreed to receive that notification.

This is why you often have the situation on push traffic that a campaign either just works or not.
The importance of the placements in connection with the landers is often not that high compared to other traffic types because often often the push collection landers are rather general and don´t collect that niche specific users.

On push traffic the funnel „Creative → Landing Page → Offer“ is much more important.

Nontheless you will find placements that just don´t work, no matter what you try and then there are also placements that show a good performance in all campaigns where you run them.

So on push traffic it´s also very important to know what placements are good and which not and target your campaigns accordingly.
Otherwise you can lose lots of money on bad placements

Usually it can take quite some time to collect enough stats for a good and significant blacklist or whitelist but let me tell you something...

For your first rough but already good working blacklist you can just take the stats from all different Geos together.

That´s the good thing about the rather non niche specific nature of the placemets, they often perform the same or similar for many different countries and aloso for different verticals.
The good placements are good in most (if not all) Geos and same goes for bad placements.

So don´t try to build seperate blacklists for each country from the very beginning.
First try to get a good general BL as fast as possible, when you have enough conversions in a Geo you can still create better BLs for the different countries.

Let me show you some examples from different sources:

Propeller Ads bad placements







Zeropark bad placements







Evadav bad placements







As you can see, it´s not 100% accurate but it´s working pretty good.
Many placements show similar performance across many different Geos.

That way you can build a first blacklist to work with much faster than collecting stats for each country separated.

For whitelists it works similar but there I rather use stats from only the countries where I run campaigns.
There are also many placements that convert similar in many different geos but for whitelists I usually bid higher than for blacklists so I want to be sure that the placements I bid on really work for my targeting.
Otherwise I pay money for nothing.

I tell a bit more about it using blacklists and whitelist the part about running campaigns or scaling.

You can also ask your AM at the trafficsource for a basic BL/WL to start and it can also help that you don´t have to start from scratch.
But from my experience my own collected lists are always better than the lists I receive from networks.

Another thing to consider a trafficsource for me as a main source for testing is the campaign approval time.
When I want to test something I want to test it now, not in 12 hours and not the next day.
Luckily many/most sources have pretty good approval times, some have even almost instant approvals but there are still some sources where approvals can take longer.
Such sources I wouldn´t use for testing where I want to work fast, they would be a better match for scaling.

Summary:

Try to focus on only 2-3 good trafficsources.
Important is to have at least one source with the option to target better quality for testing campaigns.
When your trafficsources have good traffic and high volume then 2-3 sources are enough to run $XXX-$X,XXX/day campaigns.
Start with lower bids and maybe lower volume for your tests to save money.
For scaling you can still increase the bid and and open wider targeting for higher volume.

Bonus Tip 1:

On some traffic sources you can run push campaigns on CPA.
For such campaigns you don´t per click or per impression, there you only pay a fixed amount when a conversion happens.
The volume for such campaigns is often lower and ROI is also often better when you run CPC campaigns so that the overall profit is often better on CPC.
But CPA campaigns can be a very good way to collect stats that you then can use for your CPC campaigns.

Bonus Tip 2:

Some trafficsources have specific channels or traffic for aggressive campaigns.
For some verticals that can help and result in better performance when you run more aggressive.

---

Damn! That's good. Can you share which networks have the option to run push on CPA?

---

Thanks for this! Could you clarify how to make a blacklist or whitelist?

I'm using propeller ads. Do I just write down the numbers of bad placements somewhere like notepad and paste that in when creating a new campaign?

How does white list work? ,🙏

---

There are several ones but the thing is that you mostly need a separate account or a good history with the platforms to run real CPA.

Few sources would be Hilltop Ads, Adcash, Adsterra.

On Clickadu you can run SmartCPA for push.

Yes, you take the non-converting zones end exclude them from campaigns so that you don´t receive traffic from these zones anymre.

But keep in your mind that you can only block placements because of bad quality when you have a converting funnel.

Whitelists are the opposite of blacklists.

For WLs you take the good converting placements and run campaigns where you receive traffic only from the good placements.

---

Great post! You've really cleared a few important things up for me. Can't wait for the next installment.

In response to what you said in the intro of your push guide about not being too specific and spelling everything out, I think it's a great idea. It is so much more use to me to understand the underlying principles of push and apply them myself. I want to learn push thoroughly and hopefully this guide will help me do that.

Gav

---

Would be great when it can help you, Gav.

Will try to get the next parts done ASAP.

---

Thank you for the valuable posts @twinaxe! Looking forward to the next

---

Always one nugget of awesomeness that is overlooked:

When you think about how lists are built, this has to be true. Way different than a global BL/WL for Pops. Push is a list of subscribers from wherever. I know dudes who have some of the biggest 'zones' on several of the bigger networks... and most just target everywhere, never a specific geo; just going for volume. But others, are normal affiliates collecting subs on their normal landers where they are already targeting specific geos, and when they're already targeting good traffic, they get good subscribers, etc. etc. etc.

Anyway, interesting to think about.

---

Yes, it´s a difference if you collect targeted subscribers yourself or if you buy from a network where the it´s rather untargeted/general.

---

Right, but I'm strictly talking about buying from networks. Not using your own list.

I collect subs for Advertizer and Propeller. So, I know people are using them because of the rev share. And I also know that whatever my sub id is on Advertizer must have good push traffic for US, ZA, PL, DE, FR because that's what I focus on during my normal camps.

I also have a bunch of random subs from TH, SG, IT, ES, that I have collected from random testing, but I doubt their quality is good for those geos since I don't really focus on them.

But for someone that found good subs on a US camp and wanted to try the same 'zone' for a DE camp, they would be pleasantly surprised because my list has quality subscribers in both.

---

Alright, misunderstood your previous post.

Got it now

---

I had in mind split testing push subs between Monetizer(Advertizer) and other networks like Propeller, what do you find better, Advertizer or Propeller? I also think that traffic on sources where its abit more expensive to buy clicks it might be better to join as a publisher than on cheaper sources but I have yet split-tested that, what do you guys think about that ?

---

Thanks! Nice post, alot of useful info. it was mentioned that we should start with a lower bid, and its likely to get traffic from the bad zones. but wouldnt that put an extra unknown into the testing, as compared to starting with a higher bid. Since for the initial testing, we need to go for quality. if its the bad zone, i wont know if its the zone or my funnel that was the problem?

For a sweeps offer say a payout is $2. what is the maximum CPC bid that you are willing start testing an offer on? Also, if the recommended CPC was very high, would you still suggest to test an offer over there with your own CPC bid price which might be alot lower? since my concerns are there is little opportunity for scale and receiving lower quality traffic

---

This is why for testing I only use sources with quality targeting and use high quality for it.

Of course it´s still no guarantee for success but it helps alot in the process.

As always there are no hard numbers.

For example it depends on average bid or on available traffic.

When there is huge volume available you can bid lower on high wuality traffic and still receive enough volume.
When there is only low volume available you need to bid higher to get enough volume.

Really man, from my experience the recommended bid is always higher than what I am willing to spend for tests and often also for scaling.

There´s a difference between testng and scaling.

For testing I want to have good quality and volume isn´t that important.

For scaling however I need the volume, but with a good converting funnel I can also bid higher then.

---

@twinaxe, great value!

Do you set your bids based on a certain CR% that you're expecting?
For example, the offer is 2.4$, you set your bid to 0.024 if you're expecting around 1% CR (or more) for the offer to break-even while you're testing?

---

Not really, I rather try to find a good compromise between low enough bid and enough volume for testing.

What you are doing is to expect a total CR from TS to offer.
But along the way you also have different landers, maybe different creatives.

So the 1% total could be from 10% LP CTR and 10% offer CR as well as 1% LP CTR and 100% offer CR.
It also could be 20% LP CTR and 5% offer CR as well as 5% LP CTR and 20% offer CR.

You can do such calculations when you previously had good campaigns running already with same targeting and funnel so that you know how your creatives and landers perform.
Then you can get an idea about how a good offer converts on that funnel.

But again, usually I don´t do such assumptions and set my bid based on that.

---

Thanks for your reply

I also wanted to ask you, do you collect global BL/WL data from Desktop & Mobile with in-difference?

Also, read something you said in jaybot's FA about taking zones which have lower EPC & making different lower bids for them, running in different WL for the same campaign in different bids in a sense (I think). Are you doing that on Push, too?

---

For push traffic I separate desktop and mobile.

Yes, just because several placements are profitable it doesn´t mean that they all perform the same.

That´s why it´s better to put similar performing placements in "bid baskets" and bid accordingly.

For very good placements you can bid higher than for "just" good performing placements so it´s better to keep them separated.

That way you have more control over the campaigns and the bids.

---

Nicely done @twinaxe! A lot of really userful information. We are excited to see what will come next!

---

Thanks, btw great to see that you now implemented rules for source level as well

---

A question for you How do you deal with Trafficsources like @pushground which have a relatively small amount of sources per GEO (they call their sources suppliers), but have alot of zones inside those sources? This is similar to Zeropark Push I believe. Do you then optimize these zones inside the sources by regular x payout rules like you would treat zones in Propeller? Those zones can have traffic spread out between them quite alot though, what do you do in that case?

On the other side of the fence, you have a Trafficsource like EvaDav, which can have 10,000 sources per one GEO campaign (which have zones inside but aren't optimizable to the user) And traffic is distributed a lot between them too, needing to go through quite alot of traffic to gather sufficient data per EACH source. You just increase your budget?

For the "big source" type of Trafficsources, do you ever cut those big sources? Any x payout amount? (well.. for SOI's ).

---

I treat the zones inside the sources like normal zones in Propeller Ads.

You mean that you receive only few clicks per zone but traffic from many different zones?

If that´s the case, I still treat them like normal zones but when the total performance of the supply is too bad I exclude the whole supply.

That decision is made on case by case.

Same as with Pushground, I treat the sub sources like zones in Propeller Ads but when the overall performance of the sources is too bad I exclude the whole source.

And of course you can exclude the sub sources on Evadav as well.

Excluding a source would be something like s123, excluding a sub source inside source s123 would be s123_sub12345



What do you mean with "big source type of Trafficsources"?

Do you mean big sources/supplies inside a trafficsource or do you mean sources/supplies in a big trafficsource like ZP?

---

Yep I mean I receive not alot of clicks from each zone, because those clicks are spread out across many zones.

Wow, I never knew that, I thought all of these were just sources Because TheOptimizer calls these all "sources" on their platform
Ok. Good to know that they don't actually have 10,000 sources per campaign, but rather those are sources + zones combined

For that I mean Trafficsources which there have sources which include ALOT of volume, and hence, the traffic source usually doesn't have a lot of sources per one GEO campaign.

How would you cut a Big source with a simple European SOI offer?

---

Good stuff. Looking forward to more

---

Brilliant! Great share!

---

It´s hard to give an exact answer there.

Treat the zones inside these sources like normal zones but when the whole source sucks just block it.

It doesn´t matter if there isn´t enough traffic left then or not, when it doesn´t work there´s no need to keep it running.

But you should also check if it´s really the sources performance or if it´s maybe the offer.
When you test it with an offer that is converting good on other sources then it´s probably the source that´s not converting.

When the offer isn´t converting good it would be good to re-test the source when you have a better offer.

---

thanks for sharing see you in the next chapter

---

Thanks,for the informative article!
I have a question: Do you mean we should test on sources with subscription age to find bad placements, then scale with the ones without subscription age?
If so, do you mean these sources actually share a lot of common placements?

---

Someone correct me if I'm wrong, but I think what he was getting at is that you should test on sources that have an ability to filter the subscription age so you can optimise your funnel on good quality traffic that has high activity and more engagement. Then when you have it optimised, you can then scale on other sources.

As far as I know, you can't share placement data between sources.

Hope that helps, Gav

---

Gtrat share , Especially what you mentioned about Subscriber Age factor ... this matters alot and can alone make or break your campaign .... I tested 12 Push networks this month, 7 out of 12 worked and 4 were exceptional ... ClickAdu worked amazingly well but their support was horrible, Dead Slow so I had to dump it ...

---

It depends, on some trafficsources you just can´t run different campaigns with same campaign link.

Then of course you have to duplicate in the trafficsource and in tracker.

For a better overview and cleaner stats it´s better when you only duplicate it on the TS and use the same campaign link.

The problem there is that some sources could count it as duplicate campaign (because that´s what it is) so that you don´t receive the full "new campaign" boost.

That´s why I mostly duplicate it on the TS and in the tracker.

In the tracker then I just select all of these duplicated campaigns and check their stats together.

That way I have many more campaigns in the tracker but it´s safer to get the boost then.

---

Many thanks for the great tutorial!

I'm running Push on Propeller and I've noticed a curious trend. Basically, a campaign starts off great, with some good conversions, but then tanks after only a day or two. The traffic VOLUME doesn't go down, but the quality seems to. My CTR is 1-2% (which I was told is fine/average), and I'm bidding at around the recommended bid, high user activity only. I've seen this happen with multiple geos, and in multiple verticals.

Any insights?

---

Hi,
Thank you so much for your tutorial. I've been waiting for a clear and easy to follow guide on PUSH for a long while now, much appreciated.

When you say "enough volume" for testing. What does it mean? Visits on landers?
If you could expand more detail on this a bit, I'd really appreciate it. Like how much is enough for testing and where to look on the tracker / Traffic Source provider.

Thank you very much.

---

"Enough volume" means there has to be enough volume available on the trafficsource for your targeting.

As always there are no fixed numbers to work with.

How many visits you receive to your lander has not only to do with the available volume but also with your creatives CTR.

In the end it´s only important that you have enough volume to run tests that bring you fast results.

It won´t help when you set your targeting too tight or when you bid too low so that you are left with next to no volume.

During the test stage it´s all about working fast and effective to see if something has potential or not.

Testing is not about having a perfect campaign setup.

To test fast and effective you just need some volume to get results on first day or latest second day.

When you run tests on such low volume that you need a week or so to gather enough data then it won´t get you anwhere.

To sacrifice too much volume for higher quality is just not worth it.

For tests it´s always good to use high quality traffic but the best traffic on earth won´t help when you have no volume available.

So when you target high quality traffic but volume is too low just widen the targeting a bit or increase the bid to get more volume or when you creative CTR is too low work on your creatives to get more from the already available traffic

---

Thank you so much for clear explanation. I'm on first day of testing, almost having a winning lander now (but still need to use that lander to test for winning offers right?)
My creatives CTR is about 1% and judging from reading some of comments in your tutorial it's OK.

I get it now.

1. To test need volume to get results in a day or two. (Not WEEKS)
2. Since already use high quality PUSH traffic, just get enough volume to get the test done quickly, by adjust targeting / increase the bid / work on creative CTR.

Cheers,

---

@twinaxe @vortex i have launched a camp on propeller ads and can see that there is litrally hundreds of zones that i am buying traffic from, yes a lot of them have only spent a few cents but what is the best way of optimizing to bring the zones down because i feel like it is way too much to ever be profitable. thanks

---

Propeller Ads bad placements







Zeropark bad placements







Evadav bad placements







As you can see, it´s not 100% accurate but it´s working pretty good.
Many placements show similar performance across many different Geos.
@twinaxe sorry can you explain this part little bit more. So just to confirm what exactly is this showing, is this a list of geos that a running bad...sorry as you can tell i dont get it and have read in 10 times trying to understand

---

@wifishmnoney

With this I mean that often placements show similar performance across different geos.

A good placement is often good in different geos and a bad placement is often bad in different geos.

This can help to build first rough blacklists faster compared to building lists specific for each geo separated.

---

Anyone want to start a telegram pop/ push mastermind? Or maybe there already one existing?


Sent from my iPhone using Tapatalk

---

Home > >