Home > Paid Traffic Sources > Push Traffic

---

Recent Test With Push, Advice needed (7)

---

Hi, there, I have been testing push traffic recently.
I started with Megapu.sh and gradually bought more traffic from Admaven. (Megapu.sh’s traffic is great but does not seem to be able to limit traffic according to your daily budget limit, easily spend over 3x of your budget limit. Admaven might have implemented some view through rate + impression prediction method to control the traffic, so it does not over-spend your budget too much.)
I was running pop traffic before this and I am only doing casual dating vertical (started end of last year, still new to affiliate marketing ; ) ). What made me wanted to move to push traffic was that it had way less bot traffic than pop (tested with google verification code and also device information check code in the landers) and it generally has lower cpc as compared to pop traffic, so I get more clicks to the offer at the same price .
But still there many things very puzzling about this traffic. I would like to hear more advices from people who are experts in push traffic. : )

Here are my questions:

1. I did see a trend that campaign roi drops very rapidly over time as many mentioned. I had a few campaigns started with positive roi but get to negative the very next day. To deal with this problem, some STM members have stressed the importance of frequent change of push ad creatives. But what is the mechanism behind it? Is it because of one creative is being spied by others and saturated the audience? Or is it because of the audience is small and are showed with the same creatives repeatedly?
2. For the effect of pre-lander: in casual dating vertical, I have tried to direct link with offers which already have their pre-landers and the conversion seemed to be quite ok, at times even out-perform the campaigns with my own landers. Do you also see this pattern in casual dating?
3. Is it important to blacklist sub-placements as we do for pop traffic in order to engage the more relevant audience? Some have said that they create new campaigns quite frequently, but do you bring over the blocked sub-placement id to the new campaigns?
4. Should I change vertical? I have used Anstrex to spy the offers running with push traffic and noticed that there are more sweepstake, casino and nutra offers. Does it make sense to change to a new vertical and higher payout offers?

Thanks in advance for your response.

---

Hey Man!

1) About the ROI - did you see a decrease of Clicks in general, or only in CR% - That's crucial to understand, because if your getting the same amount of clicks and imp's maybe it's the offer? Not sure tough. More details would be awesome

2) If the advertiser has their own prelander and it's a smartlink, would be a good idea to go with theirs, at least for the beginning. Most affiliates have synthesized what works to attract people and reach a converting lead. Maybe start with the advertisers landing page and further along the road create your own and do a split A/B test and see if you are able to create a better CR% with your own

3) Unlike pop, the Sub Id's have very little impact. A user subscribes from a specific site but once is registered he get's push traffic until he cancels his subscription. You need to blacklist age groups, with us it's publisher feeds, not sure how is it for megapush. Your blacklisting age groups, it's just easier to assess quality like that.

4) I wouldn't change the vertical all together, I would continue dating since it's something that has worked for our advertisers at least quite consistently. But your spy tool doesn't lie, Nutra, Casino and Sweeps is definitely verticals that works super well on push. So maybe trying some more verticals would be a good test, but i wouldn't stop the dating campaigns, but always have something lined up to test, always think of the next offer type you want to run

Cheers

---

Hello, let me answer your questions with my personal experience in mind

I think it has a lot to do with the particular algos and the way they serve the notifications ... connected with the often limited audience size. On top of that, the new subscribers get bombarded with multiple pushes so they start to ignore them. Which takes us back to the algo and how it mixes the new subscribers with the old ones.

Yup, sometimes the offers with built-in prelanders work better, but if I look at a bigger time frame, I still did better with my own LPs.

Yes, I do block placements with push too. And yes, with new campaigns (when the difference is really significant) I retest some of the already blocked placements. Especially when they have solid volumes.

Testing several verticals is pretty much a must. Based on the data I got from spytools, dating is not the BEST vertical for push in certain GEOs. And there are quite frequent quality problems with dating on push, so definitely don't limit yourself to just dating... CC sweeps can work pretty well on push for example.

---

Hey thanks for your reply

I saw a drop in clicks about 10%, which I guess is not that huge?

I use the "{pubfeed}. {source_subid}" format to create blacklist ( for example, 133828.722803). Am I doing it right? When you mentioned "publisher feeds", did you mean "{pubfeed}"? or "{pubfeed}. {source_subid}" ? I am quite interested to know more about this. The {pubfeed} is the one that categorize your inventory into different age groups? How frequently are you refreshing your inventory?

Sure! Will try some new verticals

---

matuloo, thanks for your reply.

Would it be better to test 5 traffic sources at the same time to engage different audiences and cut the non-performing sources?

When you say "difference is really significant", did you mean campaigns for very different verticals or campaigns with very different creatives?

Yes, I am going to test new verticals. I got a feeling that push traffic is something like an uncategorized facebook audience. Most of the traffic that I receive are real human. It makes sense to do more cps with real human traffic.
The only trouble is that there is no detailed interest targeting as facebook is providing. A user clicks on the Yes button on a certain webpage somewhat reflect his/her interest in a certain topic. It's only for us to filter out the more relevant audience. What I have been seeing in my tracking data is that there are placements with clicks to my landers but with 0% CTR and do not pass the google recaptcha test. This can be due to the users, after clicking the push ads, quickly closed the browser? It might indicate a very weak interest or tired of push ads?

---

Drop in clicks is Normal. When it's first going live it's new so users aren't familiar of your creative, so there is more chance they click it. After a day or 2 it's a bit lower, but makes sense, nothing to worry about.

Regarding the lists, Yep, That is how to build a black/whitelist on our platform

And regarding verticals, Yeah i agree, for instance the biggest offer type in India for us is Nutra offer, FR the best offer type is Sweepstakes. So you got tons of options

---

More sources means more time needed to optimize, more funds to spend on testing ... but also higher possible profit and faster optimization, due to the increased volume which would allow you to test more. If you can handle it, you can work with multiple sources for sure.

I meant when the performance of the new campaign would be way better. So let's say I blocked some placements with a campaign, because they were negative a bit. Later on, I might create a much better performing campaign (same or different vertical) so chances are it would perform good enough on some of the placements I blocked with the old campaign. So I try to retest those placements. To sum it up, it's ok to start a new campaign with a blacklist created with a previous campaign, but in many cases, it's a good idea to retest some of the blacklisted placements. Especially when they have solid volumes.

Yup, the targeting is not there, but the prices are still quite low so it gives us more space to test stuff

I didn't test google recaptcha with my campaigns, so not sure what could cause it. But it's not like push traffic is 100% bot free, some sellers mix in bots too, and it got more frequent lately. So could be bots too.

---

Home > Paid Traffic Sources > Push Traffic