Home > > Newbie Follow-Alongs

---

Ruriks Super Duper Pop/Sweeps Journey (Vortex) + Tips for some newbier newbies (29)

---

Hey Guys,

Here's my second follow along in the last 2 years for this Vertical. -

2 years ago I started this dream and couldn't do 10% of the things that I've learned now. -

So I've added in my insights to make this a kind of TUTORIAL/FOLLOW ALONG thread.




Why 2 years?


Conditions/Skills/Budget


I wasn't ready or had the resources to commit to do this properly, but I've been building the skills that I COULD afford to build slowly! So it hasn't really been two years of action taking. Just two years of a goal and movements toward it.



Following the advice of the experts to a tee is a MUST. (VORTEX, Mr.Payne. Mr BRAUN. ZENO, CAURMEN (RIP) - THANK YOU GUYS)

Learning some coding and getting technical skills are a real life saver. You'll need them!

I opted out because i new i couldn't make it happen unless i dedicate a good period of time to just hustling, at where money is not an issue.


NOW TODAY!

Currently i have about $2000 - $3000 USD to spend on ads per month.

Which i'm finding is actually difficult to do! - Ad spend needs to be coupled with enough actual work in order to spend that money! (If you are trying to spend it wisely, following the systematic approaches) -

There's no point in investing in landers that continuously fail, So the work keeps piling up. Work work work work work - rihana




What i've got down so far as far as skills:

What's working for me at the moment?





Enough Landers:

There's actually only about 10 - 15 uniquely different landers on Adplexity for iPhone it seems. I've ripped all the of the ones you've probably ripped. They might be adjusted for language/Model - but all in all it's limited.

#1 piece of advice as far as landing pages is "Learn To Make Your OWN" - Be original, creative!

To quote Charles NGO: "the most useful skill you should learn if you want to be an affiliate marketer is copywriting"

Here's a link a link to a really great interview with him talking about Specifically Sweeps/Pops/Newbie stuff: https://www.mobidea.com/academy/sweepstakes-offers/


What I Do as far as creating new landers:




OFFERS -


Traffic Sources




Notes on Bot traffic - I've tested 3 so far..



Current Bottle Necks/ Questions:



Current Goals



Last but not least - The Fruits of my work so far ... It seems i have a pretty promising lander out of the ones i've tested and deployed. If there weren't so many bots, I'm sure i could make this campaign work. Any tips guys?

STATS/ RESULTS - The Fruits of my labor so far... 1 lander that seems to be working well

https://drive.google.com/open?id=1zC...mrNXvs_2m3iWza

---

I'm new here. GREAT read. Thanks. Most stuff here is very instructional.

I'm still working out the Landing Page step before diving in..

---

Here's some more bot stats -

UPDATED DISCLAIMER : I wouldn't rely on this information. Also, I have tried Cuarmen's Bot Script and my personal conclusion is that it does not function properly. You can read about it here: https://stmforum.com/forum/showthrea...ampaigns/page4



3G - Switzerland - Popads - Average payout..

3G - Austria - Popads - Average payout..



3G - Germany - Popads - Average payout..


I'm going to do the math on the previous stats too:

- Germany - Adcash - 1.5x Average Bid..


3G - Germany - Popads - 1x Average Bid.. (Test a)



3G - Germany - Popads - 1.5x Average Bid..


3G - Germany - Popads - 2x Average Bid..


3G - Germany - PropellarAds- Average Bid..


3G - Germany - PropellarAds- x1.5 Average Bid..



This was the best run out of all of them



3G - Germany - Popads- Average Bid
----

---

Just got accepted into zeropark - But there's something i would like to test with pop ads as well - and that is whether or not setting a higher daily budget to your Ad campaigns trigger better traffic from the network.

This would be interesting to see.

---

Are you sure that you have setup the bot test correct? And that you are reading the results correct?

80% bots on 3G traffic seems exceptionally high.

---

yes - this was what i was wondering as well.

some points:

Reading the results : (Total amount of visits) - (Clicks on the parses bot offer) = Actual amount of visits

The click through rate of the Bot Offer --->Destination URL is always something like 70-90% - which should be significantly lower if a significant portion of the real traffic is being funnelled there too. since CTR for real traffic is 0.2-5%

What do you think?


Here's the instructions i followed for bot funnel set up:
( https://stmforum.com/forum/showthrea...live-campaigns)


1. Header code:

2. Actual link in body: Xxx.Voluum.com/click/1


3. End of page:



4. Campaign set up:

When I click on my landing page through my site link, and the campaign link and click through to xxx.voluumtrk.com/click /1 - It always leads me through the same link - the proper offer link or 404 page when not using camp link

I'll run a test without Cuarmen's bot script installed and see if that makes a difference - but this should be the correct set up (?)

IF the code IS correct, The only other conclusion that i could come up with is that this script does not work well with a large portion of devices that people use.

(I don't think that the rest of the code in the landing page has something to do with it either.. since the bot ctr of 70-90% is consistent through all 10 landing pages. Whether they are heavily coded w/ fancy scripts, or just a simple 1 page x 2 lines of text html lander)

---

Hey Rurik,

Good hustle, keep going and good luck!

---

So an FYI - the results of the parses script aren’t really concrete.

If you visit the thread associated with it and read Cuarmen’s advice, it’s supposed to collect the REAL visitors... it seems... at first.... then it doesn’t,. Until it does again... (read the thread, it’s confusing af ) - USE AT YOUR OWN RISK!!

Yet there’s a dispute whether it’s for the real visits or fake ones. Resulting in either extremely high real visitors or extremely high bots.

I’m going to stop using this script until they finally figure out how the hell to interpret and use the data above. Makes no sense whatsoever, however I encourage everyone to keep up with the script thread until we find a suitable solution... and what the hell it all means

My suspicion is that it works for some cases and not for others - based on operating system. The data seems to be collecting a portion of both bots and real visitors- which is the only way to make sense of 12% of either

It’s Most likely not such high or low quality traffic

---

Two links for other bot detection techniques and knowledge bombs from matuloo&charles ngo

http://www.matuloo.com/bot-traffic-h...o-do-about-it/

https://charlesngo.com/how-to-detect-bot-traffic/

---

Wow - thanks for sharing so many great tips! Especially all the tips on making good landers!

And a big WELCOME BACK! If you haven't given up on the idea of making affiliate marketing work after all that time, then it's meant to be. I wish you lots of fun on your journey!

Below are my 2 cents on some of what you said:

See this thread for some good networks for sweeps as suggested by Mr Green and stickupkid:

https://stmforum.com/forum/showthrea...-In-March-2018


Why not try another traffic source for that particular geo? Zeropark/Propeller are also quite big in terms of traffic volume.


You're using caurmen's method as illustrated here:

https://stmforum.com/forum/showthrea...live-campaigns

According to his tutorial, the CTR of "parses javascript" is the percentage of HUMAN traffic, not bot traffic!

Thus, if the CTR is 70%, that means 30% is bot traffic.


If you're deciding to cut placements that have for example 70% of bots or above, then you'd cut placements with a CTR of 30% or below.

No traffic network can survive with bot percentages of 80%+. That would be insane. And no advertiser/affiliate would be able to make enough conversions at that level of bot traffic by paying current bid prices.

Of the best-known pop sources, PopAds and Propeller already have the best-quality according to my personal experience. So, don't hesitate about the bot traffic anymore - just do your best to weed out the worst placements and work to make as much as the available profitable via testing landers and offers (with focus on testing offers).


I'm not seeing any statistically-significant results. Have you used the statistical calculator to check lander stats before cutting the inferior?

https://stmforum.com/forum/showthrea...Banners-Part-1

Once you DO have a good lander - test lots of offers! Talk to all your AMs and compile a list of the best offers. Look on Adplexity to see which offers have been getting considerable amounts of traffic volume consistently in the recent days (few people would keep a losing campaign running with the same or increasing levels of traffic for beyond a week). Do a search on the forums or start a new post to ask for recommendations.

I would suggest not to spend too much more time on figuring out which bot traffic detection method is the most accurate, because in the end, it would be near impossible to identify ALL bots. Caurmen's estimation is that his method can trap around 80% of bots. That's more than good enough for most purposes. We need to mind the 80/20 rule. You'd probably increase your bottom line by a lot more if you'd focus on testing and scaling instead of trying to achieve 10% more accuracy on bot tests. Just my opinion. Great detective work you did though!


Looking forward to seeing some campaign stats soon!




Amy

---

Alright -

So here goes the next test - will get the stats when i wake up.


Testing 16 offers and 17 landing pages - Iphone X - Zeropark - Germany (SOI and DOI offers)

GROUP A: ZERO PARK - Premium Tier 1 traffic - 17 landing pages X 8 Offers Campaign / Total budget for the first test = (17 x (average offer amount) x 2)

GROUP B: ZERO PARK - Standard Tier 1 traffic - 17 landing pages X 8 Offers Campaign / Total budget for the first test = (17 x (average offer amount) x 2)



Looking forward to seeing what the results are going to be.

half of the landing pages have been ripped/fixed - And have of them I made myself from scratch!

Will update in the morning.

---

HOLY SHIT!!

Kudos on testing so many things at once!

I know many people prefer to stick with small split-tests in order to reach statistical significance sooner. But to me, larger tests make more sense, because they're more efficient:

1)The more test candidates you include, the higher the probability that the range of performance would be wider.

By this I mean, when you're testing a lot of offers/landers at the same time, chances are there'd be a greater difference between the best and worst offer. This will enable you to reach statistical significance faster - saving time and money.

2)The fewer split-tests you conduct, the fewer candidates you'd need to retest. Every subsequent split-test we do, we need to include the winner from the last test - what we call the "control".

Let's say you were to test those 17 landing pages in batches of 5-6:

1st split-test - 6 landers, cut down to a winner
2nd split-test - another 6 landers + winner from last test = 7 landers
3rd split-test - the remaining 5 landers + winner from last test = 6 landers

So, instead of testing 17 landers, you'd effectively be testing 19 landers - which, again, will require more money and time.


Having said THAT....

Testing so many things at the same time will also require a strong heart, because it would take longer and more spend to cut down to a winner (compared to smaller tests - although in the end you'd be saving money and time as described above).

You got to fight the psychological urge to cut stuff before it reaches statistical significance. Sit on your hands if you have to!

Regarding test approach: Run until one of the offers shows good performance (2 conversions or above, with non-excessive spend), then pause all other offers, and just run that one offer to cut landers down to a winner.

(Try to avoid cutting landers based on conversions from multiple offers - one lottery conversion from each of several offers can throw off your stats. You can still choose to run all offers simultaneously and cut landers based on conversions from all offers - but to me that would be less accurate. Up to you though!)

After that, you can retest the other offers, and can also test new ones after that.


This is exciting! Looking forward to seeing some stats!




Amy

---

Update:

I woke up today and noticed that the campaign is doing really poorly - This was extremely discouraging at first.... But at the end of the day there was good new and bad news.


Good News/Bad News:

I made some mistakes and this made these landers perform worse than anything tested up to now... - 175$ that i might as well have lit on fire.

HOWEVER, upon finding this out. This is actually GREAT news. Because the bottom line is that I spend much more time on these landers and setting everything up.

If everything had been done correctly, this would have been realistically - $1000 worth of my time, and it also means that I didn't really try the campaign out! so I still have lots to test.

What were the mistakes that surely affected the campaign?


1. Call out scripts ... I put an extra " ; " where it wasn't necessary ie <script> document.write(getURLParameter('brand'));</script> I don't know whether this had a negative affect, but something i noticed.


2. TOKENS. Most of my landing pages NEEDED Tokens to function properly ---- As i found out, In order for the tokens to function properly.. You NEED to include the index.html link inside Voluum's lander set up - even if your website is set up to automatically read the pages HTML file, this is not good enough. (at least for my set up with AWS)

3. Symbols.... It's Very important to include <Meta charset=utf-8> inside your header if you don't want your foreign language to come out all jumbled up on the other side. also <html lang="{lang code}">

If you are having Foreign languages appear different when you load your page. this is probably it. (or another charset code, i don't know)


I should have waited a bit more and took more time too look over everything.

I've had a successful run last campaign as far as set up (no errors) and I assumed everything was set up correctly again. Wrong!


Lesson learned: Don't launch campaigns when you are tired. Don't rush through double checking. Quality work takes time. I should have stepped back - waited one more day, and QA'd my work.


Anyways - RE-launching again today.. and stats will come in tomorrow


Thanks for the tips Amy,

I'll adjust accordingly and implement them on this run!

---

If you don't do it already you should use an editor with syntax highlighting when you work on websites.
That way you can often easily spot if there is something wrong with your code.
For Windows you can use Notepad++ or Sublime.

---

Wow! I'm glad you eventually noticed these things!

twinaxe made a great suggestion - Notepad++ and Sublime Text can save your life - especially for people like me who is careless.

Also - adding the following 2 steps to your workflow will minimize this type of thing from happening:

1)Rip and fix up and test more landers in the beginning. If you only test 1 lander and there ends up being something wrong with it, you're screwed.

2)Verify that the landers both DISPLAY and FUNCTION as intended:

To verify how they display in various browsers and devices, check them using browserstack.

To verify functionality, manually browse to the lander and click around, click back etc., to make sure everything is working as intended - do this on a mobile device if that's the traffic you're targeting.

The language display thing: If you're using a script that will automatically detect the user's language and display the corresponding language, you can quickly test this by changing your browser's display language to see how your lander displays that language on your lander.

It IS frustrating that such a small oversight can result in such a big disaster. One way to guard against that would be to start a checklist to go over before launching each campaign. But then all the extra checks may result in greater cost in terms of time, than the mistakes you would make without using such a checklist. It all depends on how careless you tend to be, and whether or not you feel such a checklist would be useful.




Amy

---

Alright, so after a few weeks of setting up, joining new networks, fixing and making a large collection of sweeps landers, I'm ready for another round.


Current tests -

Just set up a singapore campaign inside Propellar ads and Zeropark - budget 25$ per day (x2 offer average payout per day)

Test 1: 15 landing pages x 12 offers - Propellar ads

Test 2: 15 landing pages x 12 offers - Zeropark Premium Tier 2

Test 3: 15 landing pages x 12 offers - Zeropark Standard Tier 2

The game plan is to run these until one of the offers hits good results (Multiple conversions/Non excessive spend) - Will see what happens.

Propellar ads camp is up and running right now, and waiting to do some quick bot blacklists on zeropark before starting that tomorrow.



UPCOMING:

Polish sweeps campaign - Pop ads

Spanish Sweeps campaign - TBD

---

Unless you are certain that these 12 offers you are pushing are doing really really well, I would go for a lower number of lander/offer bundles to split-test.
Also another important thing is to already have some data from the traffic sources you are running when it comes to placements quality and bids for the targeted geo. Info which you may have already collected previously with a successful campaign (lander/offer).

If Singapore is a new geo for you then check out Monetizer to see if you can get some info about possible potential offers there. On top of that, if you see some potential offer there, send like 20%-30% of your traffic direct-linked to a smartlink. This is a really practical way to see if your offer performs or not.

---

It's a new GEO - not sure what you mean by the monetizer link. I'm doing only iphone X/8/7 sweeps right now - Changing Geos over vertical.

Sounds useful though, VPN to singapore to see what kind of offers pop up - I'll make a note of that

As far as the offer/ lander split, what would you recommend? - I'll set that up right now and run a few tests.

---

Okay -

I further adjusted my overall campaign to try out a few more different approaches -

the large scale campaign from first look is a much bigger money dump, and I can see why smaller digestible chunks are suggested - So i'm structuring different sized tests for these offers to get a feel for the different approaches.


Current offer test:


Test 1 - Singapore : 15 landing pages x 12 offers - Propellar ads - Going to keep this going and cap at around $100-200


Test 2 - Singapore : Landers #1-10 x Offers #1-5 - Zeropark Premium Tier 2

Test 3 - Singapore : Landers #1-10 x Offers #1-5 - Zeropark Standard Tier 2

- Total offers to test are 12 for SG - however going to digest them ie #1-5, #6-10, #11-15 (as more offers are getting accepted by aff networks)


Test 4 - Poland: Landing pages #1 - 5 x Offers #1-3 - Zeropark Premium Tier 2

Test 5 - Poland: Landing pages #6 - 10 x Offers #4-6 - Zeropark Premium Tier 2


Something that's really helped recently: I created a spreadsheet of about 300 iphone x/8/7 offers across all my networks - and can now conveniently pick and choose 5 - 20 offer groups per geo to test, and since I've done a ton of leg work with landing pages.. it's just a matter or translation, adding more here and there. Adjusting. Not a big deal - but the reach to all these different offers is what I'm really excited about

it's tough finding matching sweeps offers in non-cliche geos. So the way that I went about it, was taking a week to just compile offers and add as many networks as possible. Now it's relatively easy to blast through offers

---

Holy moly! that's alot of offers to test Just curious, how many networks are you currently signed up with to get these 300 offers?

---

I'm signed up to like maybe 20-30 -

The majority of the offers are on Mundomedia, Clickdealer, Toroadvertising, Peerfly, Wewe, Addiliate and to a lesser several other networks - Between the ones I named - there's a TON of offers..

However.. The truth is that when you get specific, 300 boils down to "Just enough" to justify doing the leg work (assuming it's nichey foreign geo).. I'm still joining networks and hoping to get with Affiliaxe and advidi, who also have a bunch of sweeps offers.

---

Using a Monetizer link you'll be able to send traffic to an offer that is already working there and split test your offers against that. Also the most interesting part is that you'll be able to see how this geo is performing.

Browse through the tabs for further info

Beware of the results you get from a VPN connection. Most networks and offer owners are redirecting proxy traffic elsewhere. Singapore is available on adplexity so take a look over there. Coupled with monetizer's info it should be easier to focus on some potential offers right from the beginning.

I would say at max 3-5 landers and around 3 offers. The more landers/offers you add the more budget a specific campaign will need to be properly tested.

---

thanks Platinum - I'll set that up and include that in my next tests.

Here's little write up from my recent tests:

Test 1 - Singapore : 15 landing pages x 12 offers - (Note. Forgot to Click mobile only, so traffic is 3g + wifi :*(

Screenshot (Offers):
https://www.dropbox.com/s/8m1vflnjqg...rs%20.png?dl=0


Screenshot (Landers)
https://www.dropbox.com/s/2zvh6vwx4t...nders.png?dl=0

Test 1 b -After I realized I turned forgot to make it mobile only, I set up the same Camp and ran the test again:


Screenshot (Offers):
https://www.dropbox.com/s/w3b15jigtp...ffers.png?dl=0

Screenshot (landers):

https://www.dropbox.com/s/9ujww40dph...nders.png?dl=0


Other tests just canned, and I didn't proceed with the wide net for Zeropark, since I wanted to boil down these variables first.

Obviously, This 12 offer x 15 lander campaign should have spent over a thousand to properly run, which i'm not going to do...BUT

I did get some offers and landers which seem to have performed close to -50%, While at the same time being very diluted by all the other crappy landers and offers.


I took my 4 best landers and 3 best offers - All of which received multiple conversions - and set up a another 3 campaigns to test them .


1. 3G - Top 3 offers x Top 4 Landers - Average/high bid - w/blacklist of absolute worst CTR placements
2. wifi - Top 3 offers x Top 4 landers - Average bid
3. 3g - Top 3 Offers x Top 4 Landers Whitelisted Zones Only - High Bid


The zero park polish Campaigns Tanked so I'll just leave that for now.

Test 4 - Poland: Landing pages #1 - 5 x Offers #1-3 - Zeropark Premium Tier 2 (-100%)

Test 5 - Poland: Landing pages #6 - 10 x Offers #4-6 - Zeropark Premium Tier 2 (-100%)



Waiting to see how these offers perform and what will come out of it - and move on to testing more landers, or more offers, depending on the results.

---

Ok - So the final conclusion to all these recent tests was that I should have cut all of them a lot sooner -

The majority of the conversions in the 15 x 12 came from 1 offer and 1 landing page. This seemed to me like a good sign and I had some relatively ok performing offers in total - I took the top 3 offers and top 4 landers and got less conversions than the initial 15x12 campaigns...

A bit frustrating, but overall I feel that I should have cut more ruthlessly earlier on and not have even tested such a large batch.

Lesson learned.


What Now? - Focus on testing LOTS of offers. 6 landers x 3 offers at a time. Cutting offers at x10 payout ....Ice cold cutting....

Update soon

---

Good work @Rurik dedication! I have gave up on this tutorial after hardly any conversions with direct linking. Creating landing pages, getting Voluum tracking to work and lots of testing looked way to much of a ball ache for me.

---

Kudos on the mass-testing!!

Being willing to test so much stuff is definitely an important breakthrough in mindset.

Subsequently learning to be more efficient in mass-testing will stand you in good stead. But know that it's easier to tweak your approach to be more cost-effective, than to get past the psychological block that keeps most newbies from doing mass-testing. I've had to go through a lot of trial and error initially as well. Look at this as a worthwhile investment in education and experience, and not a loss.


Question: Are you testing the same landers and offers on all 3? If so, it's not necessary - just test using one campaign, find the winner, then scale. No point in wasting money on testing the same stuff.


Actually, I don't see any big problems with testing 15 landers at once - chances are very good that the winning lander will be a decent one (whereas if you're only testing a couple of landers, even the winner among them may be a dud).

What you can do next time, is run the offers until one of them makes 2 conversions for example, then pause all other offers and use JUST that offer to generate conversions, and use those conversions to cut landers down to a winner. This is the methodology I've outlined in the 40-day tutorial.

With this method, you'd be using a promising offer to cut landers, which will save on budget. (And, if none of the offers are converting, you'll probably want to either replace with a new batch of offers, or test another geo altogether - again, this was covered in the tutorial, so you can refer to that for details.)

Then, once you've cut landers down to a winner (using the stats calculator), you can mass-test offers - including the ones you've temporarily paused in the beginning, and any additional new offers as you like.

Including more test candidates in a split-test has pros and cons. Pros: Chances are better that you'll end up with a winner that performs well. Cons: It will require a bigger budget to test more things at once. (Although, it should actually be CHEAPER overall to test everything at the same time vs. in batches - because when you test in batches, each subsequent split-test you do will need to include the winner from the last batch, such that you'd be retesting some candidates.)

So, it would be up to you to decide how many candidates to split-test at one go. There are no rights/wrongs.

One thing I really want to stress, is the importance of running split-tests to statistical significance. If you don't have the patience or the heart to fork up a big initial budget to split-test a ton of landers or offers, then set up smaller tests, but make sure you cut offers/landers as they reach statistical significance!

Cutting based on emotions will eventually drive you insane. Pop traffic already gives volatile performance. Cutting offers/landers on a whim will only give you even MORE unpredictable results.

The only exception to that rule, is when you have an entire batch of landers/offers that don't seem to be converting, OR if the CTR/CR is so low that you feel it would be unlikely for the campaign to ever be profitable. In that case, you can justify cutting the under-performing lander/offer without waiting for statistical significance. (You may even want to consider replacing the entire batch of offers/landers, or ditching the geo+vertical altogether.)

Heeding the advice above should help you to drastically reduce testing costs. Like I've said at the start of this post, being willing to do mass-testing is major progress in itself. Don't get discouraged by the results - as your test approach can be adjusted.

Looking ahead: If you have the budget (it looks like you may!), moving into Native can give you bigger and easier profits. Native is similar to pop so would make the perfect next step, but will involve having to lose a lot more money at the onset while you cut placements to make the campaign green, but based on what I've heard from native experts, campaigns stay profitable for quite a bit longer than for pop, and profits are higher.

It would be necessary to get more experience with pop first though - it's so much cheaper to learn the same lessons with pop than with native!





Amy

---

Thanks Vortex -

I feel that my approach is not systematic enough indeed.

Statistical significance is indeed my weak point - I'll make sure these next camps address your feedback

---

Thank you for the valuable information you share, it's very interesting

I agree with others, in your tests you have to take into account that x10-15 offers must be multiplied by the number of traffic variables such as different operators, browsers, OS, etc ...

The available traffic in relation to your bid will also influence the data. The other important point that you have pointed is Motivation, holding long-term tests to have the volume without having the opportunity to really benefit from returns can be discouraging and motivation is extremely important to hold your concentration

---

Valid points indeed.

Targeting geos that have higher traffic will allow split-tests to reach statistical significance sooner, which will help one stay motivated. On the other hand though, traffic coming in too fast can make it easier to waste money, when you over-run split-test candidates way past the point of stat sig.

Bigger split-tests have the potential of achieving greater improvements in ROI, and also works out to be cheaper (than to separate test candidates into several smaller tests - as you'd be retesting the winner from the previous round each time). But it will also take longer to finish and require a bigger budget, which I guess may not be the best thing for motivation when you're a newbie.

So - pros and cons. Go with what makes you feel the most excitement and/or the least resistance.




Amy

---

Home > > Newbie Follow-Alongs