Vortex' 40-Day Tutorial 2019 (Voluum) - newbie choosing action over info overload (17)

First of all - huge thanks to @vortex for creating this amazing and comprehensive guide! I appreciate the effort you've put into this, huge thing.

Also many thanks to other members for creating follow-alongs. I'm new to the forum so only have seen those written by: @GregMorrison, @wakeboarder and @Jafar so far - very inspiring, and the reason why I decided to write my follow-along too. Maybe I'll manage to inspire someone one day too

So I'm starting now - going to apply to affiliate networks, sign up for traffic networks, get the tracker and a domain.

Day 1: Applying to various networks

I applied to several networks:

• Mobidea - instant approval.
• Clickdealer - I had to answer several questions on Skype and over the phone, after which I was approved. The AM was very nice and helpful. Gave me multiple offer recommendations good for a newbie and for use in the Tutorial.
• Gotzha - no answer so far.
• Advidi - no answer so far.
• MaxBounty - approved after a short phone interview.

• Previous experience - I had some with Pay-per-call and SEO. It's worth to create screenshots before applying, upload them to a service like imgur.com and have links ready. Everything that shows that you are doing something in the IM space is relevant and helpful. I think the worst is to not have anything to say.
• Reasons why you want to work with media buying and CPA, move from Pay-per-call to CPA - it's good to have something other than "to make more money" here. I talked about problems in the SEO, more control over the traffic in comparison to SEO's volatility, scaling potential.
• Campaign links from other networks - here I was lucky. Because Mobidea approves immediately I went ahead and set up a campaign in Voluum right away, just to get my feet wet asap. Then when this question came up during Clickdealer interview I had something to show. So applying to Mobidea first and preparing something there may be a good strategy. Even if you don't get such question during your interviews you can bring it up yourself - and that way show initiative and prove that you're already doing something.

• PropellerAds - here I had to deposit minimum $100 and they take $4 service fee for using a payment card. Couldn't use PayPal for the first payment.

• PopAds - here I could use PayPal. The initial deposit minimum is just $10. And there was no service fee.

Days 2-13: Picking offers; Configuring Voluum; Setting up first campaigns

I've spent some time familiarizing myself with the interfaces of the networks I was approved to. Chose some to run the initial direct linking from several GEOs. That way I could compare prices and available traffic between GEOs. Some were very surprising actually, so before starting with any GEO it's worth having a look at its CPM (besides the available traffic).

Choosing offers was hard. But my main goal though was to set up my initial campaigns ASAP, so I just picked some without agonizing too much, and I managed to get campaigns running. Got some conversions (ROI negative of course). The feeling of having some campaigns running and all data recording properly in Voluum (cost and conversions pretty accurate) was great and very motivating to continue.

In ClickDealer's interface of a given offer it's very easy to test the offer (enabling testing mode) and send a test conversion to verify that all is set up correctly (the little wrench icon below the screenshot). It would be cool to be able to delete this conversion after the test in Voluum but looks like it's not possible.

I've set up Voluum with custom domain - that was pretty straightforward with the instructions in the guide.

For traffic source I've used PropellerAds and PopAds. Although had to wait for a long time before PopAds campaigns were approved. Looks like PropellerAds is great for testing as campaigns there are approved within minutes.

Day 14-15: Landing Pages

As stated in the guide running campaigns without landing pages doesn't make much sense. Also my manager from ClickDealer immediately guessed that I don't use any LPs just looking at my stats. So I stopped all the initial campaigns and started looking into landing pages. Investigated the repository here on STM. Signed up for AdPlexity and learned to use the interface, particularly the seach modes. To be honest I expected to find more variety of LPs there, looks like there's a lot of repetitive stuff, but definitely enough to get inspiration. Looks like the best approach will be to collect/create English versions and outsource translating them to other languages, as suggested in the guide.

Day 16: Picking a Vertical and Applying to Offers

Looks like finding offers is my weakest point at the moment. To keep progressing with the guide I'm going to stick with Sweepstakes/Leadgen, SOI for now. My manager at ClickDealer gave me enough offers to get my feet wet with using landing pages.

I felt like I need more understanding of verticals and offer flows, and these 2 guides were very helpful for Pop and Sweepstakes:
Sweepstakes 101 Guide by @vortex
The ultimate guide about POP-traffic! by @mrbraun
(thank you guys, these guides are AWESOME!)

Also having a VPN is very helpful to be able to see the offer pages, which usually redirect to completly different offers when accessed from another GEO. For now I have ExpressVPN till I learn how to use Luminati...

Days 17-26: Downloading landers from AdPlexity; CDN; Fixing up landers & optimizing for speed

Downloading landers is easy, but choosing which one to download is more tricky. I sort by "last seen" and combine that with "days running" to figure out which may be the most tested or optimized version from all the available.

It's worth to download several identically looking anyway - some of them are hardly optimized, some are optimized to the point that there's just a single html file - everything is inside it - CSS, JS and images as "data:". Comparing versions also gives some inspiration on what can be done in the process of optimizing the lander. The Guide (+ additional links it provides) explain that in detail and some of these landers are practical examples.

Setting up CDN wasn't too hard, although the AWS interface is a bit annoying. But no major problems with the instructions in the guide and additional hints from @Jafar's follow-along (thanks!). I find using "aws" on command line most convenient for uploading/syncing the LP files.

Now I'm looking for landers, fixing them, optimizing. I think I'll use optimized and not optimized versions during campaign testing to see differences.

GTmetrix.com is great, but often claims to be overloaded and I need to wait quite a bit for the page to get tested. So I'm starting with just what's available in Firefox: Inspect element > Network. Then I upload the before and after versions and continue using GTmetrix.com. I realize that what I see in "Network" is just local, but gives a good clue on how many calls there are and what gets loaded.

I managed to get my first landing page optimized to the point that instead of 20 elements (before) it's just index.html (after) looking exactly the same and loading much faster (about 100% faster).

I also noticed that some landing pages include jquery library (which is a big file), while they don't use any jquery inside. Worth a quick test (disable it and reload the page) to find out if it's possible to eliminate it.

We've got the highest payouts on the market for Sweepstakes/Leadgen, SOI. Hit me up, if you want to increase your ROI%.
https://affiliates.gasmobi.com/register

Ieva Gasmobi
Skype live:ieva.m_1
Telegram @ievamihelsone

Thank you, Ieva! I'll contact you on Skype.

I have my first bigger puzzle regarding reading test results. I tested 1 offer with 5 versions of landing pages.

For the first 4 hours the first one was a clear winner (up to 20 conversions in total). It had ROI -35% consistently during that time, while all the others had -70% to -90%.

I left the campaign running for the night and in the morning I was not sure about the winner any more... The diffence in ROI seemed not big enough.

Here's the screenshot from Voluum:


Now, if I look at the ROI:

1) survey 4q                 : -73% << best, but not a big difference
2) grey wheel + price        : -78%
3) grey wheel - price + OPT  : -79%
4) grey wheel - price        : -80%
5) green wheel               : -84%

1) survey 4q                 : 0.55 << best, with a more significant difference
2) grey wheel + price        : 0.64
3) grey wheel - price + OPT  : 0.66
4) grey wheel - price        : 0.72
5) green wheel               : 0.89

1) survey 4q                 : 56% << best, with a significant difference
2) grey wheel + price        : 11%
3) grey wheel - price + OPT  : 26%
4) grey wheel - price        :  7%
5) green wheel               : (not enough space to calculate this)

1) survey 4q                 : 0.0009-0.0018 << best
2) grey wheel + price        : 0.0006-0.0015
3) grey wheel - price + OPT  : 0.0006-0.0017
4) grey wheel - price        : 0.0006-0.0014
5) green wheel               : 0.0006-0.0011

Hello

Originally Posted by afkuni

Setting up CDN wasn't too hard, although the AWS interface is a bit annoying. But no major problems with the instructions in the guide and additional hints from @Jafar's follow-along (thanks!). I find using "aws" on command line most convenient for uploading/syncing the LP files.

Originally Posted by matuloo

I'm looking at the numbers and there are quite big differences in the number of visits per Lander, why is that? [...]

And the other traffic source - PopCash - shows a lot of suspicious visits and clicks, high cost and hardly any conversions. So surely influenced the final (mixed) results a lot:

Although in this post:
https://stmforum.com/forum/showthrea...of-Pop-Traffic
it's recommended to test multiple bids from the start:
"DO test multiple bids from the start. 3-5 camps at staggered bids to see which one give the best ROI."
Not sure if it's the same what I did here though?...

Thank you!

Yes, I created separate campaigns. But then tried to determine the best landing page based on data from all the campaigns.

So the main takeaway is to analyze the data separately per campaign.

Now, looking at the screenshot PropellerAds LOW/MEDIUM/HIGH bids - https://i.imgur.com/e92YTjY.jpg - what would be the best way to proceed? Which campaign is the best one - would that be HIGH campaign with LP2A because it has 9 conversions, and also the lowest cost/conversion (0.24 vs 0.30 LOW and 0.32 MEDIUM for the best ROIs respectively)? And try variations of this landing page with different copy, etc?

Or with such bad ROIs should I:
1) create new LOW/MEDIUM/HIGH campaigns
2) eliminate LP2A landing page which was the worst in each campaign
3) add several new, different landing pages
and run the test again?

I took some time to read the remaining lessons. It was very helpful and answered many doubts and questions I've gathered in my initial campaigns. Definitely worth reading for me before running next campaigns, as I now see how many newbie mistakes I made in just a few days of running campaigns...

I now believe it was a mistake to run multiple campaigns in parallel the last time as my conversions were diluted between camps. And then drawing any conclusions from digging into variables was hard. Each camp had just a handful of conversions from different traffic sources and bids. Next time I'll stick to one traffic source and focus on testing first and foremost, not "optimizing". Once I find (through testing) a promising combination of lander/offer I'll start optimizing. And only if I get at least close to green I'll use other traffic sources (i.e.: scale).

I also applied to a few more networks and am now accepted to five which gives me enough offers to choose from.

And I practiced fixing and optimizing landers too, which was taking a lot of time. Started writing some code that automatically replaces all images in the html file with base64 data. Hope it will be worth sharing here once I improve it, for now it's just a quick prototype, but managed to replace all images without spoiling the lander.

Originally Posted by afkuni

Thank you!

Yes, I created separate campaigns. But then tried to determine the best landing page based on data from all the campaigns.

So the main takeaway is to analyze the data separately per campaign.

Now, looking at the screenshot PropellerAds LOW/MEDIUM/HIGH bids - https://i.imgur.com/e92YTjY.jpg - what would be the best way to proceed? Which campaign is the best one - would that be HIGH campaign with LP2A because it has 9 conversions, and also the lowest cost/conversion (0.24 vs 0.30 LOW and 0.32 MEDIUM for the best ROIs respectively)? And try variations of this landing page with different copy, etc?

Or with such bad ROIs should I:
1) create new LOW/MEDIUM/HIGH campaigns
2) eliminate LP2A landing page which was the worst in each campaign
3) add several new, different landing pages
and run the test again?