Home > The Newbie Zone >

---

My $1 Guide Experiment (14)

---

Hi Everyone!

I joined the forum quite a while ego, but never really took any serious action.
When i got the email to the guide, i decided to kick myself in the a$$ and start making moves.

I started last week, on Friday 06/14. I received from Flor an offer list with payouts and average conversion rates.
It was a little different than the list in the guide, but i was able to convert the data as i needed, chose a few geos, followed the guide and started the campaign.

Popads took quite a while to actually approve the campaign and i think it was actually started the following day.

After the campaign depleted, i went over to download the results so i can continue with the guide, but no revenue was recorded in PopAds :-(
Turns out i missed the second line in the pixel view at mobipium needing to add part of `tid,pay`.

But the data was still there, just had to go one by one matching data between popads & mobipum.
Surprisingly, the PNL was actually +$0.21 but the data showed that 95% of the traffic went to a single Geo


I felt that such amount of traffic going to a single geo does not let me buy data correctly especially since the payouts across each geo were completely different.
It started rising questions such as:
1. is my bid too low for specific geos but might be too high for other geos?
2. If i cut a publisher from geoA, does it mean it does not work for geoB?
3. If one geo is eating all my budget, can i truly make informed decisions to cut other geos because they are not preforming?

Basically, a lot of if/thans started filling up my head, there is so much data to consider, and so many way to interpret it... I think this is probably the hardest part of AM, making decisions and trusting they are the correct ones when so many variables can make a difference.

So to solve this issue, i setup a campaign in FunnelFlux, and created a sub funnel for each GEO. I then went ahead and created a campaign for each geo in popAds and linked them to the funnels.
This would allow me to allocate budget for each geo, control the bids of each geo, and also remove publishers based on the geo performance.
It took another whole day until the campaigns were authorized (I think PopAds have no one going over campaigns at night time) and when i cam in the morning to work, i started all of them.

The traffic was insane, it is quite a rush seeing the numbers going up and conversions registering. Unfortunately all this was happening while i was at work, so i couldn't dedicate much time for it. I did cut some publishers, and tried to top off the profitable geos, but couldn't really dig into the data as i wanted to.



At night, i was finally reviewing the data and made some conclusion while adding to my pool of questions in my head.

First, i noticed that a static bid of 0.0004 across all geos did not make sense.

Brazil for example, brought a lot of conversions, but only generated 0.01 per conversion making each conversion very expensive.
While other geos, didn't get much traffic to even decide which publisher to cut.

Is my bid too low for some geos? Maybe popads doesn't have much inventory for these countries? Maybe the time of day makes a difference?

The next day i started playing with the bids for each geo and looking the impact of the traffic levels & conversions correlating to the bid.
I didn't implemented much methodology in my bidding, and decided i have to make decisions based on concrete data and hypothesis rather than gut feeling.
(All this was done while i am cutting off publishers that are bring no conversions, or very low conversions)

So i went back to the excel file Flor had gave me, and analyzed the data. I decided to use her reported conversion rate (half of it) & the offer payouts as a base line to figure out what to bid for each geo, i ended up with something like this:



This gave me a starting point, not saying its a good one, since it is not my data and i don't know how it relates to the traffic that produced it, but at least i had a base line to start with.
I did one last analysis of publishers and while topping off the campaigns i excluded more publishers.
The next day was still in the red, but i was finally able to get decent traffic for most of the geos (some like the US still need higher bids which can't produce positive results, at least in my case because the payout is too low)
I was able to find some geos that showed real promise, some publishers in these geos produced 4-5 conversions per 100 visitors, but not enough traffic from them to make a serious dent in the PNL.

The following few days i was really swamped with work, finishing my day around 10PM, so i didn't top any of the campaigns and stopped sending traffic.
Yesterday (friday) i finally finished early and spent a few hours cutting publishers from each geo while topping off campaigns.

PNL was still in the red, using Flor data as the base line for bidding produced too expensive bids, but at this stage, i already knew which were my strong geos, and had enough data about conversion rates from my own purchasing.
I implemented the same excel sheet, this time using my own data for conversion rates, that gave me a much more accurate bidding strategy.
Updated campaigns once more, removed publishers, implemented new bids and went to sleep (it was already 12am when i finished)

This morning i finally woke up to a green PNL :-)



Now, its not even a cup of coffee here in NY, and the amount i already spent in total on campaign will probably make it quite impossible to turn into an overall positive one, but it was worth every penny!

I learned a lot from this exercise and discovered i like this kind of work. Got a lot of questions answered, and many more to find answers for.

I don't think I'm going to continue with this specific campaign much longer, since after cutting publishers, the winning geos don't get much volume. But ill continue to tinker with it, since the more i can raise my conversion rates, the higher i can place my bids and hopefully see more traffic.

Here are a couple questions i am still struggling with:

1. mobipium can pass back additional data in the callback, such as offer_id, carrier etc... i did set it up in the URL, but i can't find a way to view this data in funnel flux. I really wanted to see which of the offers in the smart link are actually preforming for me. Anyone know how to setup additional tracking fields for an offer source similar to how we do with a traffic source?

2. Flor asked me to stop sending traffic to some geos , she didn't really specify much (Due to low quality) and one of the GEOs was the BH which was not in my Geo list.
Is she talking about my conversion rates? If so, why does she care? We are paying for the traffic, so conversion rates are meaningless data for them.

Is it the offer owner? I signed up when testing the link to the offer, and got 5 notifications of girls telling me they love me in the first hour so, so i am not sure what kind of performance the offer owner is expecting. Also, the geos she asked me to pause had very low payouts, so it seems they adjusted their payment to take into consideration these low preforming geos by paying lower for them.

Either way, i didn't like that, since we are spending good money trying to make a geo work, and as it becomes profitable, i had to stop it.

I am thinking of taking this campaign and start collecting my own subscribers for push notifications, i think it would give me more wiggle room with bidding and budgets, and would allow me to learn a new vertical everyone seems to be talking about at the moment.

Amy, thanks for the detailed guide and the kick in the a$$ to take some action! :-) It was an exciting experiment and you guide was super detailed and easy to follow.

Yossi

---

And i just read the quality post you made :-) makes sense, the offer owner is doing his publisher cutting just like us. And i guess as long as he is paying for his testing (ie our conversions) its good business

---

Sorry I somehow forgot to reply to this!

This was one of the reasons why I suggested to set a low daily budget on the first day. I should have explained about this in the guide.

To answer your questions...

1. Yes! Every geo's available traffic volume and level of competition is different. So when you place the same bid for multiple geos, you will see different levels of traffic.

2. Yes! This is why I suggested targeting each geo individually once you know which geos are promising enough to be worth the time in doing so - so you could optimize each geo further. Targeting multiple geos in the same campaign is only good for a quick initial test.

3. As you may have suspected, the answer is no. What you can do is this:

-Run the initial $3 budget as suggested in the guide.

-Check stats to see which geos are "over-represented" - remove them from targeting, and start new campaigns for the promising ones.

-For the original campaign, add more budget (and increase the daily budget accordingly if this is still the same day you spent the initial $3, or else the traffic won't resume) and unpause to run more traffic, to give the rest of the geos a chance to perform.


Based on your intelligent questions, I know you've answered most of them in your head even before you asked them. Still good to get some validation though!

Tip: I like to ask questions, then devise tests to find answers. If you do this enough, you'll gain insights that can really give you a leg up over the competition.


Completely agree with you! But the harder something is, the more it can pay off when we figure it out.


Or you could just set up a separate campaign for each geo yes! Would just require more work upfront.


Your options if you're too busy to check stats during the day:

1)Add more budget in the smallest increments possible - which is $2.50 on popads.

2)Target multiple geos using the same campaign. That way you don't spend much on each geo (especially if you also implement the first option).

3)Set for traffic to run "throttled" - this is a setting in popads and will spread traffic throughout the day instead of giving you as much traffic as you can get until your budget is exhausted.

4)Just run at night time when you have time to babysit campaigns - which may not be the best method because you won't know which hours are the best-converting from the start, and this may be different for each geo.


This is a good trick for figuring out how much to bid - once you have some initial data with which to calculate your conversion rate.

However, do keep in mind that:

1)When you bid differently, the conversion rate will change. (This is because the best-converting placements tend to be more expensive due to increased competition - everyone will cut bad placements and increase their bid to get more traffic from the remaining, good placements, which will drive up the price. So if you bid higher you'll get more traffic from the good placements. The opposite is true also - if you bid low you'll get traffic from the bad placements. Thus the conversion rate change.)

2)Each placement can and will have a different conversion rate.

So, if you can implement what you did to each of the biggest placements instead of for the campaign, you can fine-tune each placement for profit. Sad part though is that very few traffic networks will allow us to set a different bid per placement. Zeropark being one of them.

In the end, it would be good to test different bid levels. To do this properly, we should test at least 3 different bids to start, then test in-between bids to fine-tune as necessary. I know people that test 5 or even 10 bids to see which one does best. This kind of testing can really pay off for larger geos.


Nice! So we think alike.


Congrats on the green!

You're right - it's not the amount that matters, since you're still in the learning stage. It's the amount of learning, and knowing you've figured out so many things along the way that have allowed you to optimize the campaign to profits, one step at a time.

And you may be right about not continuing with the campaigns. Everyone that has bought the guide are running the exact same thing on the same traffic source - that never bodes well. I knew that would happen, but the purpose of the guide is to help aspiring affiliates make their first conversion - the fact that many people have also found profits is just a bonus.

You can try to scale it to other traffic networks to get more volume and avoid the competition. I'm setting up some tests and will report back soon - but no need to wait on me of course, if you feel like doing some testing on your own!


Unfortunately funnelflux's postback only has the 3 tokens:

?flux_payout={pay}&flux_txid=&flux_hid={tid}

Since we're not usually using flux_txid for anything, you can use that to pass ONE of the Mobipium tokens - for example the offer ID you mentioned:

?flux_payout={pay}&flux_txid={oid}&flux_hid={tid}


You've already seen this post - but I'll elaborate a little bit anyway:

https://stmforum.com/forum/showthrea...um-and-Quality

Nobody likes it when these things happen. But really, the fact that Mobipium honors all these payouts up to the point they tell us to stop running certain geos, is already very fortunate. Not every network will do that (although the ARE obligated to).

There's no way for them to pre-determine the conversion rate, because every traffic source has different traffic quality. Even among pop traffic sources, quality can vary by a lot.

AND - even when 2 affiliates run traffic on the same traffic network from the same geo, quality can STILL differ due to the bid and targeting (placements, devices, etc. etc.).

So the only way for them to determine quality, is to allow each affiliate to run initial traffic to collect some conversions, then allow the offer owner to monetize to see results, and THEN come back to tell us whether they're getting the bang for their buck or not.

Obviously, if they don't make their money back, they can't continue to pay the aff network (and in turn, us affiliates) for the same traffic. It's not like once they agree on an offer payout they're obligated to honor that forever. (The affiliate network though IS usually obligated to pay affiliates for conversions that have already been made, unless they have proof that the traffic is fraudulent. However, every case is different, and every aff network has different TOS, which is one reason why some affiliates don't get paid.)


By all means!!

You must have seen these threads?

https://stmforum.com/forum/showthrea...on-Subscribers

https://stmforum.com/forum/showthrea...-with-Push-Ads


Thanks Yossi! Thank YOU for giving the guide a try and for all the insight you've shared by starting this thread. I'm glad you've found it helpful, and look forward to seeing you move onto bigger and better campaigns!



Amy

---

Hey guys!

So a quick update on my progress with this campaign

First of all, @Amy, thank you so much for your detailed feedback and encouragement!

So i took some of Amy's suggestion and did the following:
Like i mentioned in my first post, i created a separate campaign for each geo and tested different bidding based on the payout and actual conversion.
In excel i did something like the following:

Payout, ConversionRate, Expected Conversion Per Mill, Expected Revenue Per Mill (Conversions * Payout), Expected Earnings per Entrance, Bid (EPe * 0.55)

This allowed me to test the Geo with a more accurate bid and find out which geos actually gives me decent traffic and conversions all while cutting off publishers that don't work for that geo.

The goal was not to be profitable right away, but discover the potential of each geo.

The following day, i went ahead, and created 4 types of campaigns for each geo separating publishers by conversions rate.
For example for United States:
US-0.5%
US-1%
US-2%
US-Original Campaign (To find more publishers)

I then went ahead and grouped the publishers by their conversion rates for each Geo.
I blocked all the publishers in the groups in my original campaign to ensure only new publishers will continue to test.
And for each of the group of publishers, i calculated the average conversion rate.

It is important if you follow this style of filtering, that you assign significance based on the traffic for each conversion.
Basically: (Entrance/Total Entrances) * Publisher Conversion Rate
Then sum up all the results across the publishers.
This will ensure that small traffic publishers will not effect your average compared to the higher traffic publishers.

I repeated the same calculations i mentioned above to deduce the best bid for each group of publishers.
This way, for every group, i was able to bid more accurately and ensure that statistically, each group will remain profitable.

The reason i am only bidding 0.55 of my expected EPe is to allow some wiggle room among the publishers.
The next few days i kept cutting publishers and moving them between the groups.

A very interesting thing i discovered across my bidding, is what Amy mentioned above.
A publisher that preforms in the 0.5% group lets say at 3% conversion might preform at 0.5% at a higher bid. So its important to keep track of the moves between each group to avoid moving back a publisher you tested already at a higher bid back into the higher bids groups.

So this is my calculation sheet



And this is how it all this looks like in my flux



I must mentioned, that this process took me at least a couple hours a day.
I wish PopAds had an API to automate all this, their site is very outdated and missing good search functionality and have annoying things like if you add budget, you have again to go through all the pages to find your campaigns. (save the edit link in an excel file, will save you tons of time)

I do have an idea in my head how to automate all that using some back end code pulling data from Flux and using webscrapers to update the results and modify bids in popads (PropellerAds has cool API we can use to do all that, and i think we can even bid by publisher when i reviewed their API docs)

I think a project like that will allow me to massively test tons of offers from different providers and automate cutting of publishers and offers finding the gold nuggets. (ill call it MoneyMaking Bot :-) )

So here is the results for my campaign so far:

30 days - I finally green in total for the campaign :-)



07/05


07/06


07/07


07/08 So far today as of 5:30PM New York Time


Did you notice how the ROI keeps getting better and better?
If you can put a dollar on the floor and collect 2, how many times a day will you do that? :-)

So whats next?
1. I am going to focus on the 5-6 geos i am getting a decent amount of traffic and getting a decent ROI and stop all the rest. This will reduce the workload by a lot, and also increase the daily PNL.

1.1 I was thinking about it right now on my way back from work, we initially blocked traffic from Adult in PopAds - Gonna duplicate the successful Geos and test adult traffic separately.
The method of bidding and grouping above, i believe it should create profitable campaigns as well.
Using the strategy i am currently implementing kinda ensures that we are using statistics to our advantage ( Thanks professor :-) ) and we are always bidding according to the conversion rates.

2. I am signing up to any Pop network i could find and will follow the same process i mentioned above. First i will start with data collection campaigns, and then split it across CR % campaign. I read a few people mentioned propellerads is rejecting their campaigns, but i am sure i could find more networks that would allow this smart link.
Either way, I started doing the 40 days guide, and propeller is not giving me conversions, while PopAds does. Although like i mentioned above, Propeller does have a very cool API.

3. I am going to test some other offers and possibly smart links from mobida and monotizer, i am doing this so i can test this system of eliminating and bidding across different type of offers and learn some more while i am following the 40 days guide.

In conclusion, do not give up, keep learning and try new things.

Amy, thanks again for all your help!!!

Yossi

---

One more thing i wanted to mention,

In your main campaign, the one collecting additional data, you will get a lot of publishers with just a few visits and conversion or two.


It is not enough data to take into consideration in your calculation for the bids and might effect your bidding calculation but these are definitely worth testing on higher bids.
Currently i am moving them to 1% group and test them further.
But what i want to do, is to create another campaign bidding 0.85 * EPe of the 2% group bid.

So basically have:
US-0.5%
US-1%
US-2%
US-Original Campaign (To find more publishers)
US-NuggetsTesting - Bidding 2%Epe * 0.85

This will most likely result in negative PNL, but will help you discover faster what is the performance of these publisher and move them to the correct grouping campaign.

And always remember to block them from the datacolletion campaign to ensure you are getting the data in the right groups.

Yossi

---

May I know what is epe in the spreadsheet ?

---

Earning Per Entrance (The expected earning per entrance based on the conversion rates we calculated)

So if we have 2% conversion rate, this means we can expect 20 conversions per 1000 visitors on average.

20 conversions x Payout = revenue per 1000 visitors
revenue/1000 will give you earning per entrance (expected earning per entrance)

Take into account some click loss, bots traffic and publisher mood swings, i chose to bid about half of that.

---

Hi guys. I just joined STM a few minutes ago and this is literally the first thread I clicked.

I love the quality here. Could someone please direct me to where I can learn to create, track and scale campaigns like this?

I'm a beginner.

Thanks in advance.

---

Hi Yossi sorry for the delay in replying - it's been a crazy week and some!

First of all - kudos on using the systematic bidding method! It doesn't work well for small geos with limited volume, but can work wonders when applied to big geos with big volumes. (For small volumes, running multiple small campaigns can reduce traffic to zero.)

The main problem with this approach, as you've pointed out, is that it's time-consuming and labor-intensive. So, given the short average lifespan of pop camps, it may or may not be worthy of your time and effort. The longevity of the offer is also important for the same reason.

But if you could indeed figure out a way to automate this, then you'll definitely have an edge over the competition! 27%-70% ROI and WITHOUT a reduction in traffic volume is definitely something to be proud of!

Focusing only on geos that are promising (generating the most profit or has the biggest potential of doing so) sounds great - 80/20 rule at its best.

Really looking forward to seeing you make this MoneyMaking Bot!

Your scaling plan is flawless - I would have suggested exactly that, i.e. scale to many pop sources, then test many smartlinks. One piece of advice to get you started: Look into Monetizer.com. You'd be hard-pressed to find a smartlink that converts better. Of course you can always find something that converts better in specific geos or other targeting (non-adult vs. adult, specific OS, specific carrier etc. etc.) but overall speaking the Monetizer should outperform most smartlinks out there. If I was doing this, I would focus solely on the push smartlink and monetizer and scale to as many decent-sized pop networks as I can, before even testing other smartlinks, or individual offers (except those within Monetizer.com).

That would be the most efficient use of time IMO. Don't deviate from push smartlink + monetizer until you reach at LEAST $500/day in profits. Focus is everything.

Do let us know when you get there! And in the meantime, please feel free to keep the discussion going - you're welcome and in turn thank YOU for all the tips you're contributing! Extremely valuable and motivating!




Amy

---

Thanks for the compliment and we love having you join us!

If you're completely new to running campaigns, I'd suggest getting this guide:

https://stmforum.com/stm1dollaroffer/

Execute the guide step-by-step to get some experience on how to set up your first campaign, and then read every single post in this section and take notes on how to do more testing and optimization to grow your profits:

https://stmforum.com/forum/forumdisp...rst-Conversion

There's also the 40-day tutorial to further your learning when you want a more solid foundation:

https://stmforum.com/forum/forumdisp...l-2019-Version

(The Voluum version is completed - I'm still trying to finish the Binom and funnelflux versions. These are just different tracking solutions.)

Once you've completed the 40-day tutorial and run some pop campaigns, I would encourage you to also test push notification traffic - it's similar to pop, but involves having to create one more thing: The ad. You'll learn plenty just by going through some of the bigger threads (ones with lots of replies and/or views) in the push subforum:

https://stmforum.com/forum/forumdisp...8-Push-Traffic

Have fun and don't hesitate to start another thread to ask more questions!



Amy

---

Thank you so much, Amy. I deeply appreciate it.

---

Hi @mywebexperts. I have been following your progress and I have to admit its really inspirational and sure makes me believe this stuff is doable. I also have started my campaigned based on this guide but i am yet to be profitable though thats not the priority atm. Now my only question to you is about the excel sheet you made to aid in assessing the estimated bid prices per geo. How did u come up with it?
For me Flors table only had 3 columns: Country, Offfer Payout and CPM. how did u manage to come up with the other columns?

---

So at the beginning, i used the data provided by Flor to get the initial bids.
She actually provided me with geo, offer, payout and conversion rate %
But, you can easily reverse the calculation from CPM (Really, EPM, earning per 1000) to CR
So if you want to understand the math:

If for example an offer pays out $0.10 per conversion, and the CPM is $1 (meaning on average for every 1000 visitors we will earn $1)
Then 1/0.10 = 10 conversions, for every 1000 visits on average.
(10/1000) * 100 = 1% Conversion Rate (CR)

So to break it down in Flors terms:

CPM/Payout = Conversions Per 1000 visitors == Conversions Per 1000 visitors * Payout = CPM

And if you want to keep it real simple, without breaking your head, think of it like this:

Flor is telling you that for every 1000 visitors the average earnings is $1
You can bid $0.55 per 1000 visitors (or in popads 0.0055 per visitor) similar to how i did.

Don't forget, this is Flors data, not yours, she could already be receiving super targets (successful campaigns) coming from a million different traffic sources.
As soon as you get some of your data, calculate your own CPM using your own data.

Kill the geos which eat your budget but do not receive enough traffic (if you see some nugget's, like a publisher sending you 300 visits but giving you 20 conversions, test them on a different campaign).
Or kill Geos that the payout is too low for the cost of the traffic (for example Brazil, paying a 0.01 per conversion)

And focus on the few that you can Polish further, move the successful publishers to their own campaigns and exclude them from the main campaign.

Its a lot of work, but remember, the idea behind this exercise, and make sure you think of it just like homework, is to learn to sift through the huge amount of traffic you can buy, and focus on the traffic that can make you profit.
You will than apply the same idea, on more complex campaigns, where you will be testing, offers, landing pages, the offer landing pages, etc...
You will learn to kill offers that do not convert, kill or focus on landing pages that work/doesn't work and find the few gems that can make you (and the offer provider) a lot of money.

I am currently in Mexico, traveling for a month (life is short, no?) so i am kinda on pause, but i am following Amys 40 days guide with a little personal twist (Push traffic instead of Pops)
It is still all the same concepts, you still ask your AM for the best converting offers for X type of traffic,
You still test several landing pages and offers combinations
And basically keep testing!

Stay motivated, read A LOT, talk to some people (mastermindes etc...) and just don't give up, nothing is impossible, its just how you approach the problem you want to solve

Yossi

---

Thanks Yossi for all those details!

Making estimates like that is ingenious, and can allow you to get a feel for which campaigns are more likely to become profitable with more tweaking, and which ones would be less likely to succeed (in which case you'd cut them instead of investing more time and money into them).

Exactly!

We're running pop traffic, which is the lowest-quality traffic you can get compared to other traffic types. Most people don't respond to popup/popunder ads, which is why it's so cheap.

So if other affiliates that are promoting the same push smartlink, are running other types of traffic, their conversion rate and eCPM (aka RPM or revenue per thousand clicks) will be a lot higher. Usually though their costs of buying that higher-quality traffic would be higher as well.

Just something good to keep in mind when using Flor's numbers as a guide.


This should be printed out and taped up somewhere and read at least once a day - such great advice for every budding affiliate.

Time and time again, members have proven the fact that attitude is EVERYTHING in this business.

Thanks for the reminder Yossi!



Amy

---

Home > The Newbie Zone >