Home > Push Traffic >

---

My 1st FA: Propeller + Unbounce + Financial offer in LatAm (24)

---

Hi STM forum,

I've been lurking for a few months, so I think it's time to share a bit and I would love to receive feedback from the Push masters here.

A bit about me
I've been doing affiliate marketing for 4 years. 95% of it in the finance vertical.
I manage a team of 7-8 people right now and we leave no stone unturned when it comes to buying traffic (SEO, Google Ads, Facebook, Display).
We never had success with PropellerAds. Our tests were "okay", but not profitable at scale and we decided to focus on other things.
In the last month our Facebook Ad Account got burned and we are warming up a new one. So I thought it was a good moment to overspend money and time on Propeller and see how far we can go with it.

Choosing an offer
We have an offer that pays around 1.5 USD and we made quite some money last month on Facebook.
CVR was 13.92%. Hence EPC was 0.21 USD.
Vertial: Finance
GEO: Somewhere in LatAm
Pre-lander: needed to warm up the visitor.
I know the advertiser personally, so I know we won't have problems.
I think that for now we won't be testing any other offer.

Tracking
Propeller -> Unbounce click tracking: Simply passing $(SUBID) macro in one parameter to Unbounce
Unbounce -> Affise click tracking: Trickiest part. We sometimes use forms and sometimes use links on our Unbounce pages, so I rewrote our JS code so it takes the $(SUBID) from PropellerAds and append it to the Affise tracking link as ref_id in both cases.
Unbounce -> Propeller to report views. Pretty straight-forward config. I just had to copy-paste the code.
Affise -> Propeller to report conversions. I just needed to use two Affise macros, one to send the ref_id so Propeller can identify the conversion and another one to pass the value of the conversion (although it's fixed to 1.50USD, in this way we don't have to re-configure if payout or offer changes).

Creatives
Someone from my team prepared some creatives based on those that worked well on Facebook. But probably on Propeller we will be more aggresive.

Campaigns
I turned on two campaigns, one CPA and one CPC.

---

Depends on the geo, your targeting, your bid, your creative.
You also only run desktop now, mobile usually has higher volume so when the offer also accepts mobile traffic it would be good to also target mobile when you want to scale.

To start a WL you need more conversions, 4 isn´t enough at all.

You don´t need it yet.

Why do you use Unbounce?

Such LP builders are better for more traditional leadgens but for push traffic I would use more "CPA like" landers.

About tracking, do you Affise yourself or does the network with the offer uses Affise?

In any case I would recommend to use a tracker yourself so that you also can splittest landers (and maybe offers).

---

I guess you mean

0.005 for the "Explore" campaign (that's the minimum bid for MX)
0.008 for the "OneConversion" WL
0.010 for the "TwoConversions" WL

Btw, what user activity levels are you targeting?

---

Hi twinaxe,

Indeed my bids are like that. I missed a 0 in my post.

All. This is my targetting so far:
- Only Classic Push
- Propeller+Brokers
- High+Medium+Low

My plan is to break down the campaigns into 6:
ClassicProperllerHigh
ClassicProperllerMedium
ClassicPropellerLow
ClassicBrokersHigh
ClassicBrokersMedium
ClassicBrokersLow

And after that break add 6 more campaigns for InPage traffic:
InPageProperllerHigh
InPageProperllerMedium
InPagePropellerLow
InPageBrokersHigh
InPageBrokersMedium
InPageBrokersLow

But that's just what my intuition tells me.
What's your experience? Shall I already break down the campaigns based on activity level?

---

Yep, it´s best to start with high activity for testing, then expand to medium and when you have enough stats for a good WL run it on low activity.

You can also test low activity on BL but I mostly run it on WL because it works better for me.

I also don´t separate Propeller and broker traffic, I run them together.

---

Congrats, solid progress

One thing, can you please insert the images properly in the posts?

It´s much easier to read then.

Just check the link in my signature how to do it

Ok, so you run the campaign on CPC and offer payout is $1.80, correct?

Then you set the price you want to pay per conversion to only $0.01?
This would be way too low, usually you set the goal to something between 60% and 80% of the offer payout.

It´s pretty easy to run on high ROI with such goal but the platform will block placements and other targeting options then so fast that your campaign will run dry.
You run your campaign at $0.006 CPC or so, this means when you set the conversion goal to only $0.01 a placement has to convert each 2 clicks or it gets blocked.

That´s way too aggressive.

When you increase the goal, especially when it´s such a huge step like going from $0.01 to $1.50, then the campaign has much more room for testing.

Of course can lead to temporarily lower performance because more testing costs more money but usually performance should level out later so that the campaign runs more stable then.

I also never use the Smart Optimizations Tools myself.

I can imagine that it´s tempting to set really low CPA goals for maximum ROI but you don´t do yourself a favor with it.

Seriously, with $1.80 payout you are looking for $1.70 profit per conversion.

This would be 1700% ROI and that´s not realistic at all, at least not when you want to run on higher volume.

You would have to block placements that fast that in best case you will be left with only very few very good converting placements or in worst case don´t receive any traffic at all anymore because you killed the campaign with too aggressive optimnizations.

Don´t put too much focus on the ROI, the only important number is the profit.

A campaign that makes 10 conversions per day at 1700% ROI may sound like a huge winner but it isn´t when you can´t scale it.
Such campaign would make only $17 per day with your offer.

Set more realistic goals and rather go for highest volume instead of highest ROI and the profit from the additional conversions will surpass the advantage of a high ROI at one point.

A campaign that makes 100 conversions per day at only 20% ROI would result in $30 profit with your offer.

Is this from all 3 activity levels in one campaign together?

I don´t recommend to run desktop and mobile together in one campaign.

One reason is that they both can perform very different, another reason is that the difference in volume can be huge and bids can also be pretty different.

It´s better to run it in separate campaigns.

---

Feedback

Thank your for your feedback once again twinaxe!

Sorry about it. It should be solved now.

That's correct.

Regarding scale and profit margins. I generally agree with you. It's better to find a great volume with a small profit marging.
But you have to be very aware of a few things:
- ALL your costs. In this case Unbounce has a significant cost (that of course can be lowered). Even currency exchange rate (>0.01%).
- The more volume that you send, the more likely that the advertiser stops you if the traffic source has quality issues. And it's also most likely that the traffic source stops you if they don't like your offer.
- Your bank account cash flows. If you pre-pay and get paid from the advertiser EOM + NET60 you better run with a bit more of profit margin.

About bidding. I also agree with you. Auto Optimization with 0.01 was a mistake... but made the campaign profitable very early on. I was totally okay with this over-optimisation, having the campaign in with low volume and high margin and profiting only from the zones that looked promising early on. And with time I could have brought up the scale.
My first time back driving after a long break I also don't like to even get close to the speed limit

Correct

I agree. I've already split up the campaigns in two

Performance



I changed the reports to UTC.
Yesterday we ended the day in red. I was chaotic with the configuration of the main account and my zone bidding strategy was not working.
But today we are in record RoAS after spending 1/4 of yesterday's budget.


New bidding strategy

- As twinaxe pointed out, due to this wrong Automatic Optimization I was profiting from a very few zones that looked promising at the beginning.
- My bidding strategy (increasing bids after 1 conversion and after 2 conversions) was not working at all because I had a very few zones with many conversions.

I redesigned my "Explore" campaigns.
I tried to come up with a smarter bidding strategy.



I divided my zones into 4 groups:
- Those zones I don't have enough information about (2x the payout). So I don't do anything about those.
- Those zones that are I'm sure are unprofitable. So I will block them from all "Explore" campaigns.
- Those zones that are I'm sure are profitable. I will call this "BID2" and bid up a bit.
- Those zones that are I'm sure are very profitable. I will call this "BID1" and bid up a bit more.



I will also run 4 "Explore" campaigns to braek down for:
- Device
- No AO and Auto Optimization = 1 USD

Those changes so far looking promising. Looking forward to see how the day ends and see where I should adjust.

And of course I keep listening to your suggestions!

---

The campaign is still working well. But I'm still not confortable enough with the set up to scale it up to bigger volumes.

Some thoughts:
- Automatic Optimization is being more useful than I thought
- CPA Goal campaign is starting to bring some small profits.
- It's being difficult to monetise desktop traffic

---

Sorry, missed to reply

Yep, you are absolutely right but more profit is more pforit, no matter what the ROI is.

$1k profit from 20% ROI is still more than $200 profit from 1500% ROI

No, it just seems to be that way.

When your traffic doesn´t convert good enough for the advertiser you will get paused anyway, when you send more volume it just happens faster.

Let´s say an advertiser let you run for $200 in conversions before he decides if you can continue or not.

On low volume the $200 could take 2 weeks, on high volume it could take only one day but the result is the same.

When they don´t like it they don´t like it independent from the volume.

Sure, that´s a valid point but then I would rather try to find a solution to get better terms instead of limiting me.

Apart from that Net60 is really bad, I only know it from the Amazon affiliate program.

To be honest but this setup looks pretty complicated, at least for me and my workflow.

When it works for you then go on but keep in your mind that now were talking only about one campaign.

Imagine you run dozens of campaigns at the same time, then it will take lots of tme just to prepare such tables.

I try to keep my stuff as simple as possible and I rather risk to be not that accurate but therefor have more time available.

CPA Goal can be very nice but it can take few days to show real results.

It can also result in losses first so don´t wonder when it starts a bit bumpy and just keep it running.

When the funnel is good it will pick up.

It depends on what you run, with the right vertical and offer desktop traffic can be very rewarding.

---

Stay away from Spanish customers then


I agree. I might be over-optimising a lot for now.
But I already had my WL and BL automated using Propeller's API, which honestly was fairly simple. I thought I was going to need something like TheOptimizer.
I'm a compuster scientists, so I will probably always choose the over-complicated and automated option even if it's not worth the pain

Thank you for the advise. I think I'm going to give a good try to CPA Goal.

---

So, what happened with the campaigns?

After spending around 2500 USD and ROAS around 100% with some decent days (120%), I was preparing some new analysis to scale it up. But before that happened, the advertiser complained about the quality .
I will probably try again this offer when we are sending more traffic from other sources, doing some extra work on the funnel and the type of customers.



So after reading twiaxe's thoughts on SOI vs RevShare offers, I decided to go for a good CPA offer to continue playing with Propeller traffic. At least I won't get shut down when things start getting interesting.

Also this time I spent some time with my team building our own simple HTML pre-lander, instead of using Unbounce.

So far I think that we are going in the right direction:



Having fewer more expensive conversions make discovering good and bad zones more difficult ("the law of large numbers"). So I will have some fun thinking how to use other customer signals (e.g. click on the lander) to optimise.

---

These are the results so far:
CPA campaign: $826.993 spent. -31.78% ROI.
CPL campaign: $2,234.002 spent with +20.04% ROI

CPA campaign is now off. But we are sending again traffic to the CPL offer, hoping that the advertiser doesn't complain about it.



We have a very low volume right now, but we have some more important issues to take care of and I'm logging in the account every 3-5 days now.

---

Good to see that you´re still working on it.

---

Still making some money thanks to a good offer that we got recently!
CPA and AO work pretty well when payouts are low. So we are going to use them to run campaigns faster.
We are also working on our landing pages and building our mini landing page builder.

---

You're killing it man!

Nice spend too!

---

Wohooo! Yesterday we had a great day. 600 USD spent -> 960 USD in revenue.

I think that we are now ready to scale it up next month, ask for higher cappings and lower our margins to make more $$$.
- We have two offers with RoAS over 160%
- One of the advertisers is happy with the quality and wants more volume.
- We have the type of creatives that work for our offers figure out. One of the offers tested in all placements except PopUnder (it's in green in Interstitials, InPage and Classic push).
- We are not comfortable running the campaigns using Propeller's algos (AO and CPA) and know how to play with them.

Propeller is still small in the overall sales picture of our company. But it feels very good to win another channel

---

Sometimes CPA Goal plus TheOptimizer can work very good.

But the geo should be big enough so that it doesn´t hurt when you cut placements by Optimizer rules.

Are you still running financial?

Because I use exactly the same creative for financial campaigns for about 7 months or so across many geos

Once you found a good one I wouldn´t waste much time trying to squeeze another 2% ROI from creative splittests, better focus on scaling then.

So your running only CPC now?

In the end it´s best for maximum volume when you can run both together, CPC and CPA Goal.

---

Yes. Still running financial and not spending much time on creatives.
We still have a few things to test.

Sorry I meant that we are now comfortable with AO and CPA.
We are using CPC+AO and CPA.
CPM+AO resulted in bids that are too high to reach profitability in LatAm markets.

---

July was a good month overall.
We had one campaign that worked very well and one that worked well.
One advertiser wasn't happy with the quality of our leads, which is pretty bad news.
But now Propeller is playing a relevant role in our media buying strategy!

---

Thanks for the update.

My financial campaigns are also running very stable for many months already but last month one advertiser also wasn´t happy with the quality in the backend.

Although I already send info about campaigns and trafficsources in url parameters they can´t tell me which ones were bad for them so I had to stop many scaling campaigns.

Anyway, I keep my fingers crossed for further success

---

August was worse than July.
One of our main campaign was paused.
We were able to scale up our first offer outside the financial vertical. But once again advertiser is not happy with the quality...

Let's see how September goes. We are very constrained by the offers.

---

Thanks for your update, great to see that the campaigns are still running.

Few days ago I found a new financial network that I want to test, can let you know how it goes

---

Certainly! Also I can share some decent networks with you in case you don't know them.

---

I'll message you when I am on my pc

---

Home > Push Traffic >