Home > Paid Traffic Sources > POP / PPV / Redirect

---

What would your next step be? (6)

---

Hello fellas,

Finally got my attention back into AM since started in June then got distracted by other work.

Subscribed to Adplexity, and I feel it's definitely worth the price difference vs Anstrex. (I was using in June).

Started in June with Pushground with no results, decided to tackle direct-linked propellerads HAKA offers.

As an initial trial, I picked an offer from Adplexity to run and just tested that for a bit.

Funnelflux setup: Simple Traffic -> Offer.

Step 1)

Took the top geos reported in HAKA and ran a multi-geo campaign in propellerads, with smart CPA goal of 0.35 (100% offer payout).

Results:
Spend: $9.624 Conversions: 10 Revenue: $3.5 ROI: -63.63%

So the offer definitely converts.

Step 2)

Picked the geo that had the most conversions and lowest views (highest CR) and ran that in a separate campaign.

Still on CPA Goal, but the goal was adjusted to 0.8 since it wasn't spending for a bit. I figured I could use some data faster.

Results:
Spend: $14.648 Conversions: 25 Revenue: $8.75 ROI: -40.26%

Step 3)

Once this ended, I had a few zones that had conversions so I thought, I'll try just a whitelist campaign. Made two campaigns, one using cpa goal and the other smart cpm.

CPA Goal ($0.3) campaign results:
Spend: $4.304 Conversions: 3 Revenue: $1.05 ROI: -75.59%

Smart CPM($3.0) campaign results:
Spend: $2.008 Conversions: 1 Revenue: $0.35 ROI: -82.56%

I stopped these campaign early because it wasn't spending fast enough and I figured even if it's profitable (highly unlikely) there wouldn't be much volume at this bid.

Step 4)

So I figured I needed a bit more data, so I duplicated the campaign from step 2 but this time set a higher budget but lower CPA Goal.

Results as of yet:
Spend: $14.175 Conversions: 12 Revenue: $4.2 ROI: -70.36%


So this is where I stand with this campaign and geo. Should a winner perform better than this at start?

What I have now is a list of zone IDs (no definite winner?) each with 1 conversions (2 max). And maybe some zones with high views no conversions (500+ views enough to declare blacklist?).

What should my next step be from here? Should I test a bunch of offer/geo combinations and see if I can find a clear winner or is there something to be done here with optimizing? Whitelisting all the zones that converted?

Thanks!

Any help would be much appreciated!

---

An update to this. After some more thinking and researching. It seems I should just test more offer using CPA Goal and not worry too much about zones at the moment.

Read a bit more into how each bidding method works, it seems CPA Goal would already use AI/stats to automatically bid for users that will most likely to convert.

It seems if I am not achieving good CR with a funnel on CPA goal, no amount of optimization would turn it green?

---

It's true that with the CPA model the algo should optimize the placements on it's own, so messing with the placements/zones in the testing phase can "confuse" it and you should let it do it's job first. A properly coded algo should pause the non converting zones automatically, once they spend the defined budget without any results.

Let me tag @twinaxe here, he has way more experience with CPA Push traffic than I do, I'm sure he can share some tips.

---

Sorry for the late reply

How did you decide to put zones on the WL, what factors did you consider for it?

Also, how many zones are/were in the WL?

Did all zones in the WL perform similar?

Don´t judge the zones only by views, consider the spend as well.

You could have zones with 5k impression at $0.20 but you could also have zones with 10 impressions at $5.

First thing, stick with BL campaigns for now to collect enough stats to build a good WL.

Then you wrote this "As an initial trial, I picked an offer from Adplexity to run and just tested that for a bit."

Does it mean you tested only one offer?

If yes, test more offers.

As long as you are still testing for good offers, landers or geos I wouldn´t run CPA Goal because it can get too expensive when you run it on unproven funnels.

Don´t get me wrong, you can do it and it can work but when the funnel is not converting good enough it can also be more expensive than running CPM/CPC.

This has to do with how the CPA Goal works.

In the beginning the algo will check if the campaign has potential or not, to do so it starts with high/very high bids.

When the campaign converts it starts to adjust the bids for the different zones and after some time you can have a really good and long running campaign that compensates the high costs from the beginning many times.

On the other hand when the campaign doesn´t convert the traffic will dry out pretty fast and you will be left with the high cost from the beginning and no chance to recover it in this campaign.

In CPM/CPC campaigns you have much more control over the bids and it can help alot to keep the cost lower.

---

Thanks @twinaxe and @matuloo.

Yes, my next step is aggressively testing more geo/landers/offer combos.

On the CPAg/CPM debate, I now hold a different view after some research, for now, to be proven wrong later, I guess.

If CPAg allows me to bid for higher quality traffic without bidding high for everything else, perhaps I could be more certain that the funnel/offer doesn't work in one particular GEO. Hence I could cut that geo/lander/offer combo earlier and move on.

Well, this is what I'll be doing now. Let's see.

---

Can´t argue with that reasoning and in no way I want to say that CPAg isn´t working good but there´s a big difference between runing CPAg on untested campaigns and running CPAg on already tested and good converting funnels.

When your campaigns just don´t convert good enough then the costs from the failed CPAg campaigns can add up pretty fast.

It can make a huge difference if you pay $0.15 CPC to check for potential in the beginning of a CPAg campaign that adjusts the bid to $0.02 later but only if the campaign converts good enough to keep traffic flowing or if you pay $0.02 CPC in a CPC campaign right away and can just cut it when it´s not working.

It won´t be much of a problem when you only test 1-2 campaigns per day but at this pace it will be hard to succeed anyway.

The more you test the faster you find a winner so the main goal should be to test as much as possible.

When you do so you will pretty fast learn the difference between losing $5 on a failed campaign and losing $10 on a failed campaign

Anyway, don´t let me distract you from your goals.

I can only give input based on my very own experiences and more often we see that things work different for different people

Don´t get me wrong, I often run CPAg campaigns from the very beginning myself to save time but I don´t test random stuff and I know what to expect from it.

Keep us updated how it goes.

---

Home > Paid Traffic Sources > POP / PPV / Redirect