Follow Along Zeropark + Voluum + PeerFly. Goal = Spot 3 Profitable campaigns

kaminari
Thanks Luke and Nick for updates on my questions.

So if I am setting auto bid of $5 CPM so I(system) will bid less on target where top bid is $1 and then bid high on target where top bid is $8 CPM and on average I will pay $5 CPM? also if I set manual bid then it will bid that manual amount regardless of top or minimum bid of target?
That's correct, but with Zeropark, it's a CPC and not CPM.
 
Geozo
That should be correct. This is from the Zeropark website:

"Zeropark has a very sophisticated bidding system available. There is a pricing algorithm that ensures that you pay the minimum price required to win the visitor. Hence, most of the time you will be paying less than your actual bid. Secondly, Zeropark uses average bids. We compensate for the fact that a lot of the traffic that you will receive costs less than your bid by exposing you to more expensive traffic. Hence, most visitors that you receive will be either cheaper or more expensive than your bid. However, the average price that you pay for the traffic will never exceed the set bid. Lastly, we split traffic among top three bidders. Even if your bid is significantly lower than the top bid, you will still receive some percentage of the traffic.
To sum up: if you’re just testing the Zeropark traffic, go easy with the bids; start low and bid up once you see conversions."

EDIT: Woah that was weird. No one replied, then I submit the answer and there were 2 replies. Either way i'll leave this post, as it gives a little more info about their bidding.
 
Hey @Mrtheone14

to add to what @Nick correctly said/quoted the system will try and average out the bid to your campaign bid setting. This is to give the campaign a chance of winning higher quality visits from a source, so you get a to test a wider range of traffic.

However, when you set a custom (manual) bid, we assume that you know how much you need to pay to win the optimal amount of traffic from this source at the optimal price for your offer, so the range is much closer to the custom bid.

Thanks!
 
Last edited:
Wow thank you so much @Zeropark

So I can bid (auto bid) higher without worrying to spend more on a target with lower top bid. I mean I will spend/pay $.001 on the target if top bid of target is $.001 instead of my auto bid of $.005?
 
@Mrtheone14

It depends on your level of filtering/targeting. Say for example you were doing a RON open to all traffic types and on your current bid you were winning mainly windows desktop traffic. Then you increased the bid and you start winning IOS Carrier traffic. It is much more exspenisve, so it increases your adverge cost to match the new campaign bid amount. So yes you would still pay 0.001 for that Windows desktop bid, but you would be paying 0.005 for the new IOS traffic you were not buying beforehand. The system would balance out at your new auto bid.

In short: keep the auto bid low and raise custom bids for targets and sources that are converting.
 
Alright, quick update on my version of @markbeats campaigns since launching them yesterday:

View attachment 2588

Campaign 1 is definitely off to a rough start, but the traffic is very cheap. I only have 1 conversion to base my decisions off of, but I want to start removing sources that aren't performing well since I've already spent nearly 25x the payout. We're going to need some profitable sources/targets to be able to get this one profitable.

Campaign 1 Optimizations:

View attachment 2589

I do not have any profitable sources. That's not a shock since I only have 1 conversion. I'm going to pause the top 2 spending sources since they do not have a conversion yet. There could be winning targets in those sources that I just do not have enough data on yet, but I want to focus on the source I do have a conversion on and block out the rest of the "noise" :)

Next, let's take an even deeper look at the source, russet-lark, that has a conversion. I'm breaking it down by device type since I'm currently targeting both mobile and desktop on this campaign:

View attachment 2590

Alright, we have some green! It turns out my 1 conversion so far was from a mobile phone. We can now also group by OS and see which type of phone got that one conversion:

View attachment 2591

Well, would you look at that. Some data. So far, with Android phones from russet-lark, I have nearly an 80% positive ROI. To be fair, I don't have much data on IOS (30 clicks for $0.14) for this source. So, I am going to update my campaign so it's only targeting mobile devices.

View attachment 2592

I'll dig deeper into this source once I have more data. I've doubled my daily budget on the campaign to $20. We'll see how this next batch of traffic goes for us.

Campaign 2 Optimizations:

View attachment 2593

A quick glance at the sources I am getting traffic from on this campaign shows a lot. Without going past this screen, I know:
  • whey-wild needs to be paused
  • badious-buzzard has a ton of potential
  • conquelicot-bear could be AWESOME
This campaign, although currently ROI negative, definitely has some potential. About half my money has gone to whey-wild and it doesn't seem to want to convert with this offer. I could breakdown it's targets and see if there are any losers in there that are eating up the budget, but I'm just going to pause it for now and focus on the sources that are working.

Let's breakdown badious-buzzard by target since it's nearly breakeven with 2 conversions already:

View attachment 2594

Green is good! We have 2 targets on this source that are profitable with a good bit of traffic already. Let's see what Zeropark has to tell us about these targets:

View attachment 2595

Good news. We're only buying about 10% of the traffic of the top target and according to the Voluum report screenshot above, our EPV on it is almost $0.05. My average cost per click right now for zulu-sat-exQfZLt1 is only $0.0179. I will manually bid $0.025 right now and happily take a 50% ROI if this target can maintain a good conversion rate.

IMPORTANT: 1 conversion could be a fluke. 2 is much better.

We only have 1 conversion on each of the top targets right now so that doesn't mean we're going to be able to maintain the current conversion rate those are producing. HOWEVER, it's definitely encouraging and worth a test.

With zulu-sat-exQfZLt1, I am only buying about 10% of the traffic. From my experience with Zeropark, I know I can buy 50% with a good bid. We'll see if $0.025 is good enough and then I'l adjust my bid accordingly once I have more data.

Our conversion rate on the second target (hotel-god-Dn4BRIyY) is even better. It seems to have less traffic, but I want to buy as much of it as possible too since it appears it may be a winner. Our EPV is $0.06. Currently, my average auto bid for this is $0.015 and I'm buying about 18% of the traffic with that bid. I probably do not need to bump my bid to 50% of my EPV ($0.03). I'm going to bump this one to $0.025 too just because I like when numbers match ;)

Bumping both of these target's bids should mean I will get more of their traffic. We'll see how it goes.

I am going to leave the other source along because there's not enough data. The offer for this campaign pays $2 and besides gridelin-bear, I haven't even spent $2 on any one of the sources without a conversion (besides twhey-wild which I've paused).

Finally, let's take a look at our most profitable source for these 2 campaigns so far, coquelicot-bear.

View attachment 2596

I already have a 551% ROI no this source. BOOM. That's the good news. The bad news is that I've only been able to buy 14 of the 5,800 clicks that Zeropark is showing "Available" on this source since yesterday. My average cost for this source is $0.0219.

IMPORTANT NOTE: If your average cost is showing as more than what you are actually bidding, you are underbidding that source on Zeropark.

The reason I haven't been able to buy much of the traffic is because I'm underbidding. That's a good sign. This source must convert well if other advertisers are bidding higher for it's traffic. Zeropark was nice enough to send me a few clicks from it even though I'm bidding too low and I was lucky to get a conversion from one of those. So, we definitely need to test this source more.

I'm going to take a guess and say I need to bid at least $0.05 to get a better look at this source. My EPV right now is $0.14. I can afford to try $0.05 based on that data. So, for coquelicot-bear, I am setting a manual bid of $0.05. I will continue auto bidding the rest of the sources at $0.02.

Here's what my source tab looks like for this campaign now:

View attachment 2597

This seems like a lot of work for 2 campaigns that have lost me about $30 already right? WRONG! This is what you have to do if you want to be successful with affiliate marketing.

90% of the affiliates I know would have looked at the stats and said "well, this isn't going to work" and paused both campaigns.

They may not work, but it definitely looks like they have potential. Now, we just need to spend a little bit more money to see if we're on the right track :)
Holy detailed. I didn’t know you’d want to give such little data such importance in the early stage of testing. Seeing 1 conversation and treating it as a trend will have to be an adjustment.

I’m not sure if I understand the full scope of opportunity with this type of affiliate marketing, the numbers are smaller than I thought the would be going into this.

Learning lot looking forward to learning more. Awesome @Luke made this one public 🤗
 
I didn’t know you’d want to give such little data such importance in the early stage of testing. Seeing 1 conversation and treating it as a trend will have to be an adjustment.
Everyone has their own strategy. 1 conversion can most definitely be a fluke/luck. It most definitely does not show statistical significance. However, it's like getting a nudge in the right direction. You don't want to complete ignore it, right?

I’m not sure if I understand the full scope of opportunity with this type of affiliate marketing, the numbers are smaller than I thought the would be going into this.
You can increase the numbers with scale. Last year, I took a campaign like this and turned it into a nice money maker across multiple traffic sources. I always try to start small though. Scale with proof of concept.

Learning lot looking forward to learning more. Awesome @Luke made this one public 🤗
Glad you like it 👍
 
Top