Home > Paid Traffic Sources > POP / PPV / Redirect

---

How Could I have Done Better? Pop, PropellerAds, India. (5)

---

Hey guys, over the past week I returned to give AM a shot, again starting with POP. I ended at around -20% ROI.

I'm going to write about my choices and how I made them, maybe someone can point out where I went wrong?

1) Traffic Source: Propeller Ads, because STM guides seem to paint it as the best "newbie starting place".

2) Offer Source: Zeydoo, because the support guy told me that PropellerAds recommends Zeydoo as an offer source

3) GEO: India, because PropellerAds it's "hot right now" with Pop in their newsletter, and I wanted tier 3.

4) Offer: "Finance Survey New Feed" (Payout $0.03), because the Zeydoo support recommended it, saying it's one of the best offers they have.

5) Campaign Type: SmartCPM. I didn't really know how to choose one way or the other. CPM, CPA Goal and SmartCPM seem to have pros and cons.

6) Targeting: All suggested by the offer - Android, OS 8,9,10, Wifi & 3G, All ISPs, Chrome.

7) Optimization Goal: Told SmartCPM to optimized for $0.02.

8) CPM: I chose the recommendation, which was the maximum for all that traffic, approximately $0.43 CPM.

9) Zone Culling: The optimizer automatically removed zones, and I also culled them when they spent 2xPayout with 0 conversions, or 3xPayout with 1 conversion.

10) Targeting Culling: I looked at all data in Voluum, and culled Wifi and 1 ISP (JIO), which were both ~double CPA of others.

11) Switched to CPA Goal: I optimized the Smart CPM to about -20% ROI, then repeated the process in a similar way, and again got around -20% ROI

12) City and Zone Whitelist Cross Section: I wrote a program in MATLAB to analyze all the Voluum data and tell me a cross section of targeting that would be profitable. I ran it and was about 30% profitable, but miniscule voluum, so only made about $1 on day 1, then day 2 it stopped working.

13) Desperate attempts: I tried a bunch of other things, like CPM High Activity users, different SmartCPM CPM, different CPA Goal CPA. Nothing worked.

---

So here would be my question: I followed all the advice from PropellerAds, Zeydoo and Pop Newbie guides, and wasn't profitable. I'm an expert with data (generally), but I still couldn't get profit. So at which step have I made a mistake / misunderstood?

Thanks for any help!

-- Matt

---

Dear Matt,

Let me kindly review all of the points and share my point of view on the decisions you made.

Good. Propeller is one of the best traffic sources for Pops. If it comes to other possibilities: I've been using some Clickadu latelty that performs simillary to Propeller on some GEOs.

As I know Zeydoo and Propeller are from the same company. At least their platform marketing is not so obnoxious, I just got a popup once on the propeller with the recommendations, so this is good.

Have one imporatant fact in mind: "hot GEO right now" is not enough to indicate what to promote. Does Propeller mentioned which vertical is doing the best?

Another thing to consider is that when some offer is "hot" it's already half-way saturated.

Someone made it hot and I'd personally rather aim to be that "someone" than the person that found about it through affiliate network mass-mailing.

Financial Survey was covered nicely in their case study here, at STM. You can get your inspiration juices flowing HERE.

Whenever it's possible use SmartCPM (you can split test it with CPA Goal).

SmartCPM is much better if it comes to campaign cost compared to CPM. Always choose SmartCPM instead of CPM on the traffic sources if its only possible.

Why?

Because SmartCPM use dynamic pricing for each of the available zones that you are bidding for. This basically means that when you set up your CPM (cost per mile a.k.a cost per 1000 views) at 0.5 you will pay UP TO 0.5 per 1000 views for each of the zones. When you access the report, you will see that each zone will have different CPM. Some zones are just more wanted (meaning converting better) than the others so the pricing is adjusting accordingly (up to the price that you set).

CPM on the other side, is using fixed pricing. So when you set 0.5 you will pay 0.5 for 1000 views on each of the zones. No matter if it's a crap or a rockstar. This leads to overall higher campaign cost to get the same results.

CPA Goal is a bit different - you are setting your CPA (cost per action) Goal instead of CPM (cost per mile a.k.a 1000 views).

In that case Propeller will try to automatically optimize your campaign (disabling zones, adjusting bids) based on the conversions details to fit your CPA up to the value that you set.

Your CPA Goal should be slightly below the conversion value of your offer.

Hope this make sense.

Test Android and iOS in separate campaigns.

Same goes for carrier and wifi.

So you end up with 4 separate campaigns: Andoid WIFI, Andoid 3G, iOS WIFI, iOS 3G.

Pin submits are usually converting better on the 3G traffic.

You can even get one step further in that split testing and test Propeller vs Brocker traffic in separate campaigns (when you find an offer and targeting that works).

So you just let Propeller in that case to automatically optimize your campaign - this may be good but I would rather say it's better to optimize manually to understand better exactly how it's done by the scripts.

Try running it with disabled Auto-Optimization.

Then, analyze your tracker and disable underperforming zones (more of that in the Amy's 40 Day guide, check it out if you haven't already).

I unfortunately must leave my house right now, I'll update this post later on today and relate to all of the remaining points.

Your progress looks great so far, keep it up. -20% is good ROI for the beginner.

plutus

UPDATE:

You can try to lower / upper this in separate campaigns to see how it rolls. Changing the bid is one of the way to optimize campaign.


Good, following the rule of thumb, not much to add here.

Like I mentioned before - you can safe yourself some time and run 3G/WIFI in separate campaigns (both in Voluum and Propeller).

This way, you can cut underperforming campaign as a whole and focus on the one that is converting.

Of course this cutting apply only to the campaigns that are basically hopeless and are just adding costs. If both are converting, optimize both.

CPA Goal can do magic for you on close to no effort.

But.

The catch is that you need proven offer / funnel for it to do so.

As you probably know (given the Matlab experience) - no matter how good the algorithm is, if there won't be enough data (conversions) to process it just won't be efficient in the optimization process.

Woah. This one is geeky, I like it. I haven't actually did any level of applied mathematics on my tracker data so can't suggest anything that would optimize your algorithm.

The thing you can do is to play with bidding - incrementing it for a day, decrementing it for a day, comparing results.

Sometimes bidding higher, to the biggest surprise, can make you profitable out of the reddish.

Be careful, having too much going with your campaign can make all of your data completly worthless.

For each test like that, start fresh, with new campaign both on Voluum and Propeller.

---

The misconception that most beginners belive (including myself, few months ago) is that it just takes one or two offer to see profits flowing.

Like I mentioned above, this offer might already be saturated - meaning that dozen of affiliates already run it and milked it all, leaving nothing more than a wrinkled skin.

You were close to profits at -20% so this campaign is worth spending some more time on (check out again the stuff that I posted above)

Do have in mind possible profit margin and your current cost.

Overspending on the campaign without a possibility to get your money back in profits might not be the wisest decision to make.

Do test more offers alongside.

Start with 2-3 offers at the same time (if your budget and time allows for that).

And gradually increase that number when you feel more confident in the optimization and campaign creation process.

---

Ok thank you so far! I have some follow up replies, though I'll wait for the rest of the post so you don't have to check twice.

Great information, thank you

It's absurd how helpful people are, really. (In a good way).

---

Thank you, I just updated my previous post, check out the remaining part.

It's nice, isn't it?

---

Hi Plutus. I got halfway through writing out all the little details about how your post changed my ways of thinking, but I scrapped it to instead simply say I carefully studied what you said and you’ve given me numerous ideas for how to approach my next attempt. They will be apparent in my next write-up, which I’ll post in this thread. Thanks so much for being so helpful

---

Home > Paid Traffic Sources > POP / PPV / Redirect