Home > > Newbie Follow-Alongs

---

-56% ROI Direct link Offer. Lets see if it get become profitable. (2)

---

This is a copy of the original post from the pop tutorial thread : https://stmforum.com/forum/showthrea...N-Part-2/page2

Hello Amy!,

First off, just wanted give you a big shout out and thank you very very much for this awesome tutorial!. It gives someone like me with 0 experience an easy step by step action plan, instead of over thinking everything. It took me long enough to right this post, let alone a whole guide(God Bless you).

So, I have launched 10 campaigns so far, Im sticking with the strict 5x payout rule to decide whether to ditch or keep running the campaign. I have 1 campaign that shows some promise(or maybe just a spec). So, I continued with it, to go through the optimization process just for the experience. My goal, is just to learn how to optimize a campaign in the mean time while I continue testing other campaigns, if I could get it profitable that would be awesome too.

Attachment 17438

These are the stats. As you can see from the screen shots, All conversions came from Chrome Mobile browser which is only one third of the traffic. All other browsers(two thirds of traffic) are pretty much 100 percent loss, and I dont think have a chance of being profitable as long as I am direct linking.

Attachment 17439
Attachment 17440
Attachment 17441

When it come's to Ad placement's(Zone's), Over 50% of conversion's are coming from one placement.

Attachment 17442
Attachment 17443

And Finally when it comes to OS Almost all the traffic is coming from Android.

Attachment 17444

What I did :

1-I created a whole new campaign in propellerads(by duplicating the campaign I was running).

2-Then I Chose OS= Android

3-Excluded all Web browsers except for Chrome.

4- I Excluded 3 placements that where at 2x payout. (payout is $1.26)

Attachment 17445
Attachment 17446

Finally after running the "NEW" Ad campaign I noticed I was getting more Android traffic than Chrome?. I emailed propeller ads about it..

Attachment 17447

They responded by explaining what CPM IS ??...... (uh-hello??)
Attachment 17448

So for now I just set the CPM to a low bid where I get like 400 clicks a day (practically on pause) I got another conversion though lol.

So, in conclusion after optimizing I still got the same quality of traffic I was trying to cut.Which really sucks because I cant get any feedback/experience from the new campaign. After cutting 3 placements the number of clicks also dropped by like 80 percent. I counter balanced that by up'ing my bid, but since the traffic quality was not what I "Targeted" I had to stop the campaign.

Any advice on what to do from here?. Have you had the same issue with "targeting traffic" but still got the the same quality traffic as before?.

Finally, any advice on when to start using lander's?. I am working on ripping and editing them now. But, do you think I should first try and get a DL offer profitable first?

All feedback is welcome. Thank you very much

---

God bless you as well! Feedback is always appreciated, and not just the good kind. Thanks for giving it a try!


The 5-payout rule works well for the initial 10x payout in spend, but really, my experience is that for direct-linking, if the offer doesn't start doing at LEAST -50% ROI very early on, then it would be very difficult to achieve green just by cutting placements etc.

The reason why I "slackened" from -50% ROI to 5-payout-with-at-least-one-conversion in the beginning, is to give the offer more of a chance. The more traffic we run, the closer the CR would be to the "true" CR, i.e. the one we can expect if we continue to run the offer long-term (although nothing is long-term when it comes to pop). But we won't want to spend a lot of money on finding out whether the offer has a high-enough ROI/CR to be worth continuing. So, as a solution to this dilemma, I slackened the criterion in the very beginning to give the offer a chance, and then apply the -50% ROI rule in a later step in the process flow, once I feel enough stats have been gathered.


Very nice! Seeing lots of stats makes giving advice so easy.

It's so much more efficient to judge campaigns on a case-by-case basis and suggest an optimization approach, than to follow a one-size-fits-all approach as I've laid out in my optimization process flow.

You'll notice that in the last step of my process flow, I do suggest to target the most profitable traffic segment(s), and only as a last-ditch effort. Basically the process flow tries to keep things simple by focusing on cutting placements mainly. That approach is good in that it would be less likely to cut out big chunks of traffic like cutting a major OS would, but the disadvantage of course is that it may leave some of the big loser segments running (like Android Webview in your case). I'll make more improvements to the process flow based on our conversations - thank you!



I would start by excluding Android webview at Propeller right now, and then keep an eye on the Android Browser until it collects more data. As you cut placements, the ROI of other traffic segments such as browsers and OSs will increase. So it's recommended not to cut the bigger traffic segments unless it's doing such low ROI that no amount of cutting placements would likely make it green, OR if later on in the campaign, a traffic segment is still not green in spite of cutting the worst of the bigger placements.

As for Samsung Browser - you can try excluding the "Other" browser setting in PropellerAds, then see whether you still get traffic from this browser. I don't remember whether that would do the job. Too bad propeller doesn't pass back browser information via a token - otherwise we'd know which browser listed in propeller's campaign settings correlates with which browser listed in Voluum.


30% ROI is JUST barely a sufficiently-good ROI. Yes it's bringing in half the conversions, but also notice that it's also bringing in a good percentage of the total traffic as well.

For future reference: Although it would make sense to start a new campaign to target JUST the best placements and bid higher, on pop networks this often will not work out too well. Often, you will end up with significantly lower traffic (from the same whitelisted placements compared to the traffic amounts they were sending you in the original campaign) and CR. But if you have let's say 10+ placements to target, it would be worth a shot to set up a whitelist campaign.

Another approach to try, once you've identified the best placements, is to just cut the unprofitable placements in batches starting from the ones in greatest loss. There seems to be a filter on some traffic networks where if you're targeting too few placements (or other traffic segments), traffic would drastically decrease. So if you cut in batches, as soon as you see a drastic change, you can reinstate the last batch.

Also - for the 3 placements you've identified for cutting: If you're wanting to exclude the Android webview browser, then it may be good to look at your tracker stats Browser > Zones, then drill down into Android Webview > Zones, and look at those 3 placements to figure out whether they would still meet the cutting criterion WITHOUT Android Webview. It's always good to keep in mind that when you cut one thing, other stuff's ROI may increase. This should be the case when you cut Android Webview - the ROI of placements should increase in general.


Almost 100% of the traffic is also from android.


For OS, it would probably be good to include all OSs - they're not sending much traffic anyway so aren't really costing you. And - although I don't have proof - I find that the more segments that are included in the targeting, the more traffic the campaign will likely get. I suspect some traffic sources' algorithm will assign higher priority to campaigns that are targeting broader. But you can test this later (can leave it as it is for now).

Same applies to whitelisting Chrome as browser. (Note: You could also have toggled the "Exclude" to off, and just specified "Google Chrome" - way easier than trying to exclude everything else. But like I said, I would have tried blacklisting the worst instead of whitelisting the best. But that's something you can test later.)

And we've already talked about blacklisting those placements.


The response is completely unrelated to your question. It sounds like a reply for a publisher, not an advertiser (which you are).

I would write back to ask the question again.


Setting the bid too low in the beginning is not ideal. You need traffic quality is be good-enough such that there's a chance for the campaign to be profitable.

What you can do: Run at around average bid ($1-2+ for most tier 2-4 geos; but please don't bid on whole dollar amounts like newbies do - more competition there - bid a bit higher), wait and cut placements, then once you think you've cut most of the biggest and worst placements, clone the camp with the same placements blacklisted, increase the bid for this new camp, then wait and cut more placements. You can repeat this by bidding progressively higher. Different bids will trigger traffic from different placements. The really premium placements won't even give you traffic at low bids. In the end you could have multiple campaigns running at different bids, targeting different placements (with overlap or not).


I would suggest to stick to the blacklisting approach for now instead of whitelisting. And try to blacklist one thing at a time to make sure everything still holds (traffic levels and CR) before making another cut.


You can do either. Although, running more DL camps to learn about optimization first would be a good way to do it, because once landers are introduced into the equation, you won't be able to test for as cheaply as you could with DL camps. It will require spending more money before you can tell whether the camp has potential or not.



Amy

(P.S. There! Now we have all the info together in a separate thread. Looking forward to your updates!)

---

Home > > Newbie Follow-Alongs