Home > > Newbie Follow-Alongs

---

Constantine's 40 Day Tutorial Follow-Along (13)

---

Being fairly new to all of this, I'm starting with Vortex' 40-Day Pop Tutorial 2021 Version. Should still be current now in 2022, right? Especially since it seems to have been written towards the end of last year.

I already have Voluum and PropellerAds set up from the first push traffic tests I did before joining, so there's that.

By the way, if anyone is doing the same right now - I made a post to the Mastermind board, looking for others to take this challenge together. PM me if you're interested.


First update:

It's Wednesday today and I started over the weekend. Signed up with Haka and Mobipium. Zeydoo had a bug on their registration page.

Those two networks approved me right away on Monday. Very nice! Maybe just mentioning STM was enough for them.

Haka even added me to a bunch of Click2SMS offers right away, probably the same offers everyone following this tutorial always asks for ¯\_(ツ)_/¯

Yesterday I asked my Haka AM for a list of Geos & corresponding carriers, chose Chile as the first of those that appears on the developing countries list + has over 1 mio traffic on PropellerAds.

I also asked my Mobipium AM, but he just referred me to their market search and to sort by revenue to find the best converting offers. With a bit of excel filtering I ended up with a Malaysian offer to test with PopAds (they have very few countries that match the "above 1 mio raw traffic" requirement).

I also signed up for Zeydoo as they fixed their registration bug and was approved within hours. Looks like they didn't assign me an AM though. No matter, I'll just use those other two Network's offers to continue.

---

Great to see that you´re on your way.

Many click2sms offers work good with direct linking but if you´re looking for a good LP for download/captcha click2sms offers check THIS OUT

---

Today I first set up the traffic sources and offer campaigns inside of Voluum, and then got to setting up the first pop traffic campaign in PropellerAds.

First hurdle:

The estimated CPM is orders of magnitude higher than the sample in the tutorial and I'm not sure if I did anything wrong or if PropellerAds just got a few hundred percent more expensive since the tutorial was written.

Day 9a shows an initial CPM of 0.56 for Indonesia which changed to 0.715 after platform & carrier was selected. If I select Chile in my campaign setup it shows an initial "CPM, $" of 13.518! That is 240 times higher! I mean, maybe I chose the country poorly but even when selecting Indonesia it initially shows 3.47 (6 times higher).

Does anyone have a suggestion here? I feel like a tutorial made for CPM of 0.7 with a 10$ campaign budget won't necessarily hold up when the CPM is higher than the campaign budget. Should I just bid extremely low?

---

Thank you twinaxe, I'll definitely remember this when the lesson on LPs comes up

---

Oh yeah, I forgot that the 40 days tutorial starts with direct linking

But now you already have a good starting point when you want to work with click2sms and landing pages

---

So to just get on with it I just put the campaign up yesterday with a bid of 0.7 imitating the value given in the tutorial (and something written further down in the thread on day 9a).

After this campaign has been running for almost an entire day now it seems to have received some views but it's not clear how many exactly:

- PropellerAds seems to have another bug as they show 0 views and cost on desktop along with a red popup that an error has appeared (that has been going on since yesterday).
- When I open the same page on my mobile phone, PropellerAds shows me 469 views and $0.246 cost.
- Voluum tracked 2858 visits and thinks I spent $2.11.
- Haka report shows 351 clicks.

No clue why those numbers are so all over the place.

Edit: Wait, wrong date range selected in Haka's reporting. 2,363 clicks in total. That makes more sense then.
("Impressions" as mentioned in the tutorial does not seem to exist anymore.)
/edit

0 conversions in either case. But I guess thats what I get for buying bottom of the barrel traffic. At least I have some data to continue.

---

How many offers are you testing?

2,3k clicks isn´t that much anyway, when you also splittest different offers each offer will only receive a fraction of the 2,3k clicks.

---

That was just a single offer. Yes I had a feeling that this is going very slowly but good to get confirmation.

Sidenote: PropellerAds seem to have fixed the issue, the error message disappeared and they also now show some 2.x k views.

---

One offer is way too low.

Just to give you an example what I mean, when I started to work with Haka I took ALL of their offers and tested them all in several geos to get a better impression what works and what not.

I don´t really recommend that approach because it can become pretty pricey, just want to point out that it´s better to test few offers too much than few offers too less

---

Hi @constantine! Thanks for starting this follow-along!

You're off to a good start, asking some great questions. A few points:

1)Running one offer at a time is completely fine for beginners. @twinaxe is right - once you know what you're doing, testing massively would be the way to go. For now though, starting slow would be fine.

Beginners will make lots of mistakes, so if you set up too many campaigns, you'll make the same mistakes multiple times, which will cost you more then if you were to launch more slowly.

However, launching more campaigns at a time will increase your chances of hitting on something promising.

So - pros and cons. Just go with what you're comfortable with for now.

2)As for the bid: Are you referring to the recommended bids from Propeller? Those are typically way too high than what you could be bidding to still get decent-quality traffic. For PropellerAds specifically, I find that their traffic quality is decent even at lower bids, so don't be afraid to test. JUST - keep in mind that if the geo doesn't have a ton of traffic to begin with, bidding low may not get you much or at all.

But if you want to save on budget and don't mind waiting, you can bid low when testing offers on Propellerads. It can take longer to test each offer since traffic can be real slow, but if you don't mind the wait, you can save money. And you can keep yourself busy by launching more campaigns in the meantime (instead of twiddling your thumbs to wait for stats to roll in for a single campaign).

3)The amount of traffic you'll see in your stats on various platforms - traffic source vs. tracker vs. aff network - will usually be different. Something called "clickloss" is the most common cause of such discrepancies:

https://stmforum.com/forum/showthrea...ight=clickloss

If the discrepancy is consistently excessive - let's say 50%+ - then you'd need to do some investigation to see whether there's something wrong with your setup (settings for your domains / CDN / webhost / tracker / etc.)

Have fun and I look forward to your next update!



Amy

---

First of all, a big belated thank you Amy for explaining a bit further. Regarding (2), yes I meant PropellerAds. I actually found a solution to still get their recommended bid. See further down

UPDATE

After taking a bit of a (forced) break during a long vacation with the family, I spent quite a bit of time figuring out how to best build & deploy landers, and also honestly nerding around a bit too much experimenting with Voluum and PropellerAds APIs, doing some of the challenge steps via code that definitely didn't need it.

Some learnings that might be helpful to others:

1. When deploying to AWS and you update your landers code, you need to cache invalidate your CloudFront distribution!

This drove me nuts until I figured out this is why I wasn't seeing any changes. It might be just because I'm deploying via AWS CodePipeline from github pushes, but I actually assume that the problem should also be there when uploading directly via FTP.

2. PropellerAds' API exposes the recommended bids for any targeting combination, even though they only display the max bid in the UI.

You don't even have to do your own API calls for that! When clicking around their UI and setting up your campaign, requests are made in the background to `inventoryPlanner/`. Open the developer console(Cmd+Alt+i for Chrome on Mac), switch to the Network tab, adjust the targeting options, look for the latest call to inventoryPlanner, click that, go to "Preview" and look for the `recommended_bid`.

You can call this endpoint from anywhere though. That endpoint is under the `partners` subdomain even though the official API is under `https://ssp-api.propellerads.com/v5/` and doesn't list the endpoint in the docs - but it works just the same with your API key as Bearer token.
[hackerman.jpg]

3. If you're ever tempted to write API integrations with PropellerAds, be aware that there are some targeting parameters they don't really provide mapping endpoints for. Like languages.

It's not as easy as it first seems. If you are trying to do it, you can PM me and I can potentially give you some of the data maps you'll need (scraped them from their frontend).

--

In any case, I screwed around enough and just enabled my first campaigns that use landers. Let's see

---

That's a neat trick! Thanks so much for sharing! But I have a feeling those bids would still be higher than necessary, especially when doing initial testing (as opposed to scaling). For that I would definitely recommend bidding low if you wouldn't mind the wait.

Knowing how to code is a big asset! It can be fun figuring out ways to automate the testing/optimization process, sort/analyze data etc.


Amy

---

https://medium.com/fullstackai/aws-c...s-49c1fd3a3c31

---

Home > > Newbie Follow-Alongs