Optimization in Action: When Your Campaign Is Not Doing Its Best [Case Study]
This post is also available in:
PT 
ES
What’s this? Another insightful case study? Yes, it is!
We keep sharing the best case studies from our Follow-Along contest we organized together with the Afflift forum, and this one is a must-read. Our partner, nicknamed Viccarterann on the forum, did very hard work optimizing his campaigns, but they still seemed hopeless.
And what happened in the end? Did he manage to reach positive ROI and profit?
Read on to find out!
About the Campaign
I haven’t been getting very good results lately.
This is how Viccarterann started his story. He was more than sure that PropellerAds was giving him high-quality traffic — he’d already had some successful campaigns here. However, choosing the right offer has been a bit of a struggle recently.
So, Viccarterann was determined to find the offer that would convert, focus on PropellerAds as the main source, and spend most of the effort on campaign optimization.
Here is what he decided to work with:
| Traffic Source: PropellerAds Tracking Tool: BeMob Affiliate Network: Golden Goose, Zeydoo Type of Offer: Utilities, Fin Surveys, Sweepstakes Offer Payout: From $0.12 to $1.5 Traffic Type: Popunder ads Dates: 20/05-28/06/2022 |
Choosing the Offer
Vicarterann didn’t want the wrong offers to become a stumbling block anymore. This is why he decided to speak to the Golden Goose and Zeydoo managers first. So, he came to them with a very precise request on what he needed:
- Active profitable offers
- Payment of less than $1 to not spend much on tests
- Easy conversion flows: SOI, 1-Click, 2-Click, etc.
While Vicarterann was waiting for the managers’ reply, he decided to try one mainstream offer from Golden Goose. It seemed to be really profitable — and a pretty easy 2ClickFlow inspired even more trust.
And here was the first lesson learned: the CPA network managers know their business ? They can give really useful advice on how worthy a particular offer is.
The manager warned Vicarterann that this offer might not convert. However, Vicarterann still launched two campaigns for it — and, in eleven days, got the following results:
It was not the best choice, alas.
Vicarterann stopped these campaigns and switched to Sweepstakes offers recommended by Zeydoo managers. Besides, he decided to drop the idea of testing the 2ClickFlow — at least until he could see positive ROI for easier conversions.
Campaign Settings
So, it was time to set up new campaigns with Sweepstakes. The CPA manager recommended top GEOs for the offers (Ghana, Kenya, and more), which made things much easier — but there was still plenty of work to be done.
Here is what Vicarterann did first:
- Went to the Spypop spy tool to get some inspiration for his landing pages
- Set up landing pages: changed the copy, images, CTAs, and languages according to the chosen GEOs. Overall, he created six versions for tests
- Installed the Backbutton script with the Zeydoo Smartlink
- Added a ProPush Smart Tag for extra monetization with push notifications. Now the landing pages could collect push notifications subscribers — and ProPush would pay for them!
Here is a landing page similar to what Vicarterann used:
Once the preparation was over, Vicarterann launched three campaigns. Here is how he arranged them:
| Campaign 25 | GEO 1 | Landing Pages 1,2,3 |
| Campaign 26 | GEO 2 | Landing Pages 1,2,3 |
| Campaign 27 | GEO 3 | Landing Pages 4,5,6 |
Two of the GEOs were recommended by Vicarterann’s manager in Zeydoo, and he chose the last one himself — it worked pretty well with similar campaigns in the past.
And a bit more of the details:
- Pricing Model: CPA GOAL 2.0 with 50-60% of the offer payout
- Initial Budget for Test: Approximately $10 (150-160 times the offer payout)
- Settings: scroll the gallery below
Campaign Flow
After a 4-day test, Vicarterann shared his first results:
The conclusion he made was not really optimistic:
Looking at the ROI of these campaigns, there was little potential.
The campaigns were doing pretty fine right after the launch — but started dying after a few hours without spending even $3. What happened?
Vicarterann supposed that there was too harsh competition for the chosen GEOs. Even increasing a bid did not help.
However, he didn’t mean to surrender! Now it was time for serious optimization.
Campaign Optimization
Vicarterann decided to try optimizing his creatives and GEOs — it seemed the most obvious solution at that moment. So, here is what he did:
- Created two new improved versions of his best converting landing page. It’s the LP13 in the screenshot: the one with the best ROI. The LP1, by the way, had a better CR — so Vicarterann just kept watching how it performed.
- Tested three new GEOs. Vicarterann went back to the list of the top GEOs from his CPA network manager and picked three new ones from it.
What was the result of this? Here are the new stats for the three new GEOs:
As you see, nothing has really changed. And, what is more, not only the ROI was disappointing. The spending also got stuck: all three campaigns haven’t managed to spend even $5 in three days.
Vicarterann had no other idea but to make the next, much more drastic optimization step:
What can I do? Leave CPA 2.0 Goal and switch to SmartCPM.
Was it the right decision? It seems so — just check out the new stats:
Looks like we’ve finally found something to work with! A campaign that has potential (-47.87% ROI).
So, Vicarterann decided to focus on this single campaign. And, of course, he started with optimization again.
First, he compared how the landing pages performed. It seemed like the LP13 was still a winner:
Still, Vicarterann double-checked it using the Bayesian A/B testing tool.
- Blue line: LP4
- Violet line: LP13
- Red line: LP3
Yes, it was not more than obvious: Landing Page 13 was the best-performing one.
Vicarterann’s next steps included:
- Cutting the Zones that spend more than x2 of the bid
- Deleting all the browsers but Chrome
- Cutting all OS versions that spend the most but convert the least
Want to see the optimization details? Scroll the gallery, and see what Vicarterann has changed in his Zone ID, Browser, and iOS settings.
Finally, Vicarterann increased the bid — it was now $20.
Here are the results:
Well, it wasn’t something he expected. However, Vicarterann was already determined to get green numbers within this Follow Along!
So, he decided to try a completely different approach.
The Last New Campaign
The new strategy was very simple. Its point was to:
- Check all the latest campaigns and find the two most successful ones
- Combine them into a single CPA Goal campaign
- Set a very specific targeting: only a single browser, carrier, OS, etc.
So, here we go: the campaigns with the best CPM data were now supposed to bring income together:
And, meet the new CPA Goal campaign’s first results:
Finally we see the green and a ROI of over 100%!
Final Campaign Results
After a couple of days, the campaign kept showing profits!
Due to such specific targeting, the campaign couldn’t get much traffic — but it was all green, after all.
Vicarterann didn’t want to stop at this point, though. In the future, he decided to test more Sweepstakes offers and master his campaigns in this vertical.
I will start from scratch with what I have learned.
Not much of a profit — but plenty of useful lessons. Isn’t it already nice?
Bonus Part: Extra Monetization
Working with the offers turned out to be much of a struggle for Vicarterann. But what about his extra monetization efforts? Were they the same complicated?
We remind you: Vicarterann installed the ProPush Smart Tag and the BackButton. Let’s see what it brought to him in the end.
ProPush.me
Here are the daily stats from ProPush (scroll the gallery). Within the whole campaign flow, Vicarterann didn’t touch anything in ProPush settings after he added the tag.
And it worked! Although it was only $6 of a profit, it was completely free of time, effort, and investment:
BackButton
The Backbutton started showing great results from the very beginning. Even when the main offers seemed to be a failure, the Backbutton script with the Zeydoo direct link was the only ‘green thing’:
After some time, Vicarterann split the Backbutton campaign between Zeydoo and Monetizer — and they were both fine:
During the Follow-Along contest, Vicarterann managed to make $19,62 for ‘free,’ as he called it himself. He was pretty impressed with it!
The best thing about it is that it was passive income. Even after cutting all your campaigns, it still brings you profit — incredible!
Comments from Konstantinos Kafkalias, Senior Account Strategist at PropellerAds
Hello everyone! Let’s thank Viccarterann for a detailed and informative case study and let’s go see together the highlights, and answer a few questions!
- Choosing the offer:
It’s cool that Viccarterann searched in more than one Affiliate Network for the best match for Propellerads traffic source. As mentioned above, it’s always a good idea to check with your Account Manager about an offer, and the potential it has with the traffic source.
- Setting Up Campaigns:
Viccarterann is definitely not giving up easily, and this shows the way to a successful affiliate! We gotta work for it, guys! He chose a few different offers until he managed to nail it!
He chose the CPA Goal model for testing. I would also test the offers on CPA Goal, as it gives the auto-optimization function, as well as the option to get traffic from all the zones (sources) – cheaper and more expensive.
Let me explain a bit more here:
When launching a CPA Goal campaign, it will be competing with all campaigns (CPA Goal, CPM, SCPM) that have the same targeting. The CPA Goal Rate we set, is responsible for the amounts of traffic the algorithm can send.
For example, an offer with the rate of $50 will receive significantly more traffic than one with the rate of $1, because the creator of the campaign is willing to pay more money for each conversion. So the algorithm has more space for testing zones.
In other words, for CPA Goal campaigns, the higher rate will not exclusively mean more expensive zones, and the lower rate will not exclusively mean less expensive zones. Higher rates will get you more traffic volumes, though, for sure.
Viccarterann expresses the confusion with his example where the higher rate campaign gets lower CR, and the lower rate campaign gets traffic from cheaper zones, which ends up to higher CR and better results.
Looking at the statistics, this phenomenon was coincidentally created by starting two completely same campaigns, same targeting, but at different rates. Here is what was going on:
- 1. The algorithm started testing simultaneously, and the lower rate campaign happened to get more conversions during the test, which then drove more traffic to it and ended up being more successful.
- 2. The two campaigns were directly competing with each other in a negative way. Normally, they are competing with thousands more campaigns, but in this case they were spoiling each other.
TIP: It’s better to choose one CPA Goal rate for the offer, and the recommended rate from our side would be at least 70% of the rate you guys are being paid by the Affiliate Network. I know, I sound greedy, but stats are stats, and they never lie 🙂 One CPA Goal campaign should be tested on its own, not have two-three of the same offer, with different rates. It gets messy 😉
- Optimization:
It took some time to find an optimizable offer, but well done! And yes, it’s always a very good idea to analyze campaigns by OS, OS Versions, Browsers, and all the rest of the small details, in order to narrow down to what really worked.
And then, launch more! Let it be CPM, with best performing zones, or CPA Goal with more narrow targeting, or of course, both!
You know it works, go make money!
Excellent touch with Propush.me and extra monetization. Please get in touch with us to help you very easily set it up for all your campaigns.
Once again, a big thank you to Viccarteran for the time, effort, and details of his case study.
Until the next one, guys!
We think there’s much to discuss here. Agree?
Welcome to our Telegram chat!
Or skip the discussion and go straight to the action!
