One of the most valuable things when starting out is an example of “how to”. In this article I will walk you through my first profitable campaign, and add what I would do differently now with more experience and knowledge.
I had been trying quite a few app installs on mobile display traffic on Decisive, the go to traffic source back then. Note: They are dead now, they pivoted to something else. If you can sign up for the traffic source, don’t! Some people are still having trouble getting refunded for what they had left before they suddenly closed shop.
I initially direct linked all the offers, from the banner to Google Play page. After 4-5 apps tested, I ran into DU Battery Saver offer. It was recommended to me by an affiliate manager, I tried it out in several geos and it showed some promise in Indonesia, aka it got conversions when direct linking. ROI was still around -85% but I had data to work with. That was a huge step.
Here is a sample of my initial banners. I tried MANY variations, more than I should have in hindsight, so I will only show you a sample of the most important batches.
They are all animated GIFs, 320×50 size, since it was the highest volume banner size. I tried different headlines, different battery icons, different CTAs and also border vs no border.
These are the banners that brought some conversions, some data to start optimizing! The first thing I did then was buy a VPS subscription and start adding a landing page to my tests. Here’s one of the variants.
As you can see, text is very long, image quality is not great. I had a few other versions with different colours but all had the same text. At the same time as adding a landing page, I also started cutting placements that didn’t work so well, and cut out the worst banners, using this statistical significance calculator. I cut based on conversions, NOT CTR/clicks. Conversions matter, don’t focus on CTR, especially not initially.
Adding the landing page brought some banner + lander + offer combo close to -50%, alongside cutting some placements. I kept trying more and more banners, sample of them below:
The biggest trend here is that one of the battery icons clearly outperformed the others, and that lighter backgrounds performed better than dark.
I knew the server is not ideal, since it was located in the UK, while I was targeting users from Indonesia. I researched what solutions I could use and found that the best performance for the price seems to be Amazon S3 for hosting static pages and adding CloudFront as the CDN for it. It just so happens that we still use this setup now when getting >2M visitors to our pages daily.
Changing to Amazon S3 + CloudFront for landing page hosting put some banner designs at 0% ROI. I then ran more iterations of banners, following the idea that light background works better. I tried no-CTA banners vs CTA and also more battery animations. The main goal was to increase CTR without dropping CV. The conversion rate was very good, >7%, but the CTR was below 0.7%. It means clicks were expensive.
This batch already showed some profit. The next step was to adjust the bid and increase daily budget to the max, since this was already making money, so the risk of losing was reasonably low.
After decreasing the bid by about 10-15%, and trying out new banners, the campaign reached the point where on $200/day spent it made $300/day revenue. My absolute best banner, that lasted for quite some time is this one:
Adding something that looks like an arrow encourages clicks. I just had to think of a way to make it not look like an actual “click this banner” arrow… So I made it pull the battery.
This whole process took approximately 14 days, including coming up with headlines, waiting times to get the new banners approved, designing banners. I was grinding out new banners all day, adjusting landing page colours, trying to apply some of the optimization tips I learned. It was a non-stop grind, I was even dreaming banners and landers.
Soon after, Decisive started having technical issues and I moved my campaign to Go2Mobi. I managed to scale it more there by using more banner sizes and more geos. Once you have a traffic type + angle + offer match, you can definitely scale to more banner sizes. I strongly encourage that when you have banner campaigns, you try more formats.
The next way to scale is more geos on an international offer, and using the same traffic source.
Finally, try more exchanges within the same DSP. This is not certain to instantly work but it’s definitely worth a test where you have an exchange with decent volume.
Would this campaign make money now? Maybe, if you find the right traffic source. I think it’s unlikely you can replicate it on Go2Mobi. The offer is old, and I ran it for a long time there, until it stopped converting well.
Some things I would do differently now, having more experience:
1. Have a better hosting solution from the beginning. It can make more than 30% difference in ROI, easily. Imagine having a worse hosting than my VPS even. You end up having 0% chance of success.
2. Try fewer banners, focus more on finding an angle that works.
3. Shorter copy on the landing page. Say the same, with fewer words. Apps are an impulse, no need to tell a long story.
4. Split test the same offer on more affiliate networks. I ran it with only one affiliate network for too long. I should have tested as many affiliate networks as possible.
Keep in mind I was a total newbie. This was the first campaign I ever tried with a landing page. The thing that made it work for me was testing non-stop. If this offer hadn’t worked, I would have tried more offers, new angles, learn more everyday.
Emanuel Cinca, aka Manu
My story starts in Lipova, Romania, a town of about 10,000 inhabitants, where I was born in a working class family. Growing up in a family that had a shaky financial base and relied solely on month to month wages, and sometimes on short term loans from friends, I knew that I a) didn’t want to get a regular job EVER, and b) I want to have true financial independence.
Fast forward to my high school days, I got admitted into the number one high school of Arad county, “Moise Nicoara” National College. Don’t ask me why they call it a college, that’s just the translation from Romanian. I had a passion for technology and computers, so it seemed natural that I wanted to get into the so called informatics class.
The 4 years here taught me that:
1. Where you come from has little to no impact on your ability to learn, your skills and their value
2. Being number one high school is overrated
3. In the standard educational system, you are not allowed to excel at something, until you have proven to be above average at everything else.
4. Grades are an extremely flawed measurement of someone’s skills
5. I know myself better than anyone else, so I should make my decisions, not anyone else, and I should be OK with the consequences.
These learnings will be reflected in some of my posts here.
After high school, I quit University after 2 months in… TWICE! One time I quit Computers and Information Technology, and the second time I quit Business Administration.
During these years I also developed a passion for poker. I was a content producer at the world’s biggest online poker school for 2 years, and I wrote a book with a Maths professor from Nottingham and TT, who also blogs here. The book is called The Education of a Modern Poker Player.
I moved to Vienna in 2013, at TT’s initiative, and had several failed ventures together, all leading up to what is now Adefy.
I have a keen interest in psychology and in data, which makes performance marketing such a great industry for me. I blog about these topics, sometimes adding education, personal development, and management to the mix.
Interested in: Psychology. Marketing. Management. Investing. Startups. Entrepreneurship.
Hobbies: Reading. Fitness. Football. Travelling.
Connect with me:
Latest posts by Emanuel Cinca (see all)
- WHAT THE AFF has been going on with me? - October 27, 2018
- Affiliate World Europe 2018 – BARCELONA! - July 28, 2018
- Introducing The #1 Daily Affiliate Newsletter – What The Aff… - April 23, 2018