Lowered the CPA to 240 RUB


The client
A large platform of motivated traffic, which is used to promote App Store and Google Play apps in Russia and CIS
Our task
To increase the number of registrations on a website for mobile apps promotion
The service
Facebook targeted ads
The result
The number of new registrations on a website for the SMM channel increased by 91%. The conversion coefficient for registrations on a website increased from 0,3 to 6%. The CPA for SMM channel lowered to 240 RUB.

The author of the case study: Maryna Kalinkina, a specialist from the SMM department at the ADINDEX agency.

At the beginning of 2019, we decided to scale the budget (by 60%) to increase the number of new registrations on a website

At that moment, we used two types of advertising campaigns on Facebook to reach our goal:

  • The ‘Traffic’ campaign: targeted at the remarketing audience and optimized for the website’s target page views.
  • The ‘Conversions’ campaign: targeted at the cold audience and optimized for sessions of users who view more than one page on a website.
xpic_1_gRxBfpQ.width-1110.png.pagespeed.ic.ZjA3h7uKMR.pngImage 1. The most effective advertising for the ‘Traffic’ goal (on the left) and the ‘Conversion’ goal (on the right)

The Solution

The First Stage

The first thing we did is completely changed all the creatives and separated the audiences by different age segments. We also tested the audience of page subscribers’ friends. Due to creatives’ change and audience testing, we managed to reduce the CPC for the ‘Conversions’ advertising campaigns from 14 to 5 RUB during the first stage.

As a result of testing the audiences with different age targeting, we identified the one audience that was the most effective by conversion cost and CPC: it was the 25-34 age segment. We also disconnected an ineffective audience: it was the 35-44 age segment.

Image 2. The most effective audience by conversion cost and CPC

However, a complete change of the creatives didn’t improve the effectiveness of ads. Ads with new creatives didn’t get positive reviews from viewers. Their clickability also turned out to be significantly lower than expected: 1,57%.

During the next stage, we added offers containing promo campaigns in ad’s text to the most effective advertisement. We did that to display the most relevant information when a user decides to go to a website from an ad.

After we analyzed the effectiveness of ads’ text in the context of achieved goals and CPC, we understood that the ads without direct information about a promo are more effective.

xpic_3_XkSAmbC.width-1110.png.pagespeed.ic.9esnlorE-b.pngImage 3. Offer without a promo campaign in ad’s text (on the left), offer with a promo campaign (on the right)

Facebook didn’t bring the desired results during the first stage. Though the traffic of a current period grew by 277% compared to the previous period, ads still didn’t convert well.

xpic_4_OtOVrE1.width-1110.png.pagespeed.ic.w_4QiF1kYF.pngImage 4. Traffic dynamics of a current period compared to the previous one; comparison by Goal 1: Successful registration

We continued looking for a working model that would bring more results. After numerous tests and experiments, we concluded that we have to optimize campaigns for an event — new registration on a website. However, we discovered a problem: it took a lot of time for a client to install the Pixel events.

At the same time, we continued testing new audiences for Facebook targeting. These audiences had their code names (‘Startup,’ ‘Business,’ and ‘Lookalike audiences’) and were based on people who interacted with a post or an ad during the last 30 days.

The Second Stage

During the next stage, we set up an event tracking on a website (it tracked new registrations on a website) and set up optimization for this event.

The optimization brought us immediate results after the first month of implementation. The bounce rate lowered by 28%, and the number of successful registrations increased by 575% (while the budget increased by 42%). The traffic on a website decreased by 36%, which indicated a more accurate match with the target audience due to the optimization.

We integrated selling texts with USP in measurable benefit units into ads without mentioning the promo campaigns. This also played a significant role, along with budget scaling and optimization.

A buyer approaches a Unique Selling Proposition (USP) with the following question: ‘What will I get out of it?’. Our USP answers this question clearly: the price for app installment on Android is 4 RUB, on iOS — 14 RUB. This distinguishes our proposal from the competitors’ ones; furthermore, a user pays only for the result, which also makes our proposition even more appealing.

xpic_5_M4GgEsY.width-1110.png.pagespeed.ic.Jg01FTODsM.pngImage 5. The dynamics of a traffic period that was optimized for registration compared with the dynamics of a previous period; comparison by Goal 1: Successful registration

Our Results And Metrics

We achieved the desired result. Given that the budget was increased by 30% during the second stage compared to the previous period, we:

  • Increased the number of registrations by 91%
  • Increased the conversion coefficient for registration on a website from 0,3 to 6%
  • Lowered the CPA to 240 RUB

The Conclusion

Why did our strategy work? Which things should be taken into account?

  • It takes a lot of time to find a working model for any client. Only the testing of different hypotheses could help you see which segments and connections generate the results.
  • A high cost per registration of new users (240 RUB) could be relevant if you promote expensive services. For instance, in our project, the cost of 10,000 Google Play installations was 40,000 RUB.
  • Facebook optimizes very well for standard events; it is also great for finding the right target audience. This is proven by a high conversion coefficient of 6% that could compete with the coefficient of 3-4% in such powerful tools like Google and Yandex ads. We still maintain this conversion coefficient.