Testing is a critical phase of any advertising campaign. It helps identify the most effective advertising combinations suitable for scaling.
In this case study, the ProContext team shares how they tested two distinct audiences to promote the No One mobile app, the results they achieved, and the insights gained from the test.
Campaign goals
- 1Improve promotion results compared to the control group*:
- reduce the cost per app install (CPI) by 40%;
- reduce the cost per in-app order (CPO) by 30%.
* The control group was an audience built during a previous campaign using keyword targeting.
- 2Compare the performance of two audiences and identify the highest-converting one.
Mechanics
Campaign duration: from January 13 to February 2, 2025.
What was advertised: iOS mobile app.
Target action: app install (with additional tracking of cost per in-app order).
Bidding strategy: lowest cost.
The budget was optimized at the ad set level with daily spending limits to maintain control during testing and enable reallocation based on performance.
The promotion started by testing two audiences. For this, a single campaign was created with two separate ad sets.
Targeting
Group 1: users who had made purchases on the client’s website.
Group 2: users who had made purchases in the client’s offline stores.
User lists were exported from the client’s CRM system and uploaded to VK Ads. No additional targeting filters were applied to No One’s broad target audience, with ads shown to men and women aged 18–75 and residing in Russia.
Ad creatives
Ad banners featured photographs of clothing, footwear, and accessories. Discount information was added to capture user attention and highlight the value of purchases.
Results
Each of the two groups consisted of people who had previously made purchases either on No One’s website or in their offline stores. Despite both groups being shown identical offers and banners, the results varied significantly between them.
The group comprising users who had made purchases on the client’s website demonstrated the following results compared to the control group:
- cost per app install decreased by 42%;
- cost per in-app order decreased by 79%.
The group comprising users who had made purchases in the client’s offline stores showed the following results compared to the control group:
- cost per app install decreased by 50%;
- cost per in-app order decreased by 33%.
Interpreting the results
- 1Users who previously purchased on the website may already have had the mobile app installed, given their established habit of ordering online. Consequently, while their CPI saw a smaller reduction, the CPO decreased far more significantly compared to the offline purchaser group.
- 2The offline purchaser group was less likely to have the mobile app installed. We also hypothesize these users prefer buying clothing and footwear in physical stores, where they can evaluate items in person. This likely contributed to the more substantial reduction in CPI, but a smaller reduction in CPO compared to the online purchaser group.
Future plans
We plan to add both audiences to our core advertising campaign and refresh them at regular intervals. We will also conduct further testing:
- we will set up retargeting for users who have made purchases via the mobile app;
- we will create a new audience comprising users who have registered on the website but have not yet made a purchase.
Contents
