OMD reduces Google Ads CPA for Zain KSA by 23%

Key results:

• 23% decrease in CPA for Google Performance Max campaigns
• 202% increase in visitors with a high intention to convert through another browser, device or cookie

About OMD

OMD is one of the world’s largest media networks with more than 12,000 employees around the globe. OMD’s mission is to help its clients make better decisions faster to achieve outstanding business outcomes.

About Zain KSA

Zain KSA is a leading telecom provider established in Saudi Arabia on August 2008 as a listed company. Due to its highly developed infrastructure, the company was successful in establishing itself as a reliable telecom operator and a digital service provider. The services include telecom services, 5G networks, digital payment services, cloud computing, IoT solutions, fiber services, drones, and many others.

The challenge:

#1: Measurement and budget reallocation

OMD is running marketing activities for Zain KSA across five advertising platforms going beyond Google Ads and Facebook, including Snapchat, Twitter, and Tiktok which are mostly responsible for generating awareness about the advertiser’s products, and usually do not generate enough conversions when analysed by traditional attribution models such as Last-Non-Direct-Click.

Due to this, it was challenging for OMD and Zain KSA to accurately measure the effectiveness of their upper-funnel activities and to properly allocate the budget across their marketing mix and in-channel campaign structure.

#2: Scaling and optimisation of Google Ads Performance Max campaigns

OMD began using Performance Max campaigns, the newest type of Google Ads. This version of Google Ads gives the Google bidding engines full flexibility to optimise and target accordingly.

The challenge, however, is feeding these campaigns with high-quality data. Performance Max campaigns rely heavily on conversions for effective optimisation.

If the ad click does not result in a directly attributed conversion, no feedback signal will be provided to the Google Ads platform to learn and optimise its bidding algorithms. The lack of such conversion signals prevents Performance Max from showing its best performance and limits opportunities for efficient scaling.

Let’s see how SegmentStream helped to solve both of these challenges.

SegmentStream — solution overview:

SegmentStream is a Conversion Modelling Platform that provides a next-generation solution to outdated attribution and conversion tracking tools.

Unlike most marketing attribution tools that assign a value from the conversion retrospectively, SegmentStream evaluates each website session by predicting the user’s probability to convert in the future. When this probability is sufficient, the platform creates a Modelled Conversion that can be used for measurement and optimisation.

This guarantees that each traffic source immediately gets the value it deserves, even if a future conversion happens from another device, browser, or cookie.

Now, let’s see how this approach works in Zain KSA’s case.

SegmentStream solution architecture for Zain

Step 1: Collect website behavioural data

SegmentStream Conversion Modelling Platform uses Machine Learning algorithms to assess the impact of each website visit.

Before building the model, all behavioural events should be collected from the website first. Using SegmentStream’s platform’s “Data Sources” functionality, the team at OMD started to collect real-time data about the users’ behaviour from Google Analytics into their own BigQuery data warehouse.

SegmentStream admin panel with connected data sources

Step 2: Collect ad costs from marketing channels

The setup of data collection usually takes little time because SegmentStream already has turn-key integrations with the most popular advertising platforms.

Once the ad platform is connected, SegmentStream automatically imports all the data about ad clicks, costs, impressions, etc. into the client’s Google BigQuery.

Step 3: Conversion Modelling

SegmentStream Conversion Modelling Platform carefully assesses all actions every visitor takes and measures the users’ probability to convert in the future. Once probability is sufficient, SegmentStream creates a Modelled Conversion that can be used both for measurement and optimisation.

After a Modelled Conversion is created, it is immediately attributed to the traffic source that started this website session. Therefore, it is possible to understand which traffic sources generated more valuable traffic to the website, and send this information back to ad platforms to help them drive high-quality traffic.

Modelled Conversions in SegmentStream

To dig deeper into the topic of Conversion Modelling, read our guide for performance marketers.

Step 4: Reporting and analysis

SegmentStream provided OMD with a more holistic view of their campaigns’ performance. Knowing that the final conversions often cannot be observed due to cookie limitations, the team at SegmentStream offered OMD an alternative approach to evaluate campaign performance. With Conversion Modelling Platform, the OMD team was able to see which channels and campaigns drove the most visits with a high intent to convert in the future. Such website visitors very likely will convert from another browser, device or cookie, even though it may not be possible to deterministically attribute such conversion to the initial traffic source. For example, the OMD team was able to discover that they should redistribute budgets to Google Search and Google Performance Max since they were bringing a lot of valuable visits with a high chance to convert in the future and more cost-efficiently.

Step 5: Conversion export and Google Ads optimisation

Based on the reporting insights, OMD decided to scale Performance Max. To ensure the best efficiency, they set up automated export of SegmentStream’s Modelled Conversions to Google Ads using turn-key server-to-server integration. Once Modelled Conversions started to flow into Google Ads, OMD enabled them for optimisation across two Performance Max campaigns, promoting Zain’s two most popular voice postpaid packages: Shabab and Black.

The results:

1. 23% decrease in CPA of Google Ads (Performance Max campaigns). By using SegmentStream’s Modelled Conversions for Google Ads optimisation, OMD saw a 23.3% improvement of CPA in Performance Max campaigns.
2. 202% more users with high intent to convert. While optimisation using Modelled Conversions generated more directly attributed conversions, it also helped to attract more users with a high intent to convert in the future at a cost-efficient rate — OMD recorded a 202% increase in Modelled Conversions, indicative of a quality user that will later convert through another browser, device, or cookie, at 76% lower costs.

Customer Testimonial:

SegmentStream provides an innovative tool to measure and optimise performance marketing activities.

Thanks to the platform, we were able to optimise Performance Max campaigns in Google Ads achieving outstanding efficiency, as well as understanding the impact of awareness campaigns in paid social channels.

We are very confident in the prospects of our collaboration with SegmentStream and already see that this is producing some exciting results. We look forward to developing our partnership.

— George Achkouty, Head of Digital at OMD

The partnership with SegmentStream has been a real success. Prior to SegmentStream we knew there were opportunities for optimisation efficiencies across our media campaigns, however, we couldn’t quite pin-point exactly where they were.

Modern cookie restrictions meant we were unable to understand the real impact of our investment, and the publishers were not receiving an adequate number of conversion signals for optimum optimisation. The results speak for themselves. Within a short period of time, we could see the incremental impact SegmentStream was having on our business.

We’re thankful to SegmemStream as a technology, but also a service partner. Their in-house expertise is second to none. We look forward to continued growth.

— Abdullah Medhat, Zain KSA Channels Senior Manager