Everyone Should Know How AI-Powered Google Ads Transform Marketing
In today’s hyper-competitive food delivery market, scaling operations while maintaining profitability is a formidable challenge. At Google Marketing Live 2025, the latest advancements in AI-powered advertising solutions provided a game-changing opportunity. With tools like Performance Max, AI Max for Search, and Smart Bidding Exploration, marketer could leverage predictive analytics and automation to refine its marketing approach. Delivery Hero, a global leader that operates in over 70 countries and serves more than two billion people, confronts this challenge on a daily basis. With over 1.5 million restaurant partners and local vendors under its ecosystem, the company must continuously acquire new customers who not only make a one-time purchase but become long-term users. The stakes are high—more marketing spend doesn’t always translate to higher profits. In fact, as Gökhan Reisoglu, Global Head of SEM at Delivery Hero, notes, "At a certain point, more spending can actually lead to less profit."
To tackle this, Delivery Hero has embraced a data-driven approach, leveraging Google Analytics 4 (GA4), first-party data (1P data), and Google AI to optimize customer acquisition. This strategy has allowed the company to strike a delicate balance between reach and quality, ensuring sustainable growth without sacrificing efficiency. In this article, we’ll explore how Delivery Hero’s innovative use of AI-powered Google Ads is reshaping performance marketing in the food delivery industry.
I. The Role of First-Party Data and Google AI in Customer Acquisition
Delivery Hero’s journey toward AI-driven marketing began with a critical realization: not all first-time customers are equally valuable. Some users make a single purchase and never return, while others develop into high-lifetime-value (LTV) patrons. To distinguish between these segments, the company turned to first-party data—information collected directly from user interactions—and integrated it with GA4 and Google Advertising for predictive analytics. This allowed Delivery Hero to identify patterns in customer behavior, such as purchase frequency, order value, and engagement levels.
Google AI played a pivotal role in processing this data at scale. By feeding real-time insights into Smart Bidding algorithms, Delivery Hero could dynamically adjust bids based on predicted customer value. Emily Sutti, Head of Performance Marketing Solutions at Delivery Hero, emphasizes the importance of speed in this process: "Speed is more important than the accuracy of a model." The faster the AI could analyze and act on data, the more efficiently Google Ads could be optimized. For instance, AI models prioritized users likely to place repeat orders, allowing Delivery Hero to allocate budgets more effectively. This shift from volume-focused to value-driven strategies marked a turning point in the company’s marketing efficiency.
II. A Three-Stage Framework for AI-Driven Campaign Optimization
Delivery Hero’s collaboration with Google resulted in a three-tiered framework tailored to different market maturities and data availability. At the foundational level (Level 1), GA4 was used to segment audiences and control cost-per-acquisition (CPA) within Google Advertising. By identifying high-value user groups, Delivery Hero could refine targeting without overpaying for low-quality leads.
Level 2 introduced return on ad spend (ROAS) optimization, where short-term gross merchandise value (GMV)—the revenue from a customer’s initial orders—was correlated with long-term LTV. This allowed Google Ads campaigns to prioritize users who not only spent more upfront but were also likely to remain active over time.
The most advanced stage (Level 3) involved predictive modeling. Using Google Cloud’s computing power, Delivery Hero trained AI models to forecast customer value and adjust bids accordingly. For example, the Latin American brand PedidosYa combined these predictions with Smart Bidding, resulting in twice the expected user value compared to traditional CPA campaigns. Similarly, Foodpanda in Southeast Asia saw a 52% ROAS uplift by integrating first-party signals into Google Ads, while Talabat in the Middle East achieved a 100% ROAS increase through dynamic bidding. These achievements highlighted the capability of AI to convert raw data into actionable marketing insights.
III. Case Studies: Regional Success Stories
The real-world impact of Delivery Hero’s AI strategy is best illustrated through regional case studies. In Latin America, PedidosYa’s integration of predictive models and Smart Bidding not only doubled user value but also increased the proportion of high-LTV customers by 13 percentage points—all while maintaining CPA levels. This showed that AI was able to improve quality while maintaining reach.
In Southeast Asia, Foodpanda’s use of first-party data allowed for hyper-personalized Google Ads campaigns. By analyzing user behavior—such as preferred cuisines and order times—the brand delivered tailored ads that resonated with local tastes. The result was a significant ROAS boost, proving that regional customization is key to global scalability.
Meanwhile, Talabat’s dynamic bidding strategy in the Middle East showcased the flexibility of AI. By enabling Google’s algorithms to automatically adjust bids according to real-time data, Talabat achieved a 100% increase in ROAS. Djordje Petrovic, Google Senior Growth Manager, noted an unexpected benefit: "The composition of user groups became more similar when campaigns were optimized for long-term value." This convergence of reach and quality highlighted AI’s ability to balance seemingly opposing objectives.
IV. Balancing Reach and Quality: The Power of AI Autonomy
One of the most counterintuitive findings from Delivery Hero’s AI experiments was that less manual intervention often yielded better results. Traditionally, marketers exclude certain user segments—such as low-spending demographics—to improve profitability. However, Delivery Hero discovered that allowing AI full autonomy led to a natural equilibrium between reach and quality.
For instance, when campaigns were switched to predicted ROAS (pROAS) bidding, the algorithm independently determined when to bid higher for potential high-value users. This eliminated guesswork and reduced wasted spend on ineffective targeting. Gökhan Reisoglu admitted, "We actually expected to acquire fewer new customers. The fact that one campaign can simultaneously increase the number of new customers and profits is unusual." This revelation reinforced the idea that AI-driven marketing isn’t just about efficiency—it’s about unlocking untapped potential.
This principle aligns with Topkee’s AI-powered Google Advertising services, where autonomy and precision drive performance. For instance, Topkee’s TTO tool automates data tracking and conversion event synchronization, enabling advertisers to delegate bid adjustments and budget allocation to AI-driven systems. By associating multiple tag IDs and leveraging smart bidding strategies, the tool ensures ads reach the right audiences without manual segmentation biases. Similarly, their keyword research methodology combines AI analysis with competitor benchmarking to dynamically expand keyword lists, optimizing reach while maintaining relevance. Topkee’s remarketing strategies also exemplify this balance. Using TTO attribution tools, AI segments users based on behavior and delivers personalized ads. Data shows such targeted approaches increase purchase likelihood by over 70%.
V. Future Directions and Scalability
Looking ahead, Delivery Hero plans to refine its AI strategies further. Adaptive ROAS models will be customized per market, accounting for factors like local competition and consumer behavior. In regions with low market penetration, the focus will remain on volume, while mature markets will prioritize profitability. The company also aims to expand its AI-powered Google Ads engine globally, ensuring consistent performance across diverse markets.
Leo Salani, Director of Performance Marketing at PedidosYa, describes this evolution as an ongoing process: "The key to success lies in translating customer value insights into a format usable by Google AI." As AI technology advances, Delivery Hero’s ability to predict and influence customer behavior will only improve, solidifying its position as a leader in data-driven marketing.
Topkee’s keyword research and expansion strategies—supported by smart bidding and broad matching—will prioritize volume growth by targeting high-potential audiences. In mature markets, the focus will shift toward profitability, utilizing Topkee’s TTO CDP for precise conversion tracking and automated budget allocation across multiple ad accounts. Topkee’s AI-generated ad creatives, tailored to market trends and product specifics, ensure messaging resonates locally while maintaining global brand consistency. Additionally, Topkee’s periodic ad reports (covering ROI, conversion quality, and budget efficiency) will empower brands to iterate campaigns dynamically.
Conclusion
Delivery Hero’s success story offers valuable lessons for marketers navigating the AI revolution. By combining first-party data with Google AI, the company has achieved what once seemed impossible: scaling customer acquisition without sacrificing quality. The results speak for themselves—higher ROAS, increased customer lifetime value, and sustainable growth in a cutthroat industry.
For businesses looking to replicate this success, the path is clear: embrace AI-powered Google Advertising tools, prioritize first-party data, and trust in algorithmic autonomy. As the marketing landscape continues to evolve, those who harness the power of AI will thrive. If you’re ready to transform your Google Ads strategy, consider consulting with experts to explore how these innovations can work for your brand.
