Article

6min read

Unify GA4 with BigQuery to Strengthen Experiments

In today’s digital landscape, data-driven choices are essential for staying competitive, with experimentation as a critical driver of innovation. To support this, we recently hosted a webinar with experts from Google Cloud and AdSwerve, focusing on how Google Analytics 4 (GA4) and BigQuery can enhance experimentation strategies. GA4 is essential for all marketing teams, providing advanced analytics that, when combined with BigQuery’s data consolidation capabilities, enables more effective testing, personalization, and digital optimization.

Meet the panel

Taige Eoff, Cloud Data AI Lead at Google, has been at Google for twelve years, leading data and AI initiatives for cloud marketing. Taige focuses on developing scalable solutions that support partners like AB Tasty and AdSwerve in optimizing digital experiences.

Alex Smolin, Senior Optimization Manager at AdSwerve, brings extensive experience in media, data, and technology. As a certified Google Premium Partner, AdSwerve provides data-driven brands with solutions ranging from A/B testing to advanced analytics.

Mary Kate, our roundtable host and Head of Growth Marketing for North America at AB Tasty, leads efforts to help companies create impactful digital experiences through AB Tasty’s suite of experimentation, personalization, product recommendations, and site search tools.

AB Tasty’s integration with GA4 & BigQuery

Connecting AB Tasty with GA4 gives marketing teams insights into visitor behavior through advanced analytics on CPA, conversion rate, bounce rate, SEO, and traffic. This integration allows teams to use data from either tool to measure the impact of experiments pre- and post-rollout, generating data-backed hypotheses and fostering innovation.

Google BigQuery, a fully managed cloud data warehouse solution, offers rapid data storage and analysis at scale. With its serverless, cost-effective structure, BigQuery allows businesses to analyze large datasets efficiently, making it easier to make well-informed decisions.

With Google BigQuery, users can effortlessly execute complex analytical SQL queries and leverage built-in machine-learning capabilities.

Why is data from GA4 foundational to any CRO program?

In experimentation, data is the catalyst that drives actionable insights. Data flows in from multiple sources, and businesses generate detailed reports by working with partners to integrate tracking and tagging. But the question then becomes: what comes next? That’s where experimentation enters. Using data from tools like GA4, teams can transform hypotheses into tests, uncovering which changes impact user engagement or conversions most effectively.

GA4’s role extends further by providing a consistent framework for testing across platforms. When integrated with BigQuery, GA4 allows teams to cross-reference test outcomes with other data points, revealing not just what worked but why it worked. As Alex noted, “We gather good data, run good tests, and then verify results across disparate sources like BigQuery to see if what we tested had the expected downstream impact.”

Data accessibility and agility are also important. Trends evolve quickly, with viral content or market shifts requiring rapid adaptability. “Having partners like Google, with all data in one place, and a platform like AB Tasty, where experiments can be quickly set up, is essential for staying competitive” Alex emphasized.

“Having partners like Google, with all data in one place, and a platform like AB Tasty, where experiments can be quickly set up, is essential for staying competitive.”

Alex Smolin, Senior Manager Optimization at Adswerve

How BigQuery powers scalable experimentation

With the growing volume of data, businesses need a way to consolidate and interpret it to drive impactful decisions. BigQuery, as Taige explained, is a robust cloud warehouse that streamlines data for meaningful insights, making it a key player in the experimentation ecosystem.

“Think of BigQuery as a filing cabinet for your organized data,” Taige noted. By consolidating disparate data sources, teams can create a unified view that informs testing and optimization efforts. Through this approach, tools like GA4 and BigQuery enable accurate decision-making that scales with the business. With BigQuery as the backbone, AB Tasty and AdSwerve can build on this structure to optimize user experiences through precise experimentation.

Beyond just data storage, BigQuery integrates with various Google Cloud tools and supports a wide range of use cases—from standard reporting to advanced machine learning. For marketers, this means fewer technical bottlenecks and quicker access to the data needed to stay agile. As Taige explained, “You may not need deep technical skills to access BigQuery’s benefits; the right partnerships and data structure can give you a powerful, accessible foundation.”

Leveraging BigQuery’s built-in AI and machine learning models

BigQuery offers an array of AI models for specific use cases—from translation and personalization to customer segmentation. These models add value by automating processes, such as localization or customer behavior prediction, allowing for smoother, more targeted marketing.

BigQuery’s flexibility means that companies can incorporate custom or third-party models, ensuring compatibility with a variety of AI solutions. This adaptability helps organizations innovate and iterate on experimentation programs, expanding what they can achieve with data.

Simplifying data access for marketing efficiency

For marketing teams, BigQuery’s role as a centralized data hub allows seamless data consolidation from platforms like Google Ads, Salesforce, and GA4. This integration ensures that marketers aren’t slowed down by fragmented data sources, freeing them to focus on insights and execution. As Taige highlighted, “The peace of mind that BigQuery provides comes from knowing that all data is consolidated and accessible, allowing teams to be nimble and creative.”

With BigQuery, marketers can view performance metrics, analyze customer journeys, and refine strategies—all within a unified environment. This lets teams optimize campaigns in real time as new data insights emerge.

Next-Generation capabilities enabled by Google Cloud

Looking ahead, digital  is paving the way for more advanced experimentation capabilities.  The conversation shifts to AI and machine learning, bringing new opportunities for personalization and optimization. As Mary Kate pointed out, while AI-driven insights can revolutionize customer experiences, many brands are still years away from realizing the full potential of these tools.

True value will come not from adopting every new tool but from understanding the foundational data supporting AI and asking the right questions about how these technologies can serve customer needs. Taige added, “If you don’t have a data strategy, you won’t have an AI strategy.” While AI amplifies data power, it requires organized, high-quality data to work effectively.

By consolidating and centralizing data through BigQuery, teams gain real-time insights and can make informed decisions. This data foundation enables the current wave of omnichannel strategies and sets the stage for future AI applications. Businesses that adopt this holistic approach—consolidating data, optimizing channels, and preparing teams for AI—will unlock new experimentation opportunities and drive impactful customer experiences.

With GA4 and BigQuery, businesses have the tools to streamline data consolidation and power next-generation experimentation. Ready to join your data and experimentation? Discover how AB Tasty can help bring data-driven optimization to life.

Subscribe to
our Newsletter

bloc Newsletter EN

We will process and store your personal data to respond to send you communications as described in our  Privacy Policy.

Article

9min read

Test, Optimize, and Upsell

For e-commerce success, added revenue from existing customers can be more efficient than constantly pursuing new ones. Returning buyers are a vital piece of this strategy. We recently sat down with industry experts to discuss how optimizing customer experiences can drive upselling and cross-selling opportunities. They shared practical approaches for boosting average order value (AOV) while nurturing customer loyalty and retention.

Our speakers, each experts in testing, optimization, and conversion rate, provided insights into how brands can increase revenue through personalized, thoughtful customer engagement.

Meet the experts

In this article, we’ll explore actionable strategies from the webinar to help you personalize to existing customers and drive growth through upselling and cross-selling—not just new customer acquisition.

1. Optimizing the cart for upsells

Upselling at the cart and checkout stages can significantly increase AOV, but it requires a carefully planned approach. As Colette Carlson explains: “Before implementing anything, it’s crucial to understand how you’re going to measure success and ensure that your conversion rate is solid. When it comes to the cart and checkout process, if those aren’t optimized, adding upsell and cross-sell strategies will only introduce more noise.” Shoppers who have reached the cart are already primed to convert, so it’s important not to disrupt their momentum with irrelevant or poorly timed offers.

“Before implementing anything, it’s crucial to understand how you’re going to measure success and ensure that your conversion rate is solid. When it comes to the cart and checkout process, if those aren’t optimized, adding upsell and cross-sell strategies will only introduce more noise.”

Collette Carlson, Director of Optimization at Astound Digital

Coordination with internal teams is also important when designing upsell strategies. For instance, if an upsell is introduced at checkout, the process should be seamless – will the original product be automatically removed from the cart if the customer selects an upgrade, or will they need to make the changes manually? Likewise, if you’re offering a bundle or cross-sell, is your system prepared to handle it without disrupting the customer experience?

Effective upsell offers are relevant to the customer’s purchase. Suggesting complementary items or upgrades can boost AOV, as 80% of consumers are more likely to complete their purchase with brands offering personalized experiences. From upsell testing experience at Giftory, Jared advises against pushing unrelated or overly expensive items, which can confuse or deter customers altogether. 

Using product recommendation algorithms can streamline upselling. Automating this process ensures that customers receive relevant suggestions without the need for manual curation, creating a smoother experience for your team and the customer. AB Tasty’s product recommendation engine allows upsells based on several criteria, including most recent products, associated products, similar or more expensive items, complementary items, and top promotions.

2. Strategic product recommendations for cross-selling

To effectively cross-sell, brands must identify the right moment in the customer journey. If you offer relevant products at key points without disrupting the experience, similar to upselling. But first, establishing cross-selling metrics can lead to stronger effectiveness.

The primary metrics will vary depending on what you’re testing—whether it’s an algorithm change, a new carousel design, or a different recommendation format. There are some essential KPIs to consider: 

  • Engagement: Track how often customers interact with cross-sell offers, such as clicks or add-to-basket rates.
  • Conversion rate: Measure how many customers who engage with offers complete their purchases.
  • Average order value (AOV): Gauge how effectively cross-sell strategies are increasing the total order value.
  • Items per order: Monitor if cross-sell efforts lead to additional products being added to the cart.
  • Overall revenue: This ultimate metric reflects the total impact of your cross-sell strategy.

Once these metrics are in place, refine your strategy by determining where cross-sell offers should appear. For example, adding a cross-sell option in the mini cart or as a pop-up at checkout can add complexity, so testing can help avoid disrupting the customer experience.

Testing cross-sell algorithms in action

Nicole Story shared a valuable example of testing product recommendation carousels. Inspired by Amazon’s success, many brands rushed to implement carousels on their websites, but forgot the importance of context. Placing multiple carousels on the homepage often leads to irrelevant suggestions and a poor experience.

Nicole’s team tested various algorithms by tailoring product recommendations to the customer’s journey. On product detail pages (PDPs), carousels that showed “related product suggestions” outperformed those with generic recommendations. The tests revealed that adjusting algorithms based on context and customer behavior was far more effective than placing standard carousels throughout the site.

As Nicole explains: “Simply introducing product recommendations and checking that box off the roadmap isn’t going to deliver real value. The key is continuous optimization and discovering what works across the entire customer journey—that’s where the real value lies.”

“Simply introducing product recommendations and checking that box off the roadmap isn’t going to deliver real value. The key is continuous optimization and discovering what works across the entire customer journey—that’s where the real value lies.”

Nicole Storey, Co-Founder & Director at Hookflash Analytics

Relevance is everything

Cross-sell strategies must be highly relevant to what the customer is already doing. As Gerred Blyth from Giftory mentioned, “We have high expectations as customers and irrelevant offers can break that trust.” Customers expect brands to know their preferences and behaviors, so it’s important that recommendations feel personalized and timely. 

3. Experimentation and testing for long-term loyalty and CLV

Continuous experimentation is critical for building long-term customer loyalty and increasing customer lifetime value (CLV). Instead of relying on a single strategy, brands should constantly test and improve their approach. Colette points out that starting by analyzing existing order data can uncover natural cross-sell patterns. This provides valuable insights into which products are frequently purchased together.

For first-time visitors, bombarding them with upsell offers might backfire. Instead, let them become familiar with your brand and key products before introducing additional offers. In contrast, repeat customers may be more open to cross-sells that align with their previous purchases.

Upselling with product recommendations

According to our data a customized UX can boost revenue and increase basket size by up to 10%. Product recommendations can be seen as a form of personalization and, as our panel pointed out in the webinar, experimenting with different formats—such as carousels, quizzes, or other interactive tools—can help identify what resonates with your audience and drives engagement.

We use AI to analyze visitors’ site interactions and purchase behavior, delivering targeted recommendations, each with a specific goal. This means you can better understand which products to offer, to whom and when during the customer journey:

  • Product Page: Guide users to explore more products or categories.
  • Last Seen Products: Help users quickly resume their browsing.
  • Add to Cart: Encourage users to add complementary items to their basket.
  • Cart Page: Suggest additional items to increase order value.
  • Homepage: Showcase personalized content and help users navigate the site.

Our panel also discussed how different types of algorithms are necessary depending on your vertical. You can divide your algorithms into three distinct types and choose how you prioritize:

  • Convert: These recommendations would offer top sellers, top trending products, top converting products, top reviewed products etc.
  • Upsell: This could suggest most recent products viewed, associated products, similar products, compatible products etc.
  • Personalize: This could suggest last visited products, last bought products, user affinity or similar or associated to cart products

If you work for a beauty site, customers will replenish their favorite products, whereas home and decor might recommend accessories or similar products. While personalization drives relevance, maintaining control over the recommendation process means you can speak directly to your customer’s needs. 

Giftory: fostering loyalty with timely engagement

Giftory is beginning to focus on lifetime customer value. Their approach involves using cross-sell and upsell strategies similar to a CRM initiative, introducing customers to a broader range of products both during and after their purchase. They gather data on why customers buy gifts, such as birthdays or anniversaries, and use that information to send timely product recommendations in the future.

By reaching out to customers at the right moment, such as 11 months after an anniversary purchase, Giftory can re-engage them with relevant offers without overwhelming them with constant promotions. This creates a personalized experience that encourages long-term loyalty and repeat business.

4. Subscription models for upsell and retention 

Offering subscription products to upsell can improve both immediate revenue and CLV. The challenge is to find the right balance: How can you encourage customers to subscribe without overwhelming them, while also ensuring the offer feels relevant and valuable over time? 

Before launching a subscription model, look at your data to understand customer behavior. Consider the difference between a one-time purchaser and a subscriber. While offering a small discount for subscribing may lower the initial AOV, the long-term benefits of recurring revenue from a loyal subscriber can make up for it. 

Testing and data-driven strategy

Launching a subscription model requires more than just adding an upsell feature—it involves a data-informed approach. Starting small with a minimum viable product (MVP) allows you to test how customers respond and fine-tune the offering. Metrics like renewal rates, engagement, and overall CLV will help guide decisions about whether to scale the program.

As Gerred advises: “Walk before you run. Start with the first test—an MVP. It doesn’t have to be the final version you’ll roll out, but that initial test will help you understand the value and prove the benefits. From there, you can evolve and continuously improve. It’s easy to feel overwhelmed when you hear about advanced strategies and algorithms, but you don’t need to get there all at once.”

“Walk before you run. Start with the first test—an MVP. It doesn’t have to be the final version you’ll roll out, but that initial test will help you understand the value and prove the benefits. From there, you can evolve and continuously improve. It’s easy to feel overwhelmed when you hear about advanced strategies and algorithms, but you don’t need to get there all at once.”

Gerred Blyth, Chief Product Officer at Giftory

Offering personalized options, such as different subscription tiers or flexible renewal cadences (monthly, bi-monthly, quarterly), can make the experience more appealing to a wider range of customers. Testing, refining, and adapting based on customer feedback will ensure that the model evolves in a way that meets both business goals and customer expectations.

Wrapping up 

Just as you approach CRO with care and precision, cross-selling and upselling require a high level of attention. 

Upselling and cross-selling don’t have to be complex when you have the right tools and the right strategy. If you want the expert’s opinion – watch the webinar below: