To build a great marketing campaign in today’s landscape, data needs to be steering your strategy, not just measuring success. Developers play a key role in implementing the tools that analyze and process this data, turning it into insights, smarter strategies, and better results.
Unlock the power in your marketing data with these three developer-friendly MarTech solutions. From gathering data with unparalleled transparency and control, to transforming raw data into structured insights, or using automated A/B testing for optimal performance, here’s how developers can transform what marketing data can do.
From buttons clicked to pages scrolled, knowing how people interact with your website or app is crucial to optimizing performance. Server-side Google Tag Manager (sGTM) makes this process easier by measuring traffic and managing data flow—while opening the doors to better privacy, performance, control, and productivity.
sGTM Pantheon is a toolbox of easy-to-deploy solutions that complement the existing capabilities of sGTM in different ways:
Developers have the flexibility to mix and match solutions to create a single pipeline that can be integrated with both Google and non-Google platforms. And because sGTM Pantheon uses a server environment, the solutions run in a private, first-party cloud-secure environment.
sGTM Pantheon is a living solution and is continually growing. Want to see more tools? Explore the full sGTM Pantheon on GitHub.
Your Google Analytics 4 (GA4) marketing data holds untold stories, powerful insights, and new ways to connect with your audience—but deciphering it isn’t always easy.
GA4 Dataform is a data transformation tool that organizes raw BigQuery data into clear, modular tables, such as events, items, sessions, transactions, and more—so users of all technical skill levels can analyze data and steer data-driven campaigns. Offering both depth and simplicity, GA4 Dataform gives you the power to go beyond default settings, build your own data models, and find new ways to engage with customers.
GA4 Dataform is a Google Cloud Dataform project that provides SQL data models for transforming raw GA4 BigQuery exports. The code is essentially a starter pack to help you build models on top of the GA4 raw data exports for data-driven marketing insights.
The features available now include:
1: Building a unique user_key and ga_session_key.
2: Providing as output a digestible session table, user_transaction_daily table, event table, and more.
3: Gclid widening by mapping the GA4 GCLID to the Google Ads Data Transfer click-view GCLID (Optional setting)
4: Event level last-click attribution.
Ready to get started? Deployment is simple—explore GA4 Dataform on GitHub to learn how.
What if you could eliminate the guesswork and manual testing from your Google Ads shopping campaigns? FeedX is an open-source experimentation framework helping advertisers run A/B testing for shopping feed modifications—so they can see the result of specific tweaks against observed performance changes.
Online advertisers who want to scale optimizations across their inventories need to know their strategy will have a positive impact on performance. But without a clear feedback signal, it's hard to know whether creative changes are making the results better or worse.
FeedX solves this problem by allowing advertisers to test any changes using a reliable Python A/B testing framework. FeedX is a Python package, containing all of its logic and mechanics, as well as a set of Colab notebooks which show you how to use the package to design and analyze experiments.
FeedX uses industry best practices to ensure the experiment is as robust and sensitive as possible. With a crossover design, it adjusts for pre-experiment performance with CUPED (Controlled-experiment Using Pre-Experiment Data), and trims outlier items if necessary. Here’s an overview of the flow:
1: The advertiser starts with an item they would like to test, for example, optimizing a title or description. To ensure reliable results, the test should include at least 1000 items, and the FeedX design notebook will alert you if the sample size is too low.
2: The feed items are randomly split into two groups, a control group and a treatment group.
3: The advertiser creates a supplemental feed, containing only the optimizations for treatment items, and starts the experiment by uploading this supplemental feed to the Merchant Center.
4: Optionally, crossover experiments can be run where the advertiser swaps these groups so the treatment group becomes the control group.
5: At the end of the experiment, the performance of all items is analyzed and compared between the control and treatment groups. The result is a reliable metrics report, backed by a confidence interval and statistical significance.
Forget the guesswork. Ready to revolutionize shopping ads with data? Take a deep dive into how FeedX works on GitHub.
This is the second post of our two-part series on bridging the gap between marketing and development. To explore our gen AI MarTech solutions, check out Three MarTech solutions putting generative AI in marketing.
Keep an eye out for more updates on the Google for Developers blog, or check out our MarTech solutions guide to find even more innovative tools you can implement, today.
Adding support for Google Pay within Android WebView
What you should know from the Google I/O 2025 Developer keynote
What's new with Agents: ADK, Agent Engine, and A2A Enhancements
Gemini API I/O updates
Google Cloud announces general availability of APIM Operator for Apigee