How Nuts.com Replaced GA and Unlocked Self-Service Analytics

A third-generation business with roots dating back to 1929, Nuts.com has grown from a local family-run shop into a digital-first leader in premium snacks and pantry staples. With a product catalog ranging from cashews to chocolate espresso beans, Nuts.com is now one of the most recognized DTC food retailers online, second only to Amazon in its category.

But growth didn’t stop at DTC. Over the past year, the company has leaned into retail, B2B, and corporate gifting channels, establishing partnerships with stores like Trader Joe’s and expanding into wholesale and food service. These moves signaled a new phase of growth and a need for a far more modern and scalable data infrastructure to support it.

The Challenge: Fragmented Data and Growth Pains

As the business evolved, it became clear that Nuts.com had outgrown its legacy data tooling. The company had experienced rapid growth during the COVID-driven DTC surge, but momentum was beginning to slow. Their data infrastructure, reliant on fragmented Google Tag Manager events, inconsistent tagging, and nightly attribution batches from Google Analytics wasn't built to support the company’s expansion or enable deeper customer understanding.

Without clean, real-time customer data, the business faced growing pains:

• First-time buyer journeys were not well understood
• Returning customer behavior was hard to track
• Experimentation with new tools and channels was slow
• Marketing and product teams were without consistent metrics or customer insights

As the team set its sights on growing revenue through smarter retention, personalization, and expansion, they needed a trusted foundation and one that could unify customer signals, simplify activation, and deliver value quickly.

Partnering with RudderStack

The initial focus was clear: establish clean, consistent pipelines for collecting and activating customer data, while minimizing engineering lift and maximizing flexibility. RudderStack offered a composable, engineering-friendly approach that met the team’s dual goals of data quality and agility.

In just the first six months, RudderStack enabled Nuts.com to:

• Replace legacy GA tagging with real-time event pipelines
• Stream customer events into Redshift for modeling, while simultaneously routing to tools like Iterable, Attentive, and Algolia
• Empower marketers to self-serve analysis and build audiences
• Stand up and test new integrations in under two days

A Data-Driven Transformation

With the foundation in place, Nuts.com began evolving how data flowed across the business. The team migrated legacy analytics workflows into dbt, built cleaner transformations, and began using RudderStack Reverse ETL to activate modeled traits (like product affinities or subscription intent) back into their downstream tools.

These capabilities supported a range of initiatives:

• Modeling the first-time buyer experience
• Enabling returning customer journeys like Subscribe & Save and personalized gift recommendations
• Powering real-time search and recommendations with Algolia using RudderStack’s event stream
• Supporting new channel growth, like retail and B2B, with deeper customer insights and segmentation

While these workflows were evolving, so too was the data culture. By making data accessible and trustworthy, RudderStack helped shift the mindset of the business. What makes this transformation so compelling is how deliberately the team approached it. Nuts.com prioritized use cases that aligned directly with growth goals, driving smarter email campaigns, powering real-time personalization, and unlocking operational reporting.

Today, Nuts.com is using RudderStack as a full-circle engine:

• Teams now operate from a single source of truth
• Speed to insight and activation is radically faster
• Channel experimentation (e.g., new partners or retail campaigns) happens in days, not weeks
• Personalization infrastructure is taking shape with RudderStack at the core

What’s next? Scaling this framework, further automating audience modeling, and moving from segmentation to prediction.