StatPearls improved ROAS and launched ML-driven campaigns with trusted customer context in Redshift

StatPearls serves thousands of clinical learning experiences across a wide range of medical roles, from students to practicing clinicians. That breadth is a growth advantage, but it also creates a hard data problem: The customer journey is complex, and lifetime value is not a simple single-purchase story.

When you cannot reliably connect anonymous web behavior to a registered user, then to a subscription renewal, you end up optimizing marketing to the wrong signal. Teams can see the first conversion, but not the long-term value. ROAS looks good until it does not, and scaling ad spend becomes guesswork.

StatPearls solved this by building a cloud-first customer data infrastructure with RudderStack, centralizing first-party behavioral and transactional data in Amazon Redshift, and creating complete customer profiles that marketing could actually use. The result was better measurement, faster segmentation, and a clear path to ML-driven churn and personalization.

Main takeaways

• Centralize first-party data in your data cloud so attribution and LTV are computed from the same source of truth.
• Turn event and payment data into usable customer profiles, not one-off SQL investigations.
• Use profiles and predictive traits to move from reactive campaigns to proactive retention and personalization.
• Scale spend only after you can measure true cost per paid customer and cohort LTV with confidence.

Why customer context became the bottleneck

StatPearls wanted to answer questions that sound simple but are operationally difficult without customer data infrastructure:

• How much organic traffic to free clinical articles turns into paid customers over time?
• Which types of medical professionals buy which products and renew at the highest rates?
• What is our true cost per paid customer when we account for renewals and expansion, not just the initial purchase?
• Which cohorts are likely to churn, and what intervention actually changes outcomes?

They had the raw ingredients across systems like payments, messaging, and their product database. The problem was the workflow: reconciling disparate schemas, stitching identities, and mining the production database repeatedly to answer recurring questions.

That created three compounding issues:

1. Slow time to insight. Answers required manual work, not repeatable pipelines.
2. Low confidence. Data quality and identity gaps made results hard to trust.
3. Limited activation. Even when insights existed, it was hard to operationalize them into marketing cohorts quickly.

What changed: A single source of truth in Amazon Redshift

StatPearls implemented RudderStack to collect and standardize their first-party customer data, then deliver it directly into Amazon Redshift.

Practically, this meant:

• Behavioral events and conversions were captured reliably across the website and product experience.
• Payments data could be modeled alongside usage, so "value" included renewals and expansion.
• Data teams stopped answering the same questions with bespoke queries against production systems.
• Marketing and analytics started working from shared tables and definitions.

The underlying principle is straightforward: AI and automation raise the cost of ambiguity. If your customer journey is stitched differently in every analysis, every downstream decision gets noisier. A durable customer context layer depends on governed collection, stable identity, and consistent delivery paths into the data cloud.

From centralized data to complete customer profiles

Centralizing data in Redshift was the foundation, but the breakthrough came from turning that data into complete customer profiles.

StatPearls used RudderStack Profiles to:

• Resolve identities across anonymous visitors, registered users, and paying customers.
• Build a unified customer 360 table in Redshift that reflected the full journey.
• Add traits and attributes quickly so marketing could segment without waiting on long, repeated SQL work.

Instead of treating every new segmentation request as a data engineering ticket, profiles made customer traits a living, evolvable layer. That matters because the segmentation surface area grows as the business grows: more products, more roles, more learning pathways, more lifecycle states.

When profiles are easy to extend, teams can iterate faster without creating fragile one-off logic.

Making ROAS measurable, then scaling spend with confidence

Once StatPearls could connect acquisition to renewal and expansion, as their Head of Engineering put it, marketing became a math problem. When you spend X on a particular campaign, you know you are going to get Y yield over time.

With granular attribution signals and full payment history available in Redshift, the team could:

• Calculate true cost per paid customer by cohort, not just cost per first conversion.
• Measure LTV by customer type, including renewals and additional purchases.
• Identify which channels and campaigns produced the most valuable long-term cohorts.

That clarity is what enabled them to scale ad spend aggressively while maintaining positive ROAS, including a 3.8x increase in Google Ads budget with positive returns.

Using customer context to power ML-driven lifecycle campaigns

With complete profiles in place, StatPearls could move from descriptive analytics to predictive action.

They used Profiles as a foundation for churn modeling, then turned model outputs into traits the business could act on:

• Predictive churn signals were added as profile traits.
• Marketing used those traits to trigger timely, relevant outreach and offers.
• Personalized messaging tripled in volume because targeting was grounded in trusted customer context.

This is where the real leverage of a well-governed customer context layer shows up. The most useful automation workflows are not just chat interfaces. They are systems that continuously assemble fresh customer context, apply governance, and trigger the next best action based on reliable signals.

Results at a glance

• 3.8x increase in ad spend with positive ROAS
• 1.7x increase in customer retention
• 3x increase in personalized message volume

If you want the full story and implementation details, read the StatPearls case study.