Aeolus Data Solutions

BACKGROUND

Please describe your company and position.

I am the Founder of Grace's Towel

Describe what your company does in a single sentence.

Grace's Towel is one of the largest tectile importers in PNW region specialized in home goods and industrial needs

OPPORTUNITY / CHALLENGE

What specific goals or objectives did you hire Aeolus Data Solutions to accomplish?

• Inventory & Demand Forecasting Pipeline: Connect real-time SKU sales velocity with supplier lead times and warehouse stock levels to automate reorder point recommendations and prevent stockouts.
• Automated Customer LTV & Retention Modeling: Model customer cohorts, repurchase cycles, and RFM segments in dbt to power automated lifecycle marketing, personalized upselling, and retention campaigns.
• Unified E-Commerce & Ad Attribution Lakehouse: Build a centralized data platform integrating storefront events (Medusa/web), order transactions, and ad spend (Meta/Google) for true multi-touch attribution and ROAS visibility.

SOLUTION

How did you find Aeolus Data Solutions?

• Online Search
• Referral

Why did you select Aeolus Data Solutions over others?

• High ratings
• Close to my geographic location
• Pricing fit our budget
• Great culture fit
• Good value for cost
• Referred to me
• Company values aligned

How many teammates from Aeolus Data Solutions were assigned to this project?

2-5 Employees

Describe the scope of work in detail. Please include a summary of key deliverables.

Scope of Work

• Phase 1: Tracking & Architecture: Audit and implement server-side tracking (Meta CAPI, GA4, PostHog); design unified e-commerce & ad schemas.

• Phase 2: Ingestion & Lakehouse Infrastructure: Deploy cloud data warehouse via Terraform; build automated ELT connectors for Medusa, Stripe, and ad channels.

• Phase 3: dbt Modeling & Attribution: Build production dbt models for blended ROAS/MER, customer RFM/LTV cohorts, and inventory reorder alerts with CI/CD testing.

• Phase 4: BI Dashboards & Reverse ETL: Build self-serve dashboards (Lightdash/Metabase); sync customer segments to Klaviyo/ad platforms via Reverse ETL; deliver runbooks.

Key Deliverables

1. Server-Side Tracking Layer: Meta CAPI + GA4 setup for full-funnel event capture.

2. IaC Cloud Lakehouse: Terraform-managed modern data warehouse (Snowflake/BigQuery).

3. Automated Data Pipelines: Connectors syncing Medusa storefront, Stripe, and Meta/Google Ads.

4. dbt Transformation Models: • Attribution: Multi-touch ROAS, CAC, and blended MER. • Customer Intelligence: RFM segmentation, cohort retention, and predictive LTV. • Inventory: Sales velocity, gross margins, and automated reorder triggers.

5. Reverse ETL Pipelines: Automated sync pushing high-value cohorts to CRM/Klaviyo and ad audiences.

6. BI Dashboards: Executive overview, marketing attribution, and inventory health reports.

7. Documentation & Runbooks: Data dictionary, metrics catalog, and pipeline operations guide.

RESULTS & FEEDBACK

What were the measurable outcomes from the project that demonstrate progress or success?

• 15–20% Reduction in Wasted Ad Spend: Achieved full cross-channel visibility into blended MER and multi- touch ROAS, eliminating inefficient ad spend through server-side tracking (Meta CAPI).

• 20–25% Increase in 90-Day Repeat Purchase Rate: Boosted customer retention and repeat orders by syncing automated RFM and predictive churn segments directly into lifecycle marketing (Klaviyo).

• 30% Decrease in Out-of-Stock Incidents: Prevented stockouts across top-selling SKUs using real-time sales velocity tracking and automated reorder point alerts.

• 15-Day Reduction in Safety Stock Holding Duration: Optimized working capital by cutting excess inventory holding time without compromising fulfillment SLAs.

• Near-Zero Attribution Data Loss: Eliminated client-side ad-blocker drop-offs by routing 100% of conversion events through a unified server-side tracking pipeline.

90-Day Repeat Purchase Rate: Increase customer 90-day repeat purchase rate by 20–25% via automated predictive segmentation synced directly into lifecycle marketing tools.

Stockout Frequency & Turnover: Decrease out-of-stock incidents by 30% while reducing excess safety stock holding duration by 15 days using automated inventory reorder alerts.

Describe their project management. Did they deliver items on time? How did they respond to your needs?

 • Structured, Milestone-Driven Execution: Worked in clear, agile sprint cycles across all four project phases (Tracking → Infrastructure → Modeling → BI/Activation), supported by regular async updates, weekly milestone demos, and transparent task tracking.

• 100% On-Time Delivery: Completed every core milestone on schedule—including the cloud warehouse deployment, dbt attribution models, and dashboard cutover—with zero disruption to live storefront operations.

• Proactive & Highly Responsive:

• Rapid Iteration: Quickly adapted dbt models and dashboard metrics to accommodate custom reporting nuances and evolving marketing requirements. 
• Issue Preemption: Proactively flagged edge cases in event tracking and ad attribution before they impacted downstream reporting.
• Seamless Knowledge Transfer: Provided thorough documentation, data dictionaries, and hands-on walkthroughs so internal teams could operate the platform independently.

What was your primary form of communication with Aeolus Data Solutions?

• In-Person Meeting
• Virtual Meeting
• Email or Messaging App

What did you find most impressive or unique about this company?

• Full-Stack Data Engineering & Activation: Unlike traditional consultants who only build pipelines or static dashboards, Aeolus connected the entire loop—from server-side tracking (Meta CAPI) to dbt transformations and operational Reverse ETL syncing directly into marketing tools.

• Deep E-Commerce & Growth Acumen: They spoke the language of the business, translating complex data models into actionable commercial levers (blended MER, RFM cohort retention, inventory velocity).

• Production-Grade Engineering Rigor: Built with long-term maintainability in mind—using Infrastructure as Code (Terraform), automated CI/CD testing on all dbt models, and data contracts rather than fragile ad-hoc scripts.

• Zero Vendor Lock-In: Left behind well-documented runbooks, clean data models, and a self-serve analytics stack that our internal team could easily own and extend.

Are there any areas for improvement or something Aeolus Data Solutions could have done differently?

• Earlier BI Wireframing for Non-Technical Teams: Introducing dashboard mockups earlier during the data modeling phase would have streamlined final feedback and sign-off from non-technical marketing and merchandising stakeholders.

• Staggered Team Onboarding: Because the modern data stack (dbt, semantic layers, BI self-serve) represented a significant step up in sophistication, scheduling bite-sized training sessions throughout the project—rather than concentrating them at handoff—would have eased the team's transition.

• Third-Party API Rate-Limit Buffering: Providing clearer upfront guidelines on third-party ad platform API quirks and rate-limit constraints could have reduced minor back-and-forth during initial data reconciliation.