Snowflake and Reclaims 3,000+ Engineering Hours with Keebo

Company Overview

Cimpress is a global leader in mass customization, powering well-known brands like VistaPrint, National Pen, and BuildASign. With a decentralized, multi-tenant architecture spanning data meshes, microservices, and traditional data warehouses, Cimpress runs some of the most complex data environments in the industry. But that complexity created growing pressure on engineering bandwidth, budget control, and operational efficiency.

Key Challenges

Cimpress faced several critical challenges due to the rapid expansion of their data environments and the complexity inherent in their decentralized operational structure:

Cost Waste from Inefficient Queries: Thousands of failed queries and unused tables were driving up Snowflake bills and obscuring visibility.
Limited Engineering Bandwidth: A talented team was bogged down by manual tuning, monitoring, and clean-up, diverting focus from higher-impact initiatives.
Multi-Tenant Complexity: Distributing meaningful insights, optimization strategies, and ensuring accountability across multiple internal teams proved challenging due to the decentralized structure and varied operational needs across business units.

Why Cimpress Chose Keebo

Keebo, the AI agent for cost and performance on Snowflake and Databricks, offered exactly what Cimpress needed: a hands-free, real-time optimization layer. Keebo watches every query, right-sizes compute, tunes parameters, reroutes workloads, and flags inefficiencies—all without touching customer data.

Key Results

3,000+ Engineering Hours Saved Per Year: Automation replaced manual reviews and performance triage, giving engineers time back for product innovation.
15% Savings on Snowflake: Alongside implementing fully automated optimizations, Keebo surfaced over 250,000 failed queries and nearly 12,000 unused tables, contributing to a 15% reduction in Snowflake costs.
Enhanced Autonomy: Over six distinct teams within Cimpress gained the capability to autonomously manage their analytics and spending. This empowerment improved responsiveness, accountability, and decision-making across the organization.
Seamless Governance: Keebo runs in approval-only or full-autopilot mode—Cimpress opted for hybrid governance, striking a balance between control and automation.
Validated Internal Tools: Keebo’s insights not only complemented but validated Cimpress’s internal Snowflake Information Tool (SIFT). This validation encouraged confidence in internal innovation, provided benchmarks against industry standards, and highlighted the relevance and effectiveness of their engineering efforts.

We’ve seen 15% in Snowflake savings, plus huge time savings for our engineers. Keebo gives us confidence that we’re running efficiently and lets each team manage their footprint with visibility and autonomy.

Michael Theodoulou
Data Technology Principal

Looking Ahead

Cimpress and Keebo continue to collaborate on evolving optimization strategies. With zero security risk (Keebo uses metadata only) and zero pricing risk (pay-as-you-save model), Cimpress has not only improved efficiency but also safeguarded performance and budget predictability—ensuring teams stay focused on delivering value at scale.

Keebo gives us a unified view of Snowflake usage across a very decentralized environment. From spotting wasted compute to tuning warehouse settings and surfacing optimization opportunities we’d never catch manually, Keebo gives us actionable insights and automated results.

Michael Theodoulou
Data Technology Principal

Download PDF