Snowflake Native App

Faster Snowflake queries.
Lower Snowflake bills.

Insights automatically tunes clustering keys so your queries scan only what they need — cutting compute costs 10–50% while speeding queries up. Backed by your own query data.

Try Free on Snowflake Marketplace View Benchmarks
51%
Faster queries
76%
Fewer partitions scanned
81%
Fewer bytes scanned

Measured on the TPC-DS 10 TB benchmark — real execution, not projections.

Introducing

Insights by Goldilox

See what's costing you — and exactly how to fix it.

Insights analyzes your Snowflake query patterns and recommends optimal clustering keys for every table — each with estimated credit and cost savings, so you know the impact before making a change.

Know exactly when, how, and why to cluster. Powered by AI/ML, backed by your own query data.

Benchmarks

Proven results on industry-standard benchmarks

Measured on TPC-H — the standard benchmark for decision-support query performance — and TPC-DS, the industry's most demanding test for large-scale analytics workloads. Real numbers from actual partition pruning and query execution, not projections.

TPC-DS 10 TB scale (SF10TCL)
57 billion rows · 24 tables · 99 queries
2X-Large warehouse
51%
Faster query time
32.1 min → 15.7 min
≈ 51% lower compute cost
76%
Fewer partitions scanned
11.2M → 2.7M
81%
Fewer bytes scanned
30.9 TB → 5.8 TB
TPC-H 100 GB scale (SF100)
866M rows · 8 tables · 22 queries
Small warehouse
15%
Faster query time
104.6s → 91.6s
≈ 15% lower compute cost
39%
Fewer partitions scanned
27.5K → 16.7K
37%
Fewer bytes scanned
128.7 GB → 80.1 GB

Faster queries are cheaper queries. Snowflake compute is billed by warehouse runtime — so cutting query time 51% cuts the compute cost of that workload by roughly half. No discounts to negotiate, just less warehouse time burned.

Per-table pruning improvement

Across the largest tables in both benchmarks, Goldilox consistently removes 60–80% of unnecessary partition scans. Bigger tables see bigger wins — and they're the ones that cost the most to scan.

75%
Median partitions pruned
Across the 7 largest fact tables in TPC-DS
60–80%
Typical range per table
Multi-billion-row tables like STORE_SALES, CATALOG_SALES, WEB_SALES
71%
Average scan time reduction
STORE_SALES (28.8B rows) — fewer partitions read, faster queries

Baseline: fully unclustered tables (~0% pruning) — the starting point for tables with no clustering strategy. Read the full write-up →

“Goldilox provided clear, workload-driven recommendations for Snowflake table optimization. On evaluated production workloads, its recommended clustering strategy was associated with up to 99.9% data pruning.”
Director of Engineering, Enterprise FinOps Company Fortune 100 Customer

Quick Insights into Data Efficiency

Unlock instant visibility into how efficiently your Snowflake queries are pruning partitions. Our platform continuously monitors and surfaces actionable insights — helping you optimize performance, reduce costs, and maximize the impact of every scan.

Partition Pruning Efficiency Dashboard

Partition Pruning Coverage

  • Track partition pruning coverage and efficiency
  • Spot excessive scanning before it inflates your bill

Identify optimization opportunities across all tables

Pinpoint underperforming tables with low partition pruning efficiency. Our analysis highlights which tables are causing excessive scans, helping you prioritize optimization efforts where they matter most.

Table Optimization Analysis

Pruning Performance

  • Surface tables with poor pruning performance
  • Identify excessive full-table scans

Smart Prioritization

  • Prioritize high-impact optimization targets
  • Rank tables by estimated savings

Intelligent Clustering Key Recommendations

Get AI/ML-powered recommendations for optimal clustering keys based on your actual query patterns. Each recommendation includes estimated credit and dollar savings so you can prioritize the changes that matter most.

Clustering Key Recommendations

AI-Powered Suggestions

  • Data-driven clustering key suggestions
  • Recommendations based on real query patterns

Actionable Results

  • Estimated credit and cost savings per recommendation
  • Ready-to-run ALTER TABLE SQL for each table

Getting Started is Easy

Fully self-service — get up and running in 5 simple steps

1

Install the GOLDILOX Insights App

Install Insights directly from the Snowflake Marketplace - no external deployments needed.

2

Configure Permissions

Grant the necessary permissions for Insights to analyze your query patterns.

3

Select High-Impact Tables

Review and select recommended high impact tables to monitor and optimize.

4

Automated Analysis

Insights begins analyzing your workloads and generating optimization recommendations.

5

Reduce Costs

Implement recommendations and start seeing cost reductions within days.

Why Choose Insights?

Faster Queries

Optimize query performance without code changes.

10–50% Cost Savings

Reduce compute costs with intelligent optimization, estimated per table before you commit.

Native App Deployment

Runs entirely within your Snowflake environment — nothing to host, no external deployments.

Data Security

Your data never leaves your Snowflake account — only query metadata and telemetry are analyzed.

Quick Results

See improvements within days, not months.

You Stay in Control

Insights automates the analysis and generates recommendations — you review and apply changes at your discretion.

Simple, transparent pricing

Start free. Scale as you save.

Limited Time
Promotion
Free
Up to 5 tables
Full recommendations — no payment required while the offer is running.
  • Full clustering key recommendations
  • Dollar & credit savings per table
  • Apply changes at your own pace
Get Started
Small
$100/ month
Up to 3 tables
Ideal for proof of concept and validating ROI.
  • Automated clustering key recommendations
  • Dashboard with partition pruning metrics
  • Continuous monitoring as workloads shift
Get Started
Medium
$300/ month
Up to 10 tables
Best for growing data teams scaling Snowflake spend.
  • Everything in Small, for up to 10 tables
  • Continuous monitoring with savings estimates
  • Priority email support
Get Started
Large & Enterprise
$1,000+/ month
40 tables to unlimited
Best value at scale — plus custom terms for accounts spending $1M+ on Snowflake compute.
  • Everything in Medium, scaled to 40+ tables
  • Unlimited table coverage on custom plans
  • Volume discounts & dedicated support
Contact Sales
All plans install via the Snowflake Marketplace (sign-in required) — installing previews the Insights dashboard for free, and the promotion and paid tiers unlock full clustering key recommendations with estimated savings. 10% discount when billed annually. Upgrade, downgrade, or cancel anytime.
Songwei Tsay
About Goldilox

Decades of data platform experience, one focused product.

Goldilox is focused on one job: making Snowflake optimization automatic, measurable, and trustworthy — a problem every serious data team hits, and one that deserves a real engineering answer.

Songwei Tsay Founder & Chief Architect

Two decades of experience designing and building data, cloud, and platform systems — from query engines to large-scale financial planning platforms.

Connect on LinkedIn

Frequently Asked Questions

Insights uses trained prediction models to analyze your query patterns and telemetry data, identifying optimal clustering keys for your tables. By improving partition pruning, queries scan less data — resulting in faster execution times and lower compute costs. Each recommendation includes estimated credit and dollar savings so you can see the impact before making changes.

No! Insights is a Snowflake Native App, which means it runs entirely within your Snowflake environment. Simply install it from the Snowflake Marketplace and you're ready to go.

Typical savings range from 10% to 50% reduction in query costs, depending on your current clustering configuration and query patterns. Insights calculates estimated credit and dollar savings per table based on your actual query history — you'll see the numbers directly in the dashboard, not just during a consultation.

Absolutely. Insights operates entirely within your Snowflake environment. Your data never leaves your account, and we only analyze query metadata and telemetry - not your actual data.

Most customers see improvements within days of implementing Insights' recommendations. The analysis begins immediately after installation, and you'll receive actionable insights as soon as we've gathered enough data.

Pricing scales with the number of monitored tables, with plans starting at $100/month — and there's currently a limited-time free promotion. See the pricing section for current tiers.

Ready to cut your Snowflake bill?

Install Insights and start cutting compute costs within days.

Try Free on Snowflake Marketplace Request a Demo

Partnership or press? Get in touch →