How an e-commerce company got 3x faster analytics on AWS

The challenge

Atsell is a retailer selling on Shopee, Lazada and TikTok Shop. To grow, the team needed quick answers about their sales, their competitors and their advertising.

Before working with us, they faced:

• Slow reports that took too long to answer everyday questions.
• Data in many places: sales, competitor prices and ads on Shopee, Lazada and TikTok Shop were tracked separately.
• Hours of manual research to compare products and prices with competitors.
• Unclear ad ROI, making it hard to know which campaigns to scale or stop.
• High infrastructure cost for the analytics they had — reports ran on a general-purpose database and spreadsheets that slowed down as data grew.

What Bytenamics did

We deployed Bytenamics Analytics — our ClickHouse-based analytics platform on AWS — connected it to the team's marketplace data and added an AI assistant.

1. Bring the data together

Sales, competitor and ad data from Shopee, Lazada and TikTok Shop lands in Amazon S3. Amazon Athena handles ad-hoc queries on the raw data, and ClickHouse on Amazon EKS and EC2 serves the fast analytics. Data refreshes every hour.

2. Model for fast questions

We designed tables and pre-built views around the questions the team asks most, so queries over 1 million+ rows return in seconds.

3. Dashboards for each team

Sales, marketing and management each got dashboards for their own decisions.

4. Ask questions in plain language

A RAG assistant running on Amazon EC2 answers questions like "which competitor cut prices this week?" from the team's own data — cutting research time.

5. Right-size the infrastructure

ClickHouse compresses data and needs far less compute than their old setup. Together with right-sized AWS resources, this cut infrastructure cost by 42%.

1. Week 2: DiscoveryKey questions, data sources and success metrics agreed.
2. Week 4: Platform liveClickHouse on Amazon EKS and EC2 deployed and secured.
3. Week 6: Data connectedShopee, Lazada and TikTok Shop data loading automatically.
4. Week 8: Dashboards in useTeams using the platform day to day.

The results

• 3x faster queries across datasets of 1 million+ rows.
• 42% lower infrastructure cost for analytics.
• Less time on research: competitor and market checks that took 4 hours now take 15 minutes.
• Better ad ROI: the team moves budget to campaigns that work, faster. Return on ad spend rose 25% in the first quarter.
• Faster innovation: new ideas are tested with data in days, not weeks.

What's next

Next, Atsell plans to add AI demand forecasting and automated pricing suggestions.

Want faster answers from your data? See our modernization and RAG services.