Company
A global leader in travel technology since 2009, Peakwork specializes in dynamic packaging and distribution software. Their innovative Player Hub Technology creates a powerful network, efficiently linking global travel supply and demand for leading tour operators. Designed for high-volume traffic and extensive distribution, Peakwork's solutions streamline the global distribution of hotels and flights, enabling customized offers and maximizing market reach.
Challenge
Peakwork’s core technology manages massive traffic volumes and requires rapid search results to optimize the "look-to-book" ratio. Historically, Peakwork operated on BigQuery's on-demand pricing model and wanted to ensure a transition to capacity-based pricing wouldn’t increase costs. This raised a key decision: continue with the familiar on-demand model, or introduce capacity-based to improve predictability and value.
Solution
Peakwork integrated Alvin's autonomous agent for their Looker queries, attracted by its easy setup and immediate value. Alvin's dynamic query routing intelligently optimizes BigQuery queries, leveraging the most cost-effective pricing plan based on the query characteristics. This allowed Peakwork to combine the benefits of both on-demand and capacity-based pricing, significantly reducing their BigQuery spend from day one without manual effort or engineering overhead.
Results
- 54% reduction in query costs
- 19% reduction in average query time
- Improved visibility on BigQuery spend
- Increased predictability and consistency of daily spend
“We were impressed by how quickly Alvin was to set up and once we connected we started to see results almost immediately. We’re now seeing savings of over 50% of our total BigQuery spend on queries.”
Rouven Urfels, IT Architect at Peakwork
Billing model Optimization
Alvin continuously analyses every query executed by Peakwork, calculates the optimal billing model, and executes the query using the most cost effective. For compute-heavy queries, on-demand billing is used, ensuring Peakwork only pays for bytes scanned rather than the larger cost associated with slots. For scan-heavy query, capacity billing is used, avoiding being billed for bytes scanned. Here is what Peakwork’s savings looked like in July:
Consistency and predictability of spend
Prior to using Alvin, because of fluctuation in usage Peakwork often saw cost spikes that required additional cost-intensive review. Because on-demand is billed on bytes scanned with no upper limit, even a single query that scans a lot of data can lead to an unexpected increase in spend. With Alvin, these scan heavy queries are run on the reservation, leading to a smoothing of these spikes. As you can see from the chart below, Peakwork’s costs became much more consistent with Alvin.
Ready to Optimize Your Data Spend?
Discover how Alvin can transform your cloud data costs. Schedule a personalized demo or run a free savings analysis to quantify your potential outcomes and see immediate value.
