Data science: Next-gen performance analytics
summary
Claims collecting 425 billion beacons and keeping raw events costs about $9,000/month storage so predictive models and ad hoc queries can run on complete data; a 10M daily page-hit site yields ~0.5B yearly beacons and ~3.5B page resources. The system ranks pages by a conversion impact score, reconstructs sessions to prioritize fixes, uses 1,440 threshold pairs per day for minute-level anomaly alerts, surfaces failing $5M campaign launches in real time, and highlights that >50% of resources come from third parties.
AI-written summary · about
chapters (7)
- 0:00Start
- 1:221ST DIGITAL PEFORMANCE MANAGEMENT COMPANY
- 2:07TRADITIONAL ANALYTIC DASHBOARDS
- 8:00SINGLE CAMPAIGN LAUNCH DASHBOARD
- 9:04CROCKER CHARTS | DEVICE TYPE
- 10:02CROCKER CHARTS BROWSERS
- 10:223RD PARTY RESOURCE ANALYTICS
related talks
36mML: Predictive analytics in Ad Engineering
15mThe Secret Life of Data - Part 2
27mBatching Was Yesterday: Real-Time Tracking & Analysis For 100+ Million Visitors
34mHow We Handle Over 1 Trillion HTTP Requests ...
48m10 Billion a Day,100 Milliseconds Per: Monitoring Real Time Bidding
01h 02mAchieving High Load in Advertising Technology
53mDemocratising Attention Data at theguardian.com
01h 00mHyper Speed: When Big Data Blooms