AWS On Air Voice of the Customer ft. Amazon Ads
summary
Amazon Ads runs an ad stack that serves hundreds of millions of requests with a tight ~120 ms latency budget, recording ad events into a petabyte data lake and using real‑time attribution to report return on ad spend. Architecture uses ElastiCache (Memcached) for ~500M key lookups/sec, DynamoDB + OpenSearch to store tens of billions of campaign objects with sub‑second reads/writes and six‑nines availability, and streaming with Kinesis / DynamoDB Streams + Lambda for budgeting, near‑real‑time ML, and auto‑scaling during peaks like Cyber Monday.
AI-written summary · about
chapters (10)
- 0:00Intro
- 0:49Amazon Ads Products
- 2:03Customer Challenges
- 4:21Amazon Ads Ecosystem
- 6:16Metrics
- 7:09Return on Ad Spend
- 8:00Cyber Monday
- 11:43Persistence
- 12:45Learnings
- 14:27Challenges
related talks
09mUsing your technical skills with Amazon Ads
25mBest practices for advertising and marketing technology on AWS
44mBest practices for analytics and generative AI on AWS(ANT302)
29mUse sports data to elevate fan engagement and performance
06mAI and machine learning at Amazon
20mAI-Powered Customer Insights for Financial Services
01h 07mPower of Possibility: Navigating the tech landscape
48mBuilding Serverless AI powered Applications on AWS