Trace an event from click to analysis
end-to-end understanding of an analytics event pipeline.
client SDK captures and batches, a collection endpoint ingests, a stream and ETL enrich and load into a warehouse for analysis.
What's really being asked
This is a foundational system-design question checking whether you understand how product analytics actually flow, not a textbook description of A/B math. The interviewer wants the stages, the components at each stage, and an appreciation that ingestion is asynchronous and built for scale and reliability.
The full answer
The lifecycle begins on the client. An analytics SDK fires when the user clicks, capturing the event name, the experiment variant the user was assigned, a stable user id, a timestamp, and context like platform and app version. The SDK typically batches events and sends them over the network, retrying on failure with idempotency keys so a retry does not double-count. A collection endpoint, often a lightweight ingestion service behind a load balancer, validates the schema, attaches server-side metadata, and quickly acknowledges. From there events land in a streaming buffer such as a message queue or log, which decouples ingestion spikes from downstream processing. A processing layer consumes the stream, enriches, deduplicates, and writes the events into a data warehouse or columnar store. Finally, analysts or the experimentation platform query the warehouse, grouping conversions by variant to compute the difference between button colors. The variant assignment itself comes from a feature-flag or assignment service consulted at render time.
The mistakes people make
Proposing a synchronous insert into a production database on every click, which does not scale and couples the app to analytics. Forgetting batching and retries, so events are lost on flaky networks. Omitting the assignment step, so you cannot attribute clicks to a variant. Treating ingestion as instantly queryable rather than an asynchronous pipeline.
What usually comes next
How do you guarantee a user always sees the same variant? Deterministic hashing of the user id. How do you handle dropped or duplicate events? Idempotency keys and dedup in processing. What is the latency from click to dashboard, and why is it not instant?
A concrete example
A user opens the app; the assignment service hashes their id and shows the green button. They tap signup; the SDK records signup_clicked with variant green and a session id, batches it, and posts it to the collection endpoint, which acknowledges and pushes it to a queue. A stream job enriches and loads it into the warehouse, where the experiment dashboard later compares signup conversion for green versus the control blue.
Interview question
Why do analytics SDKs batch events and send them to an ingestion endpoint rather than writing each click directly to the production database?
- a.It is the only way to guarantee consistent variant assignment
- b.Batching plus an asynchronous pipeline scales better and decouples analytics from the app's critical pathCorrect
- c.Direct writes would reveal the user's variant assignment to other users
- d.Databases cannot store click events without a streaming buffer
Why? this is the answer
Buffering and asynchronous ingestion absorb traffic spikes and keep analytics from slowing or coupling to the app's request path. Variant assignment is handled separately by deterministic hashing, not by the write strategy.
Just read this? Test yourself on what you have been reading.
Read the original → systemdesignhandbook.com
- #analytics-pipeline
- #ab-testing
- #system-design
- #event-tracking
- #data-engineering
Put your scrolling time to good use
Learn one idea, try a quiz and save useful cards for revision. Tezvyn makes it easy to learn and stay current in your tech field, a few minutes at a time.
The iPhone app is on the way
We are building it. Until it lands, nothing here is held back from you: every interview card, your saved cards, streaks and the job board all work in Safari, plus hundreds of free practice quizzes of thirty questions each. Sign in and it all carries over to the app the day it arrives.
Want it as an icon? Tap Share at the bottom of Safari, then Add to Home Screen. It opens full screen and the cards you have read stay available offline.
We are hiring for this. Every open role lists the topics its interview covers, so you can prepare for the real thing rather than guessing.
See open roles