Design a real-time ad bidding exchange: system design interview question

How the interviewer opens it

“Design a service that runs auctions between publishers and advertising bidders for individual ad impressions. The service must enforce tight deadlines, handle bidder failures, and record auction outcomes reliably.”

The prompt is vague on purpose. The first minutes are for you. Reduce the prompt to the core features and agree on the numbers before you draw.

What to clarify first

Traffic volume
Read:write mix
Bidder calls and payloads
Latency
Auction rules
Bidder failures
Availability and consistency
Outcome durability
Retries
Scope limits

The interviewer has the answers. In a SysRound round, Alex gives these answers, as a real interviewer does.

Scope of a strong answer

Core requirements

Accept an impression request and contact eligible bidders
Select a valid winner before the deadline
Continue when some bidders fail or reply late
Record each completed auction outcome reliably

Usually out of scope

Select ads through user profiles
Manage campaign budgets or spend pace
Charge advertisers or pay publishers
Display ads or confirm ad views
Support service across multiple regions

What interviewers ask about

Auction deadlines: Each auction waits on several external bidders. One slow bidder can consume the time that the exchange needs to save and return the result.
Which bid wins: The exchange must compare only valid, timely bids. A fast reply or a large price does not alone make a bid eligible.
Auction outcomes: The exchange can save a result without delivering it. Retries must preserve one outcome despite worker failures and uncertain write acknowledgments.

Common mistakes

Wait for every bidder before selecting a winner.
Ignore the extra traffic that each auction creates through bidder calls.
Use the entire deadline for bidder responses and omit persistence time.
Return a winner before the service confirms a durable outcome.
Treat a retry as a new auction or treat an auction win as a confirmed ad view.
Add campaign budgets, payments, or user profiles before solving the stated auction problem.

The bar by level

Senior (L5/E5): Builds a complete latency budget and estimates bidder traffic. Bounds failures, validates bids, and preserves one durable outcome across retries and uncertain writes.
Staff (L6+/E6+): Defines failure boundaries across bidder calls, durable state, and publisher responses. Quantifies capacity limits and overload behavior. Addresses zone failures, abandoned auctions, and rule changes. Report export and repair are optional stretch criteria.

Sources · 1 reference

These records explain where the question appears. A company tag or candidate report does not guarantee a future interview question.

SysRound · 1 record

No company association
Design a real-time ad bidding exchange
Reported date not specified
Source snapshot: Nov 12, 2026
View source

How the interviewer opens it

“Design a service that runs auctions between publishers and advertising bidders for individual ad impressions. The service must enforce tight deadlines, handle bidder failures, and record auction outcomes reliably.”

The prompt is vague on purpose. The first minutes are for you. Reduce the prompt to the core features and agree on the numbers before you draw.

Scope of a strong answer

Core requirements

Accept an impression request and contact eligible bidders
Select a valid winner before the deadline
Continue when some bidders fail or reply late
Record each completed auction outcome reliably

Usually out of scope

Select ads through user profiles
Manage campaign budgets or spend pace
Charge advertisers or pay publishers
Display ads or confirm ad views
Support service across multiple regions

What interviewers ask about

Auction deadlines

Each auction waits on several external bidders. One slow bidder can consume the time that the exchange needs to save and return the result.

Which bid wins

The exchange must compare only valid, timely bids. A fast reply or a large price does not alone make a bid eligible.

Auction outcomes

The exchange can save a result without delivering it. Retries must preserve one outcome despite worker failures and uncertain write acknowledgments.

Common mistakes

Wait for every bidder before selecting a winner.

Ignore the extra traffic that each auction creates through bidder calls.

Use the entire deadline for bidder responses and omit persistence time.

Return a winner before the service confirms a durable outcome.

Treat a retry as a new auction or treat an auction win as a confirmed ad view.

Add campaign budgets, payments, or user profiles before solving the stated auction problem.

The bar by level

Senior (L5/E5)

Builds a complete latency budget and estimates bidder traffic. Bounds failures, validates bids, and preserves one durable outcome across retries and uncertain writes.

Staff (L6+/E6+)

Defines failure boundaries across bidder calls, durable state, and publisher responses. Quantifies capacity limits and overload behavior. Addresses zone failures, abandoned auctions, and rule changes. Report export and repair are optional stretch criteria.