Auction Architecture: A Terminology Reference Guide

CloudX is a complete AI-powered monetization platform. It can be used as a standalone solution, or in concert with existing ad infrastructure. 

Across digital advertising, publishers have long benefited from running multi-platform setups. Multi-platform stacks have been standard in web, CTV, and non-gaming mobile for years. Mobile gaming is catching up, and publishers are exploring how to structure their stacks to increase control, flexibility, and yield.

CloudX is designed to flexibly support various standard multi-platform architectures. Each has its own tradeoffs, and the right setup depends on your inventory, your partners, and your goals. Multi-platform stacks can deliver meaningful benefits, but complexity can come at a cost. We're investing in tools to give publishers the upside without the overhead.

The terminology used to describe these architectures isn't always applied consistently, so we've laid out what each term traditionally means in digital advertising.

True First Look

One platform receives the ad request first and runs its auction. If a bid clears the publisher's configured floor, the impression fills and the request goes no further. Impressions filled here never reach subsequent platforms.

First-Look: CloudX Fills

First-Look: CloudX Passes

Pre-Bid

One platform runs an auction before the next ad server is called, generating a winning bid. That bid then competes in the next ad server auction alongside other demand sources, but the other participants do not know the pre-bid platform's winning bid before they submit their own. They compete independently.

The defining feature: the two auctions are informationally independent.

Parallel

Two or more platforms receive the ad request simultaneously and run independent auctions. Each generates its own winning bid. The publisher's ad server determines which to serve. Neither platform has visibility into the other's auction while it's running.

Post-Bid / Last Look

One platform runs and generates a winning bid. A second auction then runs, but the bidders in that second auction do know the first platform's winning bid before they submit their own. The second auction participant isn't competing blindly; they see the price they need to beat before deciding whether to bid.

This is also called last look. This is the critical distinction from pre-bid: in pre-bid, the second auction is blind to the first auction's result. In post-bid, it isn't. Whether the second auction knows the first auction's winning bid in advance, that's the line between the two architectures.

Summary