Auction Insights: Watching Competitors Without Spyware

Auction insights is the report where Google shows you the other bidders. For a campaign, ad group or keyword, it lists the domains that entered the same auctions you did and compares performance on the observable outcomes — who showed, who showed higher, who took the top of the page. It is competitive intelligence with three unusual properties: it is first-party, measured from real auctions rather than panel estimates; it is free; and it is symmetrical — everyone in it sees their own version of you.

How auction insights works

Six metrics, one shared comparison set:

Metric	The question it answers
Impression share	Of these auctions, how often did each domain actually show?
Overlap rate	When you showed, how often did this domain show too?
Position above rate	When you both showed, how often were they higher?
Top of page rate	How often did a domain's ads sit above the organic results?
Absolute top of page rate	How often in the first position specifically?
Outranking share	How often did you rank above them — or show when they did not?

Read together, the metrics profile each rival's posture. High overlap plus high position-above is someone consistently outgunning you. High top-of-page on modest impression share is someone bidding hard on a narrow slice. A domain present everywhere at low intensity is spraying the category broadly.

What it cannot tell you

The report shows outcomes, never inputs. Not budgets, not bids, not targets, not spend, not conversion economics — and not intent: a rising domain might be a strategic assault on your category or one intern experimenting with broad match. It also covers only auctions you entered; a competitor dominating auctions you never qualified for is invisible here. Treat it as radar rather than an intelligence file. It tells you something moved, roughly how large it is, and in which direction — the why still takes thinking.

The monthly pull that turns the report into radar

The report's chronic problem was cadence. Most advertisers open it twice a year, usually mid post-mortem — yet its value lives in deltas, and deltas need a rhythm. A monthly conversational pull fixes that at near-zero cost: "auction insights for the core campaigns against last month — new entrants, exits, and the biggest movements in overlap and position-above." The month a new domain appears at meaningful overlap, you hear about it while the response is still cheap.

When an aggressive entrant does show up, the same conversation grades the threat and drafts the response. Is your lost-to-rank rising in step — the cross-check that anchors the "why did performance drop?" workflow? Are they pressing your brand terms specifically, or the broad middle? Measured answers separate "defend the brand campaign now" from "note it and watch" — and either way the decision rides on evidence rather than the cold feeling of seeing a rival's name. Standing prompts for exactly this cadence live in the prompt library.