← All questionsLimitations: Budget, Bids & Search Volume

How do I estimate the volume I'm missing out on due to budget and rank limitations?

A quick primer

Impression Share (IS) is the share of impressions a campaign got out of the total number it was eligible for (that is, all the impressions its ads could theoretically have competed for in the auction). Google Ads breaks the unmet share down into two separate, independent components: Search Lost IS (budget) — the share of impressions missed specifically due to a budget shortfall, and Search Lost IS (rank) — the share missed due to an insufficiently high ad rank (usually tied to bids, the target, or ad/page quality). The official description of this metric and its formula is in Google Ads Help's article on Impression Share.

Here's a key practical rule: the sum of Impression Share + Lost IS (budget) + Lost IS (rank) should come out to roughly 100% (allowing for rounding). This means these three numbers aren't scattered metrics — they're parts of one whole, and reading them together gives a precise picture: if, say, IS = 40%, Lost IS (budget) = 45%, and Lost IS (rank) = 15%, that clearly shows budget is the dominant constraint, and that's where to start (more on how to act in that case in Google Ads Help's article on fixing the "Limited by budget" status). If the ratio were reversed, priority would shift to the target/bids, and it'd be worth checking bid strategy statuses. It's also worth remembering these metrics are also calculated at the exact-match level (Search Exact Match IS) — a narrower but sometimes more telling picture specifically for core keyword segments.

What to check before you touch anything

  • The exact current values of all three components — Impression Share, Lost IS (budget), Lost IS (rank) — over the same period, so they can be correctly compared to each other.
  • What period these metrics are calculated over — weekly and monthly figures can differ noticeably if there's seasonality or recent account changes, so it's important to look at comparable-length windows.
  • Are the metrics calculated at the whole-campaign level, or is it worth breaking them out by ad group/keyword? An account-wide aggregate can mask the fact that one specific group is severely constrained while others aren't.
  • Could a high Lost IS (rank) be the result of several factors at once — the target, technical bid limits, low Quality Score — since rank-lost by itself doesn't separate these causes from each other?
  • How stable are these figures over time? A one-off spike in Lost IS during a single week could be a temporary factor (competition, season) rather than a persistent, structural constraint.

Possible approaches

  • Use the simple arithmetic rule (the three components sum to roughly 100%) as a quick check — whichever type of Lost IS is numerically bigger is the more significant limiting factor at the moment.
  • If Lost IS (budget) dominates, move toward questions about the necessary budget increase, rather than touching bids or the target.
  • If Lost IS (rank) dominates, move to more detailed diagnostics: is it caused by the target (Bid strategy constrained by target), technical bid limits (Limited by bidding strategy), or poor ad/page quality (Quality Score)?
  • Break the metrics down not just at the campaign level, but at the level of specific ad groups or keywords if the overall aggregate seems contradictory or not informative enough to make a decision.
  • Evaluate these metrics regularly (say, weekly) as part of routine monitoring, not only once a problem is already suspected — this helps you catch a shift toward one constraint before it becomes critical.
  • Manually pulling all three Lost IS components for every campaign, group, and keyword, then cross-referencing them against each other and against several weeks of history, is routine, high-volume work across dozens of active campaigns. Our tool (DataMind) calculates and breaks down Lost IS by budget and rank for every campaign automatically, immediately showing which factor dominates and how stable that picture is over time — instead of manually pulling and assembling data at every level.