Almost every audit you run on an ad account looks for something going wrong. This one looks for money left on the table. A campaign limited by budget while beating its ROAS target is the cheapest growth available in the account: the demand is proven and the only thing stopping it is a number you set. The trap is that average ROAS above target doesn't mean the next dollar will clear it. So finding budget-capped winners is only half the job. The other half is proving the win is real and the cap — not the audience — is what's binding.

In this guide: Why it matters · The two formulas · Four gates · By dimension · False positives · How to scale

What does "campaign limited by budget" actually mean?

On Google Ads, "Limited by budget" is a campaign status meaning the average daily budget is below what's needed to capture all available impressions at current settings. Google responds by showing ads less often, not by stopping them. On Meta there is no equivalent status label, but the behavior is the same: spend flatlines at the daily budget early in the day, day after day.

Neither signal tells you whether the extra delivery would be profitable. A budget-capped winner is the intersection of two conditions:

1. The cap is binding: spend is at or near the budget, consistently.

1. The campaign is winning: its cohort ROAS, measured by your MMP on matured cohorts, is above your target or well above the account norm.

Platform status gives you condition 1 on Google only. Condition 2 needs MMP cohort data joined to the campaign — which is why most teams never run this check systematically.

Why it matters: a worked example

From the PvX Delta audit spec's sample results (a fictional account; illustrative numbers). An iOS US campaign targeting value-optimized users:

• D30 ROAS vs account norm: 68% vs 41%, or 1.66×, the 88th percentile among tier-1 UA peers.

• Consistency: held for 30 days on matured cohorts.

• Daily budget: $800.

• Last three days' spend: $796, $800 and $791, for 99.5% utilization.

• When it caps: spend flatlines in the early evening, every day.

• Largest single payer's share of revenue: 9%, so no whale distortion.

• Headroom, new reach over 14 days: +11%, still growing.

• Headroom, frequency: 1.9 → 2.0 (+5%), far below a +25% saturation line.

• Last budget raise ($600 → $800): CPI moved only +4.8%, so marginal economics held.

Three campaigns are capped. Only one is a scale call. Illustrative data.

The action: raise the cap 25% to $1,000 a day in a single increment, not uncapped. Then watch CPI, frequency and early cohort quality daily, and stop raising when marginal ROAS stops clearing the 40% target.

The two formulas

Budget utilization = spend ÷ budget, over the same period. Read it daily over at least three days. A useful starting threshold is ≥95%.

ROAS vs norm = entity Dx ROAS ÷ account (or peer-group) Dx ROAS, on matured cohorts only. D7 for fast-monetizing apps, D30 where payers convert later — use the day your ROAS target is set on. See D7 ROAS explained.

Budget sits in different places depending on structure, so compute utilization at the level where the budget actually lives:

• Google App Campaigns: the budget lives at the campaign level. Compute campaign spend ÷ campaign budget.

• Meta with a campaign budget (Advantage campaign budget, formerly CBO): the budget lives at the campaign level. Compute campaign spend ÷ campaign budget; Meta decides each ad set's share.

• Meta with ad set budgets: compute ad set spend ÷ ad set budget.

• Ad networks with daily caps: the cap sits on the campaign or line. Compute spend ÷ cap.

Four gates before you call something a budget-capped winner

1. Is the win real? Consistent over 30 days on matured cohorts, above the statistical volume floor, not carried by one whale or one lucky geo. Young cohorts look bad for maturity reasons, and thin cohorts can look great for whale reasons (minimum sample size).

1. Is the cap actually binding? Spend flatlines at the cap early in the day, several days running. A campaign at 90% utilization that never hits the ceiling isn't capped; it's delivery-limited by something else.

1. Is there headroom? New reach still growing, frequency low, and CPI held last time spend rose. A capped campaign with rising frequency has audience as its real constraint — raising the budget buys repeat impressions, not new users.

1. What's the funnel role? A capped retargeting campaign has a finite pool. Raising its cap mostly buys frequency.

How to find campaigns to scale, by dimension

Campaign

The main reading level. Rank all campaigns at ≥95% utilization by ROAS versus norm, then run the four gates on the top of the list.

Ad set

Ad set ROAS clears the validity floor less often than campaign ROAS, so artifact risk is higher. Pool several weeks before you trust it. And with campaign-level budget optimization, the ad set's own cap may not be the limiter at all.

In a second spec sample, an ad set was beating a 40% D7 target by 14 points across four weekly reads, but the parent campaign's budget optimizer kept it pinned at about $110 a day. The fix wasn't a bigger budget. It was isolating the ad set into its own campaign and funding it directly.

Geo and OS

A capped campaign whose ROAS win comes from one country should be scaled in that country, not everywhere the campaign runs. Split ROAS by geo and OS before raising a multi-geo budget.

Network

Headroom varies by network. On Google App Campaigns, a budget increase gives the algorithm more room across Search, Play, YouTube and Display at once. On an ad network with thinner inventory, the same raise may saturate quickly.

False positives

• Immature cohorts. A campaign launched ten days ago can't show a real D30 ROAS. Wait for maturity.

• Whales. In a thin cohort, one large payer can double ROAS. Check the top payer's share of revenue.

• Retargeting mix. Retargeting and re-engagement campaigns often post high ROAS on a finite, already-converted pool. Their caps are rarely a scale opportunity.

• Deliberate caps. Test campaigns, cash-constrained months and geos under legal review are capped on purpose.

• Audience-limited, not budget-limited. Rising frequency plus flat reach means the pool is the ceiling. See audience saturation.

How to scale a budget-capped winner

1. Raise in steps of roughly 20–30%, not by uncapping. Large jumps reset learning on most platforms and can blow out CPI.

1. Watch CPI, frequency and early cohort quality (D1 retention, D1/D3 ROAS) after each step.

1. Stop when marginal economics stop clearing the target — the ROAS of the incremental spend, not the campaign average.

1. If the constraint is actually the audience, broaden targeting instead of raising the cap.

1. Prioritize this above most fix-it work. A capped winner is proven demand; most other findings are damage control.

When to scale is a broader question — when to scale a campaign covers order and pacing. The inverse problem, campaigns that can't spend their budget, is in budget pacing. For how to set the ROAS target these comparisons depend on, see our post What Should Your ROAS Target Actually Be?

Delta runs this check nightly — connect your data. Free, read-only, official APIs. See how PvX Delta works →

Sources

• Google Ads Help: fix "Limited by budget"

• Google Ads Help: recommended budgets for "Limited by budget" campaigns

Related checks

• When to scale a campaign

• Budget pacing: underdelivery and overdelivery

• ROAS dropping? Localize it by cohort, geo, network

• The PvX Delta morning decision brief

• The complete mobile UA audit checklist

• PvX Delta vs a manual UA audit