An unprofitable campaign is not always a failed hypothesis. Early tests may have few conversions, while advertisers can take time to approve leads. But that is no reason to keep increasing spend without a plan. A stop decision should be tied to conditions defined before launch. Pause decisions for Meta Ads and Google Ads require channel-specific signals plus an overall risk limit.
Set a stop-loss first
A stop-loss is the maximum spend or loss allowed for a hypothesis. It should reflect available risk capital and offer economics, not a desire to recover previous losses. Record the total cap, maximum spend without an approved event and a mandatory review point.
For example, an approved lead pays $30 and the target media CPA is at most $22. If an agreed amount of spend produces no approved leads and tracking is verified, the test needs a critical review. The exact threshold depends on expected conversion probability and acceptable risk.
When to stop immediately
Pause if the landing page fails, tracking stops, the offer closes, its cap is reached, the promoted product violates applicable policies or traffic becomes clearly invalid. Spending through a technical failure does not create useful evidence.
An unprofitable campaign is not always a failed hypothesis. Early tests may have few conversions, while advertisers can take time to approve leads. But that is no reason to keep increasing spend without a plan. A stop decision should be tied to conditions defined before launch. Pause decisions for Meta Ads and Google Ads require channel-specific signals plus an overall risk limit.
Set a stop-loss first
A stop-loss is the maximum spend or loss allowed for a hypothesis. It should reflect available risk capital and offer economics, not a desire to recover previous losses. Record the total cap, maximum spend without an approved event and a mandatory review point.
For example, an approved lead pays $30 and the target media CPA is at most $22. If an agreed amount of spend produces no approved leads and tracking is verified, the test needs a critical review. The exact threshold depends on expected conversion probability and acceptable risk.
When to stop immediately
Pause if the landing page fails, tracking stops, the offer closes, its cap is reached, the promoted product violates applicable policies or traffic becomes clearly invalid. Spending through a technical failure does not create useful evidence.
Reaching the agreed loss cap or a payout reduction that invalidates the economics is another clear stop signal.
When continuing is justified
If leads are pending approval, do not classify them as either approved or rejected. A small sample may also be inconclusive. Continued testing only makes sense within the budget and with functioning measurement.
In some cases waiting for existing leads to mature is more informative than buying additional clicks.
A pause is not always the end
A broken form can be fixed and the experiment rerun. Persistently expensive approved conversions, however, require a different hypothesis—not merely more budget.
A practical decision framework
For each campaign track spend, pending and approved events, cost per approval, allowable loss, rejection reasons and data reliability. At each checkpoint choose one action: continue, repair, pause or close the hypothesis.
Distinct stop signals for Meta Ads and Google Ads
In Meta Ads, rule out Pixel/CAPI and landing-page failures, then review frequency, creative performance and approved acquisition cost. Deterioration after repeated edits may reflect delivery changes rather than the creative alone. In Google Search, inspect queries, ad groups, traffic quality and conversion actions; other campaign types need different checks. A CPC threshold from Meta cannot be copied to Google without conversion and approval context.
If one source reaches its stop-loss, the other does not automatically need to stop: acquisition costs and approval maturity may differ. However, a shared budget cap should limit total exposure. If Search has expensive clicks but low approved CPA while Meta has mostly pending leads, compare mature cohorts and resist spending more merely to recover prior losses.
Bottom line
Knowing when to stop spending is as important as knowing how to launch. Verified evidence and pre-set risk limits should outweigh the temptation to chase sunk costs.
Reaching the agreed loss cap or a payout reduction that invalidates the economics is another clear stop signal.
When continuing is justified
If leads are pending approval, do not classify them as either approved or rejected. A small sample may also be inconclusive. Continued testing only makes sense within the budget and with functioning measurement.
In some cases waiting for existing leads to mature is more informative than buying additional clicks.
A pause is not always the end
A broken form can be fixed and the experiment rerun. Persistently expensive approved conversions, however, require a different hypothesis—not merely more budget.
A practical decision framework
For each campaign track spend, pending and approved events, cost per approval, allowable loss, rejection reasons and data reliability. At each checkpoint choose one action: continue, repair, pause or close the hypothesis.
Distinct stop signals for Meta Ads and Google Ads
In Meta Ads, rule out Pixel/CAPI and landing-page failures, then review frequency, creative performance and approved acquisition cost. Deterioration after repeated edits may reflect delivery changes rather than the creative alone. In Google Search, inspect queries, ad groups, traffic quality and conversion actions; other campaign types need different checks. A CPC threshold from Meta cannot be copied to Google without conversion and approval context.
If one source reaches its stop-loss, the other does not automatically need to stop: acquisition costs and approval maturity may differ. However, a shared budget cap should limit total exposure. If Search has expensive clicks but low approved CPA while Meta has mostly pending leads, compare mature cohorts and resist spending more merely to recover prior losses.
Bottom line
Knowing when to stop spending is as important as knowing how to launch. Verified evidence and pre-set risk limits should outweigh the temptation to chase sunk costs.