A prospecting campaign and a retargeting campaign can both look disciplined in their own dashboards. Give each one a cap of three impressions a day and a browser eligible for both may still receive six. The question to settle before launch is whose exposure you want to limit and which campaigns share the same counter. The number in the cap field only makes sense once you know those two things.
Start with the overlap
Write down the line items that can reach the same audience at the same time. A recent site visitor might be eligible for broad prospecting, a product retargeting line and a short promotion. The campaigns may live under separate insertion orders because they have different budgets. That separation can be useful for reporting, but it can also hide the fact that the same advertiser is competing for the same person's attention three times.
For each line, record the audience rule, the places it can serve and the identifier available for counting. Then pick the level that should share a cap. Microsoft Monetize's guide to applying caps illustrates how a parent cap can apply to its child line items as a group. Check the inheritance rules in the buying platform you actually use. Some settings apply across everything under a parent, while others only apply to the item you are editing.
Decide how the shared cap should be spent
Suppose the buyer wants to limit three overlapping lines to four impressions per day for the same recognized identifier. A shared daily cap of four prevents the lines from independently spending twelve impressions on that identifier. It does not reserve any impressions for a particular line. If one line tends to win the morning auctions, it can spend most of the shared allowance before another line gets a chance to serve.
Use line-level limits when that difference matters. If prospecting should use at most two of the four impressions, give it its own cap under the shared limit. A limit of two is still a ceiling. It does not guarantee the other lines will win the remaining impressions. Budget, bid and inventory still determine which eligible opportunities turn into delivery. In a self-serve setup, use a planning note alongside the FroggyAds DSP workflow to record the intended limits before entering them. That gives a second buyer something concrete to compare with the final campaign settings.
Set spacing separately if a burst is the concern. Four impressions in five minutes and four impressions spread across a day both fit the same daily limit. A minimum interval between impressions can prevent the burst when the platform supports a spacing rule at the right level. Record whether the daily cap resets at a calendar boundary or uses a rolling window. Those settings can produce different results just before and after midnight.
Check what the reports actually count
Start with a settled reporting period and use the same definition of an impression as the cap where the data is available. Google's guide to measuring reach and frequency explains that Display & Video 360 counts viewable impressions for display line item caps, while some reports also include impressions that were not viewable. A report can exceed a cap on paper simply because its numerator includes a different event. Make that comparison before opening a delivery incident.
If you can access an appropriate frequency distribution, look for the share of eligible identifiers above the intended threshold. An average alone can hide a small group receiving many more impressions than the rest. Separate traffic with a usable identifier from traffic the platform cannot consistently recognize. A browser, a device and a modeled person are different units. A cap enforced in one buying system also cannot account for impressions served through a separate system unless the systems actually share a counter. Document gaps without joining records across devices or sources that you do not have permission to combine.
When delivery drops after a cap change, compare the affected lines with a similar line you left alone. Check the time of the change, how much of the audience is still eligible and whether pacing or bids changed at the same time. A tighter cap can reduce impressions by design. It is more useful to see which repeat exposures disappeared and whether reach or conversions shifted than to treat the drop as a fault. Keep the revised cap when the distribution moves in the direction you intended and the resulting delivery still meets the campaign's purpose.