A frequency export may give you only rows marked “1+”, “2+” and “3+”. Each row includes everyone who reached the next threshold. To find how many people saw the ad exactly twice, subtract the “3+” count from “2+”. With a few rows, you can reconstruct the exposure groups the export supports and work out the fewest impressions they could account for.
Read the bucket labels before adding anything
Frequency is the number of counted impressions per person or identifier in a defined period. Reach is the number of distinct people or identifiers represented in that report. Use the platform's definition for both. The word “user” in an export can refer to a modeled person, an account or a device, depending on the measurement method. Keep the same definition, date range and campaign filters in every column you compare.
Look at the symbols next to the counts. A row labeled “3+” usually means everyone who received three or more impressions. A person in that row is also in the “2+” and “1+” rows. A row labeled “exactly 3” is an exclusive bucket. Someone belongs to that bucket only if they received three impressions in the selected period. If the export has ambiguous labels, open the metric definition before deciding how to calculate shares. One documented example of cumulative frequency metrics is the Display & Video 360 reach and frequency guide.
Rebuild the distribution from a small example
Suppose an export shows 1,000 people at “1+”, 420 at “2+”, 180 at “3+” and 60 at “4+”. To get the count for exactly one exposure, subtract 420 from 1,000. That leaves 580. Subtract 180 from 420 to get 240 for exactly two, and subtract 60 from 180 to get 120 for exactly three. The last 60 people stay in an open-ended bucket because this export stops at four or more. The four groups add up to 1,000.
You can now say that 58 percent of the measured audience saw the ad once and 6 percent saw it at least four times. The first group accounts for 580 impressions, the second for 480, the third for 360 and the last for at least 240. That adds up to a minimum of 1,660 impressions. Any fifth or later exposure adds to that minimum. If the same report shows 1,900 comparable impressions, the remaining 240 can come from exposures beyond four. If it shows 1,500, pause and check the inputs before passing the report along.
Match the impression count to the reach count
Use the lower-bound check only when the impression total and frequency buckets describe the same set of counted events. One report may include all served impressions while another uses only viewable impressions or reports reach for a narrower set of placements. Modeled or rounded figures may also differ slightly. Save the metric labels with the export so the next reader can see whether the figures are meant to reconcile exactly.
Check the dates and filters next. A Monday-to-Sunday report and a seven-day rolling report can cover different hours, even when both display seven dates. A placement filter can change the audience behind an otherwise familiar campaign label. If you have device or placement breakdowns, keep those rows for diagnosis and use the platform's deduplicated total for overall reach. A person who appears in two rows can be counted in each row.
Explain what the distribution can support
Put the single-exposure share, the top reported threshold and the definition of reach beside the average. Say how far the buckets go. “Six percent received four or more exposures” is useful even when the export cannot tell you whether those people received five or twenty. If you are comparing two periods, use the same bucket threshold for both and check whether the audience size or the measurement coverage changed.
Suppose the top share rises while reach stays flat. That gives you a specific next check. Look at whether a placement or audience segment took a larger share of delivery during the period. If the measurement holds and you want to test a delivery limit, review the serving scope and counting window before setting one. The FroggyAds guide to frequency capping covers those setup decisions.
Close the report with a line someone else can audit. “Same campaign and filters, compared across two complete weeks. The four-or-more share rose from 6 to 9 percent while measured reach was steady. Next check is the placement mix.” Keep the original export beside that line. If a teammate challenges the conclusion, you can trace it back to the exact row and count that prompted the check.