FroggyAds Insights
A campaign dashboard shows $420 for Tuesday, and the warehouse export shows $384. Before changing a bid or asking for a billing correction, check whether both reports describe the same slice of delivery. The two numbers can disagree because one report closes at a different hour, excludes a campaign, or was pulled before the last rows arrived. Start with the boundaries of the reports and work down to the rows that explain the gap.
Pin the day before comparing totals
Write down the account, currency, report time zone, date range and time of export for each file. Google's guide to account time zones explains that time segments in Google Ads use the account time zone. If your warehouse groups the same activity by UTC, events around local midnight can appear on different calendar dates. An export labelled Tuesday therefore needs a time-zone label too.
Choose one reporting zone for the reconciliation and convert the underlying timestamps before grouping by day. Use a named region when you have it, since the offset can change with daylight saving time. If the platform only gives you daily totals in its own time zone, compare those totals with a warehouse export rolled up to the same zone. A daily file has already lost the hour-level detail needed to split the amount across a different midnight.
Compare the same campaigns and the same cost measure
Pull one completed day at account and campaign level. Use stable IDs as the join keys, then attach names for readability. Names can change and two campaigns can have the same name. Check that both exports include the same campaign statuses, markets and buying accounts. If an account holds several currencies, keep them separate until you have a conversion rate and a documented conversion date.
Check the units in both files. If an API column stores millionths of a currency unit, convert once before adding values. When the dashboard displays rounded figures, use unrounded source values where available and round after summing. Keep taxes, service fees, billing adjustments and exchange differences in their own columns so a delivery cost can be compared with another delivery cost.
For source-level checks, the FroggyAds reporting overview describes spend reporting by source and zone. Start at campaign level even when those details are available. Drill into the source and zone only for the campaigns carrying a meaningful residual. This keeps a few late or duplicated rows from getting buried in a large export.
Run a settled-day check and explain the residual
Check one adjacent day on either side of the date in question. A time-zone mismatch often produces a shortfall on one day and an excess on the next. Add the adjacent days together and compare that total with the equivalent range in the other report. If the combined total agrees and the hourly rows move across the expected cutoff, you have evidence for a boundary issue. If the combined total still differs, keep investigating the source rows.
Keep a small reconciliation sheet with one row for each check. Record the difference before and after the change to the comparison. That makes it possible to see whether a corrected filter genuinely explains part of the gap or simply offsets another error.
Save the original downloads before refreshing them. Pull the same closed day again at a consistent time the next morning and record both retrieval times. Google's data freshness guide explains that reporting metrics arrive on different schedules and can be adjusted later. Comparing two pulls of the same report shows whether the gap is moving as data settles.
For example, suppose the $36 difference includes $25 from activity near the reporting boundary, $9 from a campaign filtered out of the warehouse view, and $2 that appears in the second pull. In that example the whole difference has an explanation, even though the original headline totals never match. Record the affected dates, campaign IDs and source files so the next person can repeat the check. Keep the numbers in the example separate from the results of your own account.
If a material difference remains, choose the smallest slice that still reproduces it. Look for an import missing one hour, a duplicated line after a retry, or a report view that combines costs from several accounts. Compare one campaign over the same time window with both raw extracts open. A row that exists twice on one side or never arrived on the other is more useful evidence than a screenshot of the aggregate.
Send the person who owns the affected report the exact account and campaign IDs, the zone, the report filters, the export times, and the remaining amount. Add a check for the specific cause once it is confirmed. A reconciliation record should let the next operator distinguish a recurring reporting boundary from a genuine change in spend before touching the campaign.