Practical framework / 5 MIN READ
Explain impression differences at the point where each system counts
Build a delivery discrepancy bridge before treating one platform as the truth.
AI-assisted research and analysis for the SNS Creative Fieldbook
Sources checked 2026-11-06
A buying platform and an ad server may record different stages of ad delivery. Their impression totals can diverge even when both systems follow their own definitions, while an unusual gap can also reveal a technical or reporting problem that needs investigation.
What you will leave with: Create a matched delivery comparison with scoped differences, supporting evidence and an accountable next step.
Use this when
- An agency and publisher disagree about delivered impressions.
- A new creative or tag produces a sudden reporting gap.
This is a measurement investigation, not an invoice ruling or a universal acceptable-discrepancy threshold.
Align the comparison before calculating the gap
Match campaign, placement, creative, date range, timezone and inventory format. Use reports with comparable processing maturity and document any excluded traffic. Check whether one extract includes defaults, tracking-only ads or a broader set of placements than the other.
Write the discrepancy formula with its denominator. If one system reports 10,000 impressions and another 9,000, the gap is 10% of the first count but about 11.1% of the second. Neither percentage explains the cause. Show both counts so readers can reproduce the calculation and understand the practical scale.
Map the counting events
Google’s Display & Video 360 documentation lists counting points, processing, filtering, latency and scope among reasons systems disagree. Use that guidance to ask what event each platform records for the specific format and integration being compared, rather than assuming every impression means the same technical milestone.
Draw the delivery sequence from request through creative response and rendering. Mark where failures, navigation away or blocked requests can interrupt it. Keep the explanation specific to verified behavior. A generic list of possible causes is a starting checklist, not evidence that a particular gap has been explained.
Localize the difference with comparable slices
Break the matched comparison into useful units such as day, placement, format or creative version. Look for a change point connected to a tag release, hosting change or reporting revision. Avoid slicing into tiny groups where a handful of records creates a dramatic percentage.
Use approved diagnostics and sanitized identifiers. Compare a permitted synthetic delivery test with the production pattern where appropriate. Do not repeatedly click live paid ads or bypass privacy controls to force agreement. If the discrepancy appears downstream of a third-party tag, give its owner the minimum evidence needed to investigate.
Resolve the decision without forcing equality
Classify the gap as explained methodology, reporting mismatch, confirmed defect or unresolved difference. Record the evidence and whether it affects optimization, campaign acceptance or a commercial discussion. A vendor’s general statement that discrepancies are common is not an agreed billing tolerance.
Follow the contract’s measurement basis through the authorized commercial owner. Do not overwrite one count with the other merely to close a report. After a fix, compare a new matched period and preserve the historical note. A useful resolution explains what each number means and what action follows.
Make the reasoning concrete.
A hypothetical buying-platform report shows 10,000 impressions and the ad server shows 9,000 for a matched period. Most of the 1,000-impression difference appears on one newly introduced creative format. The pattern narrows the investigation but does not yet prove a rendering defect.
Apply a generic tolerance
Opening: Dismiss the gap because someone says platform discrepancies are normal.
Hypothesis: No further analysis is required.
Risk: A format-specific defect or contractual issue may remain hidden.
Trace the affected slice
Opening: Verify counting definitions and inspect the new format’s delivery path.
Hypothesis: The team has a testable explanation and a responsible owner.
Risk: The eventual commercial treatment still depends on agreed terms.
Investigate the concentrated difference and document the verified cause or remaining uncertainty. Keep the technical conclusion separate from any invoice or compensation decision.
Take the questions into your next brief.
Use these prompts in your own notes. Nothing is submitted from this page.
Scope
Do both extracts cover the same delivery population?
Formula
Which count is the discrepancy denominator?
Milestone
What technical event does each impression represent?
Pattern
Where and when does the gap concentrate?
Resolution
What evidence changes the technical or commercial next step?
Watch for the wrong conclusion.
- Comparing preliminary interface totals with mature reports.
- Hiding the denominator of a discrepancy percentage.
- Using a generic tolerance as a contract term.
- Forcing counts to agree without explaining the difference.
Your next action
Build one matched impression comparison and trace its largest material discrepancy to a specific delivery stage.
Sources & scope
Frameworks, worksheets and hypothetical examples are original editorial synthesis. Linked primary sources support only the stated concepts. Source checking is not an independent audit of effectiveness. No named human author or reviewer is claimed.
- Google Display & Video 360 Help: Reporting Discrepancies →Primary technical documentation · Supports: Documents differences from counting points, processing, filtering, latency and reporting scope. No generic discrepancy percentage is adopted as an acceptable commercial tolerance.