This case belongs to the advertising delivery reconciliation hub, which covers what to inventory before an ad platform replacement. This page does one job: walk through a single hard case so a partner can recognize the same pattern at a client.
The case in one paragraph
Illustrative and fictional. A 260-person regional digital publisher sells a regional hardware chain a spring garden center campaign of 2,000,000 impressions at a $12 CPM, worth $24,000, on a named list of local news and high-school sports placements. At the end of the flight the ad server reports 2,040,000 impressions. It looks like over-delivery. After the recount it is a shortfall of 140,000, repaired by a makegood the following month, and the invoice stands at $24,000.
Step by step: what went wrong and who decided
Step 1: the reported number
The ad server shows 2,040,000 impressions on the line item. The trafficker marks the line complete. Record captured: the line-item delivery report, with its date range and time zone. Nothing is wrong yet on paper.
Step 2: two findings that contradict it
The advertiser's verification report and the publisher's own audit disagree with the ad server. A targeting change mid-flight let 120,000 impressions run on a recipes and celebrity section that was not in the order. Separately, the verification vendor removed 60,000 impressions as invalid traffic. Records captured: the targeting change log, the order's placement list, the verification report and the audit note.
Step 3: the recount
The ad operations analyst subtracts the 180,000 ineligible impressions: 2,040,000 minus 180,000 leaves 1,860,000 eligible. Against the 2,000,000 ordered, that is 140,000 short, worth about $1,680 at the contracted rate. Record captured: the reconciliation worksheet or adjustment log, ideally with one row per removal and a reason.
Step 4: the decision
The advertising operations director chooses between a $1,680 credit and a 140,000-impression makegood on the contracted placements the following month. She proposes the makegood because the hardware chain has a summer patio flight already booked and the sports pages have unsold inventory in June. The chain's media agency accepts by email. Records captured: the proposal, the acceptance email and the new makegood line item.
Step 5: the checked result
The makegood line delivers 140,000 eligible impressions on the contracted placements. Eligible delivery now equals the 2,000,000 ordered, and the $24,000 invoice stands with no charge for the makegood. Records captured: the makegood delivery report, the final reconciliation and the invoice.
The decision log as an evidence asset
Illustrative. The same case, laid out as the company would want to find it years later:
| Step | What happened | Who decided | Record that captured it |
|---|---|---|---|
| 1 | Ad server reports 2,040,000 impressions | Trafficker marks line complete | Line-item delivery report |
| 2 | 120,000 off-contract and 60,000 invalid impressions found | Verification vendor and publisher audit | Targeting change log, placement list, verification report, audit note |
| 3 | Eligible recount 1,860,000; shortfall 140,000 | Ad operations analyst | Adjustment log with reasons |
| 4 | Makegood of 140,000 chosen over a $1,680 credit | Advertising operations director; agency accepts | Proposal, acceptance email, makegood line item |
| 5 | Makegood delivers; invoice of $24,000 stands | Billing lead confirms | Makegood delivery report, final reconciliation, invoice |
Why this sequence is useful evidence
A single delivery report is a number. The sequence is a decision under rules: a reported figure tested against a contract, two different grounds for removal, a choice between remedies with a reason, a counterparty's acceptance and a verified outcome. That is the shape of work AI developers want when they train or evaluate systems that handle exceptions, as described in reconciliation exceptions as AI agent training data and the broader exception handling guide.
Each record in the case has to state one detail, or a reviewer years later cannot redo the arithmetic. The table pairs this case's records with that detail and the published guidance behind it.
| Record in this case | The detail it must state | Why, per the published guidance |
|---|---|---|
| Verification report removing 60,000 | Whether the 60,000 were general or sophisticated invalid traffic | The Media Rating Council separates routine list-based filtration (GIVT) from cases needing advanced analytics or human review (SIVT) |
| Audit note comparing counts | The date range, time zone and line item it used | Google's discrepancy guidance says counts are only comparable on matching settings |
| Recount to 1,860,000 | That it starts from the settled count, not the first ping | Google's Open Bidding guidance says the billing ping fires before invalid-traffic and discrepancy adjustments |
| The 140,000 makegood line | A pointer to the line it repaired | Ad Manager's formal makegoods, offered on Programmatic Guaranteed deals and accepted by the buyer, keep the original rate and quantity limits; a direct-sold makegood like this one is an ordinary line, so the pointer is often only in its name |
The sequence is only useful if it survives intact. Three breaks are common: the adjustment log records a total without reasons, the makegood line item has no reference to the original, and the acceptance sits in a departed account manager's mailbox. How each link is joined across systems is mapped in advertising delivery reconciliation records across systems.
How to recognize similar cases at a client
Listen for these in conversation with an ad operations or revenue lead:
- "Reported delivery looked fine, but the verification numbers came in lower."
- "We had to back out impressions that ran on the wrong placements."
- "We usually offer a makegood before a credit."
- "Finance keeps an adjustment sheet with a reason for every change."
- "Agencies sign off on remedies by email."
Each phrase points to one link in the chain. Then ask four metadata questions: how many years of closed campaigns still have reported, eligible and billed counts; whether removals carry a reason; whether makegoods reference the original line; and who keeps the acceptance emails. Do not ask for reports, advertiser names or rates.
The person to start with is usually the advertising operations director; who can authorize a licensing discussion is a separate question, covered in who can license advertising delivery reconciliation records. The company must meet the baseline: 50+ full-time employees at peak (contractors excluded), documented operating history, rights to license and an authorized contact. Companies in the U.S., Canada and the EU qualify, as do companies in the UK, Australia, New Zealand, India, Singapore, South Africa, Latin America, the UAE and Saudi Arabia when at least 70% of the data is in English. Unknown until SourceX reviews it: buyer interest, the rights result, what must be de-identified and any offer. Publishers sit with other sectors under media and production.
Next step
Refer a company. No cap on rewards. Default reward on a company's first qualifying transaction: the greater of $50,000 plus 5% and 15% of Actual Supplier Payout, with no cap; further qualifying transactions first agreed within 12 months of the first Closing pay 15%. Eligibility and payment conditions apply, including the USD 100,000 Minimum Transaction, and a signed Schedule A or addendum can differ. A lead, meeting or signature alone does not trigger payment.
If a publisher client recognizes this sequence in years of closed campaigns, register as a partner, check the company against who qualifies and read the program terms before introducing its decision-maker.