Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Programmatic Advertising › Intermediate

Checkpoint: Ad Serving, Floodlight and Verification

Scenario questions on how an ad server counts and serves, how Floodlight activities should be counted and attributed, how to investigate a gap between publisher and ad server counts, and how verification and reconciliation decide what is paid for. Each answer gives the reasoning and the guide that teaches it.
CheckpointIntermediate levelModule 2: Ad serving and verification3 minUpdated

What this checkpoint covers

The questions draw on the ad serving basics guide, the Floodlight tags guide and the media auditing and verification guide. They ask what you would check, in what order, and what you would conclude.

Task

Write a short answer to each question before opening the explanations. Count an answer as right only if the reason matches, not just the conclusion. For every miss, note the fact you overlooked and reread the linked section.

Questions

  1. A publisher's invoice shows noticeably more impressions than CM360 recorded for the same placement. The account manager wants to dispute the invoice immediately. What do you do first?
  2. A direct placement shows impressions in CM360 but almost no clicks, while the publisher reports a normal click rate. What is the likely cause?
  3. A placement is serving only its default ad. What does that tell you?
  4. An online shop's Floodlight sales activity shows more orders than the order system. The thank-you page is often refreshed. What is wrong and how is it fixed?
  5. CM360 shows many more firings of a booking activity than DV360 attributes to its line items. The client asks which number is right.
  6. A display campaign reports a very low cost per acquisition, almost all of it from view-through conversions on cheap placements. What do you look at before celebrating?
  7. Verification shows a share of a guaranteed buy was invalid traffic, and the contract names the advertiser's ad server as the counting source. How is the invoice settled?
  8. An advertiser's agency holds the ad serving and verification contracts and supplies only summary reports. The advertiser asks whether that matters.

Answers explained

  1. Investigate before disputing. Publisher and ad server counts always differ, because they count at different moments. Confirm the time zones match, compare day by day to see whether the gap is constant or started on a date, break it down by placement, then inspect the implicated tag on the live page, and compare against the documented acceptable range (why publisher and advertiser counts differ).
  2. An unfilled click macro. If the publisher's ad server did not replace the click placeholder, or the tag was trafficked with it empty, the two systems are no longer counting the same click. Check the tag as pasted (ad tags, redirects and counting).
  3. A trafficking error. The default is a safety net that serves when no other ad qualifies; serving only the default points to ads that are mis-dated, mistargeted or unassigned (rotation, weighting and targeting rules).
  4. A random ord on a sales tag. Without the order identifier in ord, CM360 cannot deduplicate repeat loads, so every refresh becomes an order. Pass the order number and confirm with a reload test (counter versus sales tags and counting methods).
  5. Both, measuring different things. The CM360 activity count is the source of how many bookings fired; DV360 reports only the share it attributes to its own line items under its own rules. The difference is largely conversions display media did not touch (attribution and conversion windows).
  6. The view-through window. A long window on cheap, high-volume placements is the classic source of inflated conversion counts. Check the window was set deliberately and read click-through and view-through separately; Attribution credit is not proof of effect (attribution and conversion windows).
  7. On verified, valid delivery under the contract's terms. The ad server count is the basis, invalid traffic the contract excludes is removed, and any shortfall against the guarantee is settled by make-good or credit (reconciling delivery, invoices and the insertion order).
  8. Yes. The party being checked should not be the only one holding the evidence. The advertiser should hold the contracts directly or have contractual access to raw data and audit rights (independent media audits).

Where to go next

If you missed the Floodlight questions, redo the Floodlight specification exercise for a car rental company. Then continue to the measurement module from the intermediate programmatic level.

Pages this lesson applies

Terms used on this page

Also related

Related