Ads on Live Blogs: Keeping Requests and Timestamps Straight

I run ads on a couple of live blogs, and the hardest part is not selling the inventory. It is keeping the request log readable when a post updates every few minutes. A live blog fires new ad calls as readers refresh, as modules lazy-load, and as editors publish. If I do not label those calls carefully, my reports turn into a pile of near-identical rows.

My first rule is to treat every request as an event with a timestamp I can trust. I use the ad server’s time zone consistently and note it in the report name. When I compare a spike in requests with an editorial update, I want to see the same minute, not a guess. I also keep a simple sheet of update times during big events. That sheet has saved me from chasing duplicate traffic that was really just a busy live post.

Duplicate requests are the next issue. Some come from a slot refreshing in place. Some come from a reader scrolling back to the same module. I log the slot name and the refresh reason so I can tell them apart. If a slot refreshes on every update, I ask whether that is worth it. Say a plan costs $150 for a test; I would rather spend it on two clear placements than one noisy slot.

For reports, I filter by page section, device, and timestamp window before I judge performance. Mobile readers often refresh more often and see fewer slots per screen, so I check mobile separately. I also keep a short list of what I verify:

1. Confirm the time zone in every export.
2. Tag live updates with a consistent label.
3. Separate refreshes from new page views.
4. Check mobile traffic on its own.

When numbers look odd, I go back to the raw request notes before I change a campaign. I keep my own live blog ad notes for that review, and I check the provider’s current terms for anything I am unsure about. The goal is simple: fewer mystery rows, cleaner decisions.

— Daniel Hayes