prebid / headerbid-expert Goto Github PK

Hi

What is the license of the files in this repo? I want to use it for my school project if allowed by license.

Thanks
Joe

It seems that the Header Bid Expert shows way to high measurements when a delivery through yieldlab is done.

The measurement is correct as long, as only the header bidding call is done, but when the delivery call is added and the counting pixels are fired it seems, that they are all included in the measurements, ending in a way to high time span.

To reproduce it:

• open the developer tool bar of your browser and go to the network tab
• filter on "yieldlab"
• navigate to https://kachelmannwetter.com/de

If you only see a /yp call than the given time span will most likely be ok. But once an actual ad is delivered with a /d call, than the timespan will be far to wide.

If you got problems getting a delivery, I recommend visiting https://www.otto.de/ or https://www.zalando.de/herren-home/. Just look at some items of these stores and you will hopefully get their ads.

According to [individual ad unit refresh] (http://prebid.org/dev-docs/examples/adunit-refresh.html), suppose that we have 6 ad units on page.
All of them are loaded through RefreshBid function apart from each other.
3 of them are loaded on page load; the remnant are lazy loaded.
On page analysis HB Expert is always throwing a warning, complaining that the page is under monetized; but as the bids are concurrent, it seems to me that the time bars (blue and orange) are cumulative and do not keep track of separate auctions.
This hypothesis seems to be confirmed when the lazy loaded units are loaded (screenshot).
So, my question is: in this particular case is HB Expert analysis reliable?

New Prebid Server Endpoint:

https://prebid.adnxs.com/pbs/v1/openrtb2/auction

Amazon UAM Bidder:

https://c.amazon-adsystem.com/e/dtb/bid?src=xxxx&u=xxx&pr=xxxx&v=7.43.01&t=3000&slots=xxx&pubid=xxxx

just some random text...

no really, that's all it is... :(