Closed hillslatt closed 6 months ago
Attribute names related to Interest Group Auctions are prefixed with "ig", might be clearer and easier to recognize them if they were prefixed with "iga".
Agreed with @bmayd, I believe we should use:
igab
instead of igb
(since it corresponds to InterestGroupAuctionBuyer
)igas
instead of igs
(since it corresponds to InterestGroupAuctionSupport
)In the same vein, I believe we should use pbps
instead of ps
(since it corresponds to perBuyerPrioritySignals
)
Consider omitting Auction from the object names, as it's redundant. Everything in ORTB, and particularly this subset, ultimately involves an auction.
Consider omitting Auction from the object names, as it's redundant. Everything in ORTB, and particularly this subset, ultimately involves an auction.
Yes, I think that makes sense so long as we don't anticipate other sorts of Interest Group related naming that might make "ig" too terse.
Regarding Interest Group Auction Signaling. What about perBuyerExperimentGroupIds? Should there be a field as part of the InterestGroupAuctionBuyer that allows to propagate experimentGroupId? Maybe something like eid field.
Hello. perBuyerExperimentIds
were discussed, but omitted from this initial, provisional spec. "pending conversation around async calls with GAM, Prebid and SSPs."
See comments history in the working draft.
cc: @hillslatt
Should poddedupe also apply to Object: Audio ? We send bid requests to DSPs asking for multiple imp objects to go into a single ad pod in an audio stream and will de-duplicate any responses that are for the same creative or advertiser domain etc. I'm not sure if any of the DSPs currently responding with duplicates would appreciate the extra signal and do anything with that information.
Wrt Interest Group Auction Signaling - shouldn't "biddable" default to 1? Today, should a buyer receive for example a Native/Video/Display request or some other - it is implied that the buyer is allowed to participate/bid. Should this behaviour need be different - wseat/bseat and similar controls allow seller to signal that explicitly show who is allowed to bid, albeit on a callout/bid request level, not a specific format
@piwanczak I replied/opened a thread re. biddable
on the PR.
Repeating this comment here:
Object: AuctionGroupIntent Object: InterestGroupAuctionBuyer
assume that the openrtb request is made by an ssp to a dsp requesting per buyer signals to include on its paapi request; but often an openrtb request is made to a supply platform by a publisher, or between supply platforms. eg (https://github.com/prebid/prebid-server/issues/2411)/
In this case we'd need a response object with a complete auctionconfig to be submitted or merged with another, not the limited subset of aucionconfig in thie response object here. We're seeing responses vary enormously by prebid implementors
rtbh does this https://github.com/prebid/Prebid.js/blob/7d3ff582753179b6fb8272122a2d6dea8b0cbab2/modules/rtbhouseBidAdapter.js#L156
magnite puts it here https://github.com/prebid/Prebid.js/blob/7d3ff582753179b6fb8272122a2d6dea8b0cbab2/modules/rubiconBidAdapter.js#L731C32-L731C43
openx puts it here https://github.com/prebid/Prebid.js/blob/7d3ff582753179b6fb8272122a2d6dea8b0cbab2/modules/openxBidAdapter.js#L105
Openrtb responses need a place for complete configs and we shouldnt publish the per buyer signals response standalone spec until we propose one for the complete object as well
I propose Object: InterestGroupAuctionSeller (PROVISIONAL) with pretty wide open formatting
I opened #89 to express some feedback we saw Criteo share about ae not matching in place definitions.
Please comment on this issue with feedback on updates to the OpenRTB updates in Public Comment
ifa_type
in Object: Device associated to List: Ifa_Typeaint
attribute to Object: Imp,inttype
attribute to Object: Bid both associated to List: Interaction TypesInterest Group Auction Signaling
igs
attribute in Object: Imp