Protected Audience auctions are splitting your reporting
Confirmed: on-device Protected Audience (formerly FLEDGE) auctions resolve separately from the contextual auction, and GAM reports them in a way that can double-count or strand impressions if your setup isn't reconciled. Ops teams see 'phantom' delivery gaps that are really just two auctions stitched badly.
— Separate Protected Audience line items in your delivery reports
— Reconcile component-auction wins against rendered impressions
Impact: split auctions break naive revenue reconciliation. Confidence: confirmed.
Confirmed: on-device Protected Audience (formerly FLEDGE) auctions resolve separately from the contextual auction, and GAM reports them in a way that can double-count or strand impressions if your setup isn't reconciled. Ops teams see 'phantom' delivery gaps that are really just two auctions stitched badly.
— Separate Protected Audience line items in your delivery reports
— Reconcile component-auction wins against rendered impressions
Impact: split auctions break naive revenue reconciliation. Confidence: confirmed.
SSP consolidation is collapsing your duplicate paths
Heard: when two SSPs you both work with merge, the acquirer often de-dupes overlapping supply post-close — meaning a publisher who relied on two paths to one DSP wakes up with one, and bid density drops without any setting changing on your end.
— Map which DSPs you reach through more than one SSP
— Flag the overlaps that share an acquirer for path resilience
Impact: M&A quietly thins your demand redundancy. Confidence: rumor.
Heard: when two SSPs you both work with merge, the acquirer often de-dupes overlapping supply post-close — meaning a publisher who relied on two paths to one DSP wakes up with one, and bid density drops without any setting changing on your end.
— Map which DSPs you reach through more than one SSP
— Flag the overlaps that share an acquirer for path resilience
Impact: M&A quietly thins your demand redundancy. Confidence: rumor.
One publisher ID, twelve seller_ids
Watching: a single SSP relationship can map to multiple seller_ids — one per pub property or per integration. Buyers de-duping at the seller_id level see twelve sellers where there's one entity, inflating your apparent footprint and triggering quality models that penalize sprawl.
— seller_id sprawl reads as low-quality long-tail supply
— Consolidation under owner_domain fixes the optics
Impact: map all your seller_ids to one declared owner_domain so audits see one clean entity. (sourced)
Watching: a single SSP relationship can map to multiple seller_ids — one per pub property or per integration. Buyers de-duping at the seller_id level see twelve sellers where there's one entity, inflating your apparent footprint and triggering quality models that penalize sprawl.
— seller_id sprawl reads as low-quality long-tail supply
— Consolidation under owner_domain fixes the optics
Impact: map all your seller_ids to one declared owner_domain so audits see one clean entity. (sourced)
Client-side vs server-side Prebid: the cookie tax
Heard: everyone moving s2s 'for speed' is quietly eating a 20-40% match-rate haircut from server-to-server cookie syncs.
— Client-side: best match rates, full SPO control, but 8-12 bidders maxes out browser CPU and tanks CLS on mobile.
— Server-side: scales to 30+ demand partners, lighter page, but you trust the s2s vendor's auction and lose per-bidder timeout granularity.
— Hybrid: top 4 revenue bidders client-side, long-tail s2s. That's where the real operators live.
Impact: Don't go full s2s on a cookie-dependent stack without measuring match-rate first. (sourced)
—
Если merchant feed shutdowns — твоя тема, посмотри @deal_wire
Heard: everyone moving s2s 'for speed' is quietly eating a 20-40% match-rate haircut from server-to-server cookie syncs.
— Client-side: best match rates, full SPO control, but 8-12 bidders maxes out browser CPU and tanks CLS on mobile.
— Server-side: scales to 30+ demand partners, lighter page, but you trust the s2s vendor's auction and lose per-bidder timeout granularity.
— Hybrid: top 4 revenue bidders client-side, long-tail s2s. That's where the real operators live.
Impact: Don't go full s2s on a cookie-dependent stack without measuring match-rate first. (sourced)
—
Если merchant feed shutdowns — твоя тема, посмотри @deal_wire
The owner_domain field nobody fills out
Watching: sellers.json has an optional
— If you resell through 3 SSPs, blank owner_domain hides the duplication from DSP de-dupe
— Buyers reading it as a quality signal, not just a curiosity
Impact: fill owner_domain before a buyer's path-dedup logic decides your seat is the redundant one. (sourced)
Watching: sellers.json has an optional
owner_domain field that ties an SSP entry to the legal entity that controls it. Almost nobody populates it. Buyers running supply-path audits now use it to collapse the same parent across multiple SSPs — when it's blank, your inventory looks like a fresh seller instead of a known reseller.— If you resell through 3 SSPs, blank owner_domain hides the duplication from DSP de-dupe
— Buyers reading it as a quality signal, not just a curiosity
Impact: fill owner_domain before a buyer's path-dedup logic decides your seat is the redundant one. (sourced)
DIRECT relationships that are quietly RESELLER
Heard: a recurring audit finding — publishers list an SSP as
— ads.txt says DIRECT, sellers.json says RESELLER = laundered path
— DSPs increasingly drop bids on the conflict rather than guess
Impact: reconcile your DIRECT/RESELLER tags across both files or watch eligible demand silently shrink. (sourced)
Heard: a recurring audit finding — publishers list an SSP as
DIRECT in ads.txt while that SSP's sellers.json marks the same seat as RESELLER. The mismatch is the tell. It usually means a managed-service team is reselling under the pub's name without the pub knowing.— ads.txt says DIRECT, sellers.json says RESELLER = laundered path
— DSPs increasingly drop bids on the conflict rather than guess
Impact: reconcile your DIRECT/RESELLER tags across both files or watch eligible demand silently shrink. (sourced)
