Overheard during the Digiday publisher town hall

"I don't think we've necessarily seen a material change in the cadence of budget flow ... There's definitely shifts and changes - tariffs impact auto, tariffs impact everything. So as macro changes happen, that definitely has an impact on individual marketing plans. But I don't think it's materially - year-over-year - vastly different from where it was before."

"We've seen people prepping for the [4th] quarter, and it's been... a good amount of success for [Q4]. I think with the ups and downs with AI, it's more so how can we pump it down and make sure that we can stay true to our commitments, and make sure we can deliver on all of them?"

"The open auction side of things - because that's directly proportional to the traffic that you receive - has been a bit tricky [to forecast or understand]. Unless you have a really, really high sell-through rate, the traffic drops definitely affect you."

"We lost about 50% of our traffic, stabilized, and have been there for a long time... My job [in rev ops] is, how do we extract more money per user? And that's where we've done a decent job. Have we recovered it all? No, but ... we've improved our revenue per user by 40% ... We're not back where we were. But that's the other way of trying to figure out how to offset that."

Ad budget planning shows relative stability with advertisers maintaining cadence despite macro shifts and tariffs affecting specific verticals. Many publishers report successful Q4 preparations while focusing on delivering commitments amid AI-driven volatility. Open-auction revenue has become harder to forecast because traffic declines directly reduce yield unless sell-through rates remain very high. Several publishers report dramatic audience losses — in one case about 50% — and respond by increasing revenue per user, achieving roughly 40% lift in RPM though overall revenue remains below prior levels. Executives prioritize optimizing and diversifying revenue, managing DSP/SSP relationships, and identifying AI use cases.

Read at Digiday