Are Publishers Getting the Most from Their Programmatic Advertising?
To get the most from programmatic, publishers need to stop focusing on adding tools and start focusing on how those tools are governed. Vijay Kumar explains more...
For many publishers, the past year has made one reality unavoidable: traffic growth can no longer be assumed. AI-driven discovery, fewer direct referrals, and changing consumption habits have made audience volume more volatile and, in some cases, structurally constrained.
That shift has elevated programmatic advertising from a background monetisation channel to a core strategic lever. When audience growth is uncertain, the ability to extract more value from existing traffic matters more than ever.
The challenge isn’t access to technology or demand. Most publishers already have sophisticated stacks. The question is whether those systems are actually aligned with how the market operates today and whether they’re capturing the full value of the audiences they serve.
An overview of programmatic in news publishing
Across the industry, most news publishers run broadly similar programmatic architectures:
● Google Ad Manager as the central control layer for ad serving, pricing rules, and reporting
● Header bidding to introduce non-Google demand into each impression decision
● A portfolio of SSP (Supply-Side Platform) partners, typically with a small group driving the majority of revenue and a longer tail enabled for coverage
● Pricing rules and floors designed to protect yield and manage volatility
● Refresh logic and layout decisions that influence session value and total inventory
● Brand safety, privacy, and compliance controls that increasingly shape buyer participation
From an infrastructure standpoint, this looks mature.
Where publishers diverge is not in the tools they use, but in how those tools are governed. Many strategies reflect decisions made under very different market conditions: when traffic growth masked inefficiencies and pricing precision mattered less. Over time, these decisions accumulate. The stack continues to function, but strategy often lags behind buyer behaviour.
Meanwhile, buyers continuously adapt. DSPs (Demand-Side Platforms) reprice impressions in real time based on competition, user context, pacing, and performance goals. When buyer decisioning evolves faster than seller strategy, value doesn’t vanish but slowly shifts away from the seller.
Key performance indicators: looking beyond surface metrics
At the executive level, the issue is rarely a lack of data. It’s a lack of signal.
Metrics like total revenue or average CPMs are essential for reporting, but they don’t explain why outcomes change. They don’t show whether auctions are competitive, whether demand is being excluded unintentionally, or whether impressions are consistently under-clearing.
More informative indicators tend to sit closer to auction behaviour:
● RPM (page and session): a direct measure of how effectively audience attention is being monetised
● Fill rate: highlights demand coverage and where inventory is effectively going unsold
● Win rate by channel: helps identify whether outcomes are driven by real competition or default behaviour
● Bid density and bid distribution: reveals how competitive auctions actually are and whether revenue is overly concentrated among a small set of buyers
These metrics don’t require executives to manage tactics. But they do enable better questions: Are we seeing healthy competition? Are some buyers winning disproportionately cheaply? Are changes improving auction quality or just redistributing revenue?
Without this lens, stability is often mistaken for optimisation.
A practical 30-day testing framework (Q1, lower risk)
Optimisation doesn’t require constant change. It requires clarity.
A disciplined 30-day framework allows publishers to identify where value is leaking and which changes produce durable impact, without introducing unnecessary risk.
A publisher-friendly way to structure this work:
Week 1: Baseline and controls: Define success metrics (RPM, fill, win rate, bid density). Establish stable control groups. Document where revenue is concentrated by partner, device, geography, and content type.
Week 2: Auction health review: Identify where auctions are thin versus competitive. Flag impressions that clear cheaply despite strong competition signals. Highlight segments with inconsistent fill or outcomes.
Week 3: Limited-scope experiments: Run small, reversible tests designed to influence auction behaviour rather than force outcomes. Keep scope tight. One change at a time. Clear hypotheses.
Week 4: Readout and decisions: Determine what moves into production, what gets rolled back, and what requires further testing. Document learnings so they compound rather than reset each quarter.
Q1 is often an ideal window for this work. Traffic patterns are steadier, budgets are resetting, and the cost of learning is typically lower than during peak monetisation periods.
Real-world data: what tends to work
Across anonymised data from publishers in news, sports, and consumer verticals, several consistent patterns emerge.
First, the most durable revenue gains rarely come from simply pushing prices higher. They more often come from improving auction competition and reducing situations where impressions clear below plausible value despite strong demand being available.
Second, “silent leakage” is common. This occurs when pricing and demand rules don’t reflect real-time market conditions, allowing certain buyers to win cheaply by default rather than through competition.
Finally, strategies aligned with buyer behaviour tend to hold up better over time. Buyers adapt quickly to static constraints. Systems that more accurately reflect real-time value tend to produce steadier, more sustainable outcomes.
In most cases, long-term lift comes from better alignment, not tighter control.
Executive takeaway for 2026
Programmatic advertising is no longer just an operational concern. It is a strategic one.
As traffic growth becomes less predictable, the ability to maximise audience value becomes a defining advantage. The publishers best positioned for 2026 won’t necessarily be those with the most tools or partners, but those with the clearest understanding of how their auctions actually function and the discipline to ensure strategy keeps pace with the market.
In a constrained growth environment, that clarity matters more than ever.
About: Mile is an AI-powered programmatic revenue optimisation platform for publishers, headquartered in New York with offices in London and Bengaluru. Mile maximises yield from every impression by refining floors, shaping traffic, and enriching bid requests using publishers’ unique auction data. The platform is trusted by premium publishers including Times Higher Education, Graham Media Group, Salon, and Sportsnaut.
Regarding the topic of programmatic, this piece really hits different. It's so on point about how publishers need to rethink their stack usage. I remember your previous take on traffic volatility, and this is a great follow-up. The AI angle totally resonated, obviously!