Why Media Planning Still Needs a Human Strategist

Automation has changed how media gets bought. Algorithms can evaluate millions of impressions in the time it takes a planner to open a spreadsheet. But speed isn’t strategy, and the industry’s own research shows what happens when the two get confused.

The cost of unsupervised buying

The Association of National Advertisers (ANA) audited programmatic campaigns with the help of participating brands. Only 36% of the ad spend studied went to impressions that were brand safe, outside made-for-advertising (MFA) sites, measurable, and viewable. MFA publishers took 15% of programmatic spend and accounted for 21% of tracked impressions.

The supply chain’s scale compounds the problem. Study participants’ ads ran on an average of 44,000 websites and apps, while most people visit only 10 to 15 sites on a typical day. A buying system that optimizes only for cheap reach will drift toward whatever inventory is plentiful, not whatever reaches the right people. MFA supply grew from roughly 5% of web auctions in early 2020 to nearly 30% by mid-2023.

The ANA study reflects 2023 data, and conditions have improved since. Jounce Media found that MFA supply as a share of web bid requests fell by half between June 2023 and March 2024. That improvement came when advertisers and their agencies started asking harder questions about where their money was going.

Invalid traffic remains a significant drain

Estimates of fraud and invalid traffic (IVT) vary by methodology, and many come from vendors who sell detection tools, so the figures are best read as a range:

• Lunio puts the average invalid rate at 8.51% of global paid ad traffic, or about $63 billion in annual wasted spend.
• Fraudlogix analyzed 105.7 billion impressions from 2025 and found a global IVT rate of 20.64%.
• Spider Labs estimated advertisers lost more than $32.6 billion to ad fraud in 2025, and found fraud risk up to 2x higher in AI-optimized campaigns than in average campaigns.

The methods differ, but the direction is consistent: a meaningful share of spend never reaches a real, relevant person. Automated systems optimize toward the signals they are fed, and bad traffic corrupts those signals. In a Lunio survey of senior marketers, 16% cited junk traffic feeding misleading signals into automated campaign types, and 14.5% cited skewed ROAS and CPA metrics.

The oversight gap

Marketers know the risk exists. More than three-quarters of senior marketers surveyed estimated that invalid traffic takes over 5% of their monthly performance budget, yet only 5.3% use a dedicated tool to stop it. Awareness is high and action is low. Closing that gap takes someone accountable for the plan’s outcome.

Placement is a brand decision

Where an ad appears shapes how people judge the brand behind it. In a 2019 survey by the Trustworthy Accountability Group and the Brand Safety Institute, 90% of respondents said it was very or somewhat important for advertisers to keep their ads away from dangerous, offensive or inappropriate content. When shown a hypothetical ad for a product they regularly buy next to Neo-Nazi propaganda, 87% said they would reduce their spending on it and 58% said they would stop buying it altogether.

Consumers also read placement as intentional. In separate research reported by Marketing Dive, consumers who saw ads beside negative content were 4.5 times more likely to think the brand didn’t care about them, and 50% less likely to recommend it. Those are reputational costs, and no click-through rate captures them.

What a human planner contributes

• Context. A planner can tell when a cheap, high-performing placement conflicts with a brand’s values or a current news cycle.
• Audience understanding. A planner thinks about the person behind the segment, including what they care about and where they pay attention, not only the targeting category.
• Supply path judgment. Legal advisers who helped shape the ANA study recommend considering inclusion lists rather than exclusion lists, since exclusion lists can’t keep pace with the number of MFA sites. Curating a list of trusted placements is a human, editorial decision.
• Accountability. When something goes wrong, a person can explain why and fix it.

The right model: human-led, machine-assisted

None of this argues against technology. Automation is invaluable for reporting, bid management and pattern detection. The strongest teams assign it those jobs and keep strategy, audience definition and brand-fit decisions with people who understand the business.

The question for any media plan is simple: does a qualified person own the decisions about who we reach, where we appear, and why? If the answer is no, the budget is working for the algorithm’s objectives, not the brand’s.