Most 2027 marketing budget debates are stuck on tools, channels and AI efficiency. New payroll data suggests the real question is quieter: who will be around to run the function once today’s directors move on?
What the Stanford data shows
A working paper from Stanford’s Digital Economy Lab, revised Aug. 12, 2026, uses anonymized ADP payroll records from about 3.5 to 5 million employees a month through June 2026. The authors find that employment for workers aged 22 to 25 in AI-exposed occupations is 19% below where it would be if it had kept pace with peers in less-exposed roles.
The raw shift is stark. Since November 2022, the two most AI-exposed quintiles among young workers fell roughly 11%, while the three least-exposed quintiles grew roughly 10%. Older workers show no comparable gap; every older group grew, with 35-to-40-year-olds up as much as 11%.
The mechanism is the part marketers should highlight. The decline runs through reduced hiring, not layoffs. Separations among young workers fell at least as much in exposed occupations as in less-exposed ones. That points to entry-level seats not being filled, rather than teams being cut.
The paper is careful about its limits. The authors call the findings “early, descriptive indicators,” not a causal estimate. Controlling for education shrinks the effect and weakens its significance. The research also covers occupations broadly and includes no marketing or SEO-specific figures.
Why this matters for marketing budgets
The column argues that many CMOs are funding experienced editorial and strategy roles while quietly leaving junior seats empty. Gartner data has shown labor rising as a share of marketing budgets, from a mean 21.9% to 24.5%, but it does not split that spending by seniority. That makes it easy to say the team is “intact” even when the entry-level pipeline is not.
An intact team tells you who is still there. It says nothing about who was never hired. If fewer juniors enter, the next generation of managers has fewer people learning the judgment work that AI still cannot own.
Put the role in a learning bucket
Kevin Indig’s 2027 planning memo offers a useful frame. Proven work like technical SEO and digital PR gets full funding, while uncertain bets get a capped learning bucket with one owner and a decision date. Community work, which is hard to hand to a model, deserves a named owner and fixed weekly hours.
For 2027, the column’s advice translates into a simple planning sequence:
- Sort every junior task into “AI can substitute” and “needs judgment and relationships.”
- Write the junior role as a learning bet, with fixed hours for community, citation tracking and off-site mentions, plus a decision date.
- Add a line to the scorecard showing seats you chose not to fill, next to hours saved and outcomes delivered.
- Track exposure, behavior and business outcomes monthly, and read them together.
The metric to watch
Forget “intact.” The better test is whether you are still training. If AI pays for itself, some of the savings should show up as a junior seat with real work attached. Otherwise, the team looks healthy on today’s headcount report and hollow when the next director search begins.
Source: Search Engine Journal
