Photo by Vitaly Gariev on Unsplash
The Association of National Advertisers' 2023 in-housing study put adoption among major brands at 82%, up from 42% in 2008 (ANA, via Adweek, 2023). The headline reads as a generational shift away from agencies, but the same dataset shows that 92% of in-house teams still partner with external agencies, and 88% report increased workload that the in-house team alone is not absorbing (Adweek, 2023). The shift is layering, not substitution.
The economics drive the layering. Marketing budgets compressed to 7.7% of company revenue in 2024, down from 9.1% in 2023 and below the pre-pandemic 11% norm (Gartner CMO Spend Survey, 2024). Two mid-level performance hires at $130k base run $324k to $364k fully loaded, where a comparable agency engagement lands at $25k to $45k per month (MarketingProfs, 2024). The arithmetic does not make in-housing cheaper at all scopes; it makes some functions cheaper.
Two functions that consistently in-house well: programmatic media buying and data-driven optimisation. Bayer reportedly saved $10m in six weeks moving programmatic in-house; AT&T claimed $100m annually by doing the same (MarketingProfs, 2024). Sprint's 2017 case at Harvard Business School documented the same pattern: real-time bidding, data instrumentation, and audience operations belong close to first-party data, where vendor markup costs you the most (HBS, 2017).
Two functions that consistently outsource well: creative direction and senior strategy. Both depend on cross-account exposure (more brands seen, more pattern recognition) and senior talent retention is harder in-house. The right framework is not in-house versus agency but a per-function decision tree: data, real-time, tactical to in-house; cross-account creative or strategic to agency. The ANA study tracks brands running this framework already.