7 marketing lessons from Google’s Media Lab

As marketers, we are writing the future of our practice together in real time. It’s an uncomfortably exciting moment, and, at Google’s Media Lab — the team that oversees media strategy, planning, and buying for Google’s own marketing efforts — we wanted to share what we’ve learned so far. The impacts of AI in marketing across creative, production, media, and measurement are real and scaling fast. It’s not yet seamless, but a full end-to-end approach is coming into view.

Here’s how we’re leveling up our marketing strategy to continue driving gains across our marketing stack.

1. Let AI take the wheel

We used to plan media — selecting sites, times, placements, matching creative, and more. Now AI campaigns do most of the driving. Last year, over half of our on-stack media investment decisions were improved by AI. A marketer’s job isn’t to steer but to provide better fuel to feed our algorithm: stronger data and a richer, AI-ready suite of multimodal assets, especially short-form video. Across creative, production, media, and measurement, we are seeing real AI impact and gains.

2. A thousand flowers bloomed. It’s time to make a bouquet

Media experiments are constant, especially as AI-powered media becomes more capable. In 2025, we established a disciplined learning agenda with graduation criteria that resulted in triple the learnings from 2024.

We evaluated more than 95 hypotheses, selected 22 for rigorous testing, and generated nine scalable learnings over three quarters. The key is to define what success looks like at the start and then measure pilots, experiments, and tests so you can scale what works and shutter what doesn’t.

3. Demand more incrementality

Incrementality remains the gold standard of direct-response marketing, but AI tools have often felt like a double-edged sword. AI-driven feedback loops boost returns, yet the black-box nature of these systems often complicates measurement. Though imperfect, we’ve seen steady reporting improvements in response to advertiser feedback, such as Performance Max transparency reporting. We’ll continue pushing the need for more transparency to drive better incremental business outcomes.

4. Everything is video

Video is, first and foremost, a direct-response channel now. Your social feed is video. CTV is video. Even your podcasts are video. And now, Gen AI is accelerating video’s dominance. Plan your creative approach accordingly.

5. Creators should be part of your creative mix

Social boosting isn’t a silver bullet, but it has shown promising results when used to complement a regular media campaign. We call this the “two-voice model,” carefully picking creator voices to complement our own marketing voice. But if the creator’s content doesn’t stop the scroll organically, paying to promote that content won’t suddenly make users engage.

6. AI is forcing a ‘remerging’ of marketing

AI is forcing reappraisal of the separation of creative, media, and production. A “rebundling” is happening, and there are opportunities for the truly integrated AI agency to deliver exceptional value across every layer.

7. Change management is the toughest part of AI transformation

The toughest and slowest part of AI transformation, like every technological transformation previously, is the change management needed to put AI to work. To get the most out of any new technology, you need to change how people work and how organizations function. This is a challenge requiring leaders to play with the tech themselves, celebrate the learnings, not just the wins, and lead with optimism.