2025 Meta Performance Marketing Summit - 6 key learnings

1. AI remains central, but advertisers are getting more control back

We’re still in the era of AI-driven performance, but Meta is slightly shifting its stance. After years of aggressively pushing automation (think Advantage+), we’re entering a new phase where advertisers are being handed back more levers to guide performance toward business goals.

This new balance “AI at the core, human input at the edge” is a nuanced but important evolution. It opens the door for smarter strategy and deeper alignment with client objectives.

2. Creative volume and diversity still at the forefront 

Creative remains the most powerful lever for performance, and Meta continues to invest in tooling to support more volume and variety. 

This is paired with the push of Meta to further support their creator community and make creator content the core of their 

One small update: Partnership Ads can now feature just the creator’s handle, creating a more “native” look in-feed. This small but strategic tweak - combined with new experiments in testimonial formats and catalog integrations - makes Partnership Ads an even more compelling format. 

3. Incrementality presented as the ‘north star’ by Meta 

Incrementality has been the north star for most serious growth marketing organizations but it was still rarely put on the forefront by Google and Meta. This changed at this summit - it was mentioned slide after slide. For the first time, they’re positioning it as the foundation for aligning media with business outcomes, acknowledging that ROAS doesn’t tell the full story.

We’ve been bullish on incrementality for years (see our first article on this from 2019!), and it’s exciting to see Meta finally building products that support it, like incrementality attribution campaigns. This is a major validation of the direction we’ve been supporting and we’re glad that Meta is creating tools to enable easier measurement for it. 

4. Value Optimization continues to be a core focus for Meta’s ad product 

Meta continues to push value-based optimization - including profit-focused bidding strategies - and we have seen it drive great value for some clients (but not all). For businesses with many SKUs and no strong internal prioritization, the model can work very well. But in cases where certain categories or SKUs matter more than others, it’s not always the right fit because it can send the wrong signals to the algorithm. 

We’re continuing to refine our own POV on when and how to use value optimization. For now, it’s a promising lever - but one we apply selectively.

5. High-quality signal is more important than ever 

Meta continued to emphasize an ongoing message: signal quality is the engine behind AI performance. Especially for non-Shopify brands, poor signal quality is still the biggest limiter of campaign success.

Meta introduced a framework to help brands assess and improve their data signals - and their new performance scorecard includes signal quality as a measurable input. 

6. Meta’s new performance scorecard: helpful, with context

Meta introduced a new performance scorecard, complete with “opportunity scores” and recommendations. But while these can be helpful diagnostics - especially for tracking signal quality - they’re designed with the long tail of advertisers in mind.

For brands spending at scale, these tools need context. We’ll be integrating the parts that make sense (like signal health tracking) but keeping a critical eye on the rest.