Most marketing teams treat Meta ads like a slot machine. Pull the lever, watch the dashboard, hope for a payout. When performance dips, they audit creatives, debate lookalike percentages, and tweak attribution windows. The real problem sits upstream: they're optimizing for platform metrics instead of revenue outcomes.
The gap between what Meta reports and what actually hits the P&L has widened. And if you're presenting ROAS to your board without a clear line to incremental revenue, you're building on sand.
The Incrementality Question Your CFO Is Already Asking
Here's the number that should anchor every Meta conversation: across 640 incrementality tests, Meta campaigns delivered an average 19% lift to primary KPIs like total revenue or new customer acquisition. Remove Meta from your media plan and you'd reduce these outcomes by nearly one-fifth.
That's a meaningful number. It's also an average, which means your specific lift could be 8% or 34%. The only way to know is to test, and most teams don't because incrementality testing feels expensive and complicated. It's neither, but it does require you to stop treating platform-reported ROAS as ground truth.
For omnichannel brands, the math gets more interesting. About 32% of Meta's attributed impact happened outside direct-to-consumer channels, driving in-store or Amazon sales that wouldn't have occurred otherwise. If you're only measuring DTC conversions, you're systematically undervaluing the channel and misallocating budget.
Signal Quality Before Creative Testing
The instinct when performance drops is to launch more creative variants. That's usually the wrong move.
In 2026, Facebook relies heavily on data signals to power ad delivery. When purchase data is missing, duplicated, or delayed, the algorithm cannot identify the right buyers. This leads to learning phase resets, unstable delivery, and unpredictable return on ad spend.
Your optimization foundation should ensure that purchase events fire reliably, order value and currency are accurate, events are not duplicated, and server-side tracking supports browser events. Clean signals allow Facebook to understand who converts and why. Once this foundation is in place, every other optimization becomes more effective.
I've seen teams spend six figures on creative production while their pixel fires inconsistently. That's not a strategy problem; it's a plumbing problem. Fix the plumbing first.
The Advantage+ Trade-Off
Advantage+ campaigns were adopted by 93% of brands in one test group and account for about 39% of Meta ad spend. The appeal is obvious: less manual work, faster optimization, broader reach.
The trade-off is less obvious. Advantage+ excels at finding high-intent users and delivering immediate impact. It's efficient at harvesting demand that already exists. What it doesn't do well is create new demand or reach audiences who aren't already in-market.
This matters for your growth model. If you're running Advantage+ exclusively, you're likely over-indexing on bottom-funnel conversions while starving the top of your pipeline. The dashboard looks great until you notice that new customer acquisition is flat and you're paying more to convert the same shrinking pool of high-intent buyers.
Beyond Purchase Conversion Campaigns
One operator's experience captures a shift worth understanding: "When we started using non-sales campaigns to drive revenue, people thought it was stupid. Even me from 2 years earlier would have thought it was stupid."
The insight is that purchase conversion campaigns optimize for a narrow definition of a buyer: someone who clicks and buys in a 1-day window. That's a small slice of your actual market. When you ask the platform to look for different intents, you're still hitting people who are in-market but whom you weren't reaching with purchase-optimized campaigns.
At the highest level, brand growth requires two things: reaching more net new people, and getting your most memorable content in front of them so they remember you when they come in-market. Every campaign type and optimization objective is just a different tool to achieve those two things.
The huge opportunity for performance marketing-led brands who know they need to invest in brand is that the ad auction has massively bid up the cost of purchase-optimized inventory. Non-purchase objectives often reach similar audiences at lower CPMs.
The Omnichannel Measurement Gap
If your business sells online and in-store, optimizing Meta campaigns for website purchases only hides a big part of your real impact. Meta Omnichannel Ads help optimize for website and in-store outcomes in one Sales campaign.
To make this work, advertisers need a reliable, ongoing feed of offline events, typically implemented via Conversions API for offline events. The setup isn't trivial, but the alternative is making budget decisions with half the data.
Fresh Clean Threads ran a 2-cell lift test with Amazon orders as the primary KPI. They found Meta was responsible for a 23% lift in Amazon sales while simultaneously driving a 21% lift to DTC sales. Without that test, they would have attributed zero Amazon revenue to Meta and likely cut spend that was actually profitable.
A Two-Week Pilot Plan
If you're presenting Meta performance to your board next quarter, here's what I'd prioritize in the next two weeks.
First, audit your signal quality. Check that purchase events fire reliably, values are accurate, and server-side tracking is active. This is a one-day technical review that often reveals significant issues.
Second, run a holdout test. Even a simple geo-based holdout for two weeks will give you directional incrementality data. The goal isn't statistical perfection; it's getting a number that's closer to truth than platform-reported ROAS.
Third, map your campaign mix against the funnel. What percentage of spend is Advantage+ versus manual? What percentage is purchase-optimized versus awareness or engagement? If you're 90%+ on bottom-funnel, you're likely harvesting demand faster than you're creating it.
The risk of doing nothing is that you continue optimizing for metrics that don't correlate with revenue. The risk of this approach is that you discover your actual incrementality is lower than reported, which means you've been over-investing. Both are better than not knowing.
Your CFO will eventually ask the incrementality question. Better to have an answer before they do.
