Meta isn’t broken, but a lot of Meta ad accounts are. Teams keep applying tactics that used to work without realizing how much the platform has changed. Where advertisers once won through precise audience targeting, Meta has shifted its system to embrace machine learning.

In many cases, accounts struggle not because the product is bad or the channel is dead, but because the account isn’t set up for how Meta now learns and optimizes. Below, we’ll break down three setup issues we see holding many Meta ad accounts back—and what to do instead.

Looking for support with your Meta ads? Primer can audit your account’s tracking, campaign structure, and creative system and recommend a path forward.

How advertising on Facebook and Instagram used to work—and what changed

In the 2010s, Facebook and Instagram were among the most efficient digital acquisition channels. Many brands found success through:

  • Narrow interest targeting
  • Lookalike audiences
  • Reliable attribution via the Facebook Pixel

Costs were typically lower, competition in the auction was lighter, and even relatively simple creative could generate strong returns across nearly every industry. Over the last several years, however, changes in privacy, user behavior, and AI have reshaped how Meta advertising works.

Data and privacy regulations

While the concept of data privacy isn’t new in the grand scheme of things, it’s gained enormous attention over the last decade as platforms and apps collected more user data. 

Most notably, the Cambridge Analytica scandal—in which a political consulting firm harvested the personal data of millions of Facebook users without their consent—brought Facebook under intense scrutiny and accelerated tighter controls on data access.

Then Apple’s iOS 14.5 update introduced App Tracking Transparency in 2021. The new privacy feature required apps to ask for permission to track users across other companies’ apps and websites—effectively reducing the amount of observable conversion data available to ad platforms.

These developments led to a shift toward server-side tracking to bypass ad blockers and privacy restrictions. For Meta, that meant the introduction of the Conversions API (CAPI) (originally known as “Server-Side Events”) to let advertisers share web, app, and offline events directly from their servers. 

AI and machine learning advancements

Even before the changing tides around data and privacy, Meta began investing in machine learning. In 2013, it launched the Facebook AI Research Lab (FAIR), declaring its goal to “understand the nature of intelligence so that we might create intelligent machines.”

Since the creation of FAIR, Meta has made good on this promise by applying its research to ads. In 2025, CEO Mark Zuckerberg shared his vision for a new world of advertising, where AI does everything:

“In general, we’re going to get to a point where you’re a business, you come to us, you tell us what your objective is, you connect to your bank account, you don’t need any creative, you don’t need any targeting demographic, you don’t need any measurement, except to be able to read the results that we spit out. I think that’s going to be huge, I think it is a redefinition of the category of advertising.”

By automating decisions about audiences, placements, and delivery, Meta’s Andromeda and Advantage+ are a step in this direction. No longer about manually finding the “perfect” audience, these developments emphasize giving Meta’s algorithm the right inputs so it can learn what works and scale it.

Changes in user behavior and attention

Finally, Meta users’ behavior has changed, especially as Facebook no longer dominates the social media space the same way it did 10 years ago. 

To be clear, Facebook remains the world’s largest social media platform with more than 3 billion active users per month. But users’ attention is now fragmented across other platforms like TikTok, YouTube Shorts, and Snapchat. 

It’s also worth noting that Facebook users have evolved over the years, from a youthful demographic in the mid-2000s to a more mature audience. And though the platform once primarily centered around social networking, it’s now used for other purposes, like reading news and shopping. 

These changes have pushed Meta toward a more recommendation-driven platform—and its advertising tools have evolved in the same direction.

The 3 ingredients of a successful Meta ad account

So how can you set up your Meta ad account for success in this day and age? From our experience, there are three crucial elements:

  1. A strong data foundation
  2. A campaign structure that works with Advantage+
  3. Differentiated creative

Let’s break each down.

1. Strong data foundation

Many advertisers assume that setting up the Meta Pixel is enough. But because of iOS privacy, browsers, and ad blockers, relying on the Pixel alone can mean missing a substantial amount of events—some estimate as much as 30-40% of data. As a result, brands only using the Meta Pixel end up judging their campaigns on incomplete data.

Meta’s algorithm can only optimize for what it can accurately observe. And the more signals you can send back to Meta, the better your performance will be. If events are duplicated, missing, or miscategorized, Meta optimizes toward the wrong people and your reporting becomes unreliable.

A strong data foundation for today’s Meta ads accounts includes:

  • The Meta Pixel and events are installed and configured correctly, including deduplication so you aren’t double-counting conversions.
  • CAPI is enabled so Meta receives more accurate conversion signals, including offline conversions.
  • For B2B companies and lead generation campaigns: You send lead stage updates back (lead → qualified lead/opportunity → sale) so Meta learns what a high-quality lead looks like—not just who fills out a form.

With clean signals in place, your Meta ads account has a better foundation for learning. This makes downstream optimization decisions easier.

2. Campaign structure that maximizes Advantage+

When it comes to campaign structure, many accounts over constrain Meta with tightly defined audience segments. In other words, they try to force outcomes with narrow targeting, exclusions, and micromanaged setups.

However, achieving success with Meta ads now requires letting go of manual control and leaning into automation, particularly Advantage+ . From audience selection to budget allocation, Advantage+ can automate almost every aspect of your campaigns. The system’s machine learning algorithms study audience behavior in real time to predict which users will convert, and then automatically allocates budget/placements/targeting based on what it’s learning.

Here’s what we recommend:

  • Enable Advantage+ as much as possible across:
    • Budget (where applicable)
    • Targeting
    • Creative enhancements
  • Avoid over-segmenting with layers of interests and behaviors unless you have a clear, evidence-based reason.

Of course, there may be exceptions. For example, if you know you only want to reach parents or retirees, it makes sense to put guardrails like parent status or age. But for many offers, broad tends to be better. 

For truly specific B2B targeting, consider importing third-party audiences (e.g., job title/industry data not available natively).

With this type of setup, you generally make fewer manual adjustments over time. It’s a shift from the old Facebook playbook—one that requires getting comfortable with “letting go of the wheel” so the algorithm can do its job.

3. Differentiated creative that does the targeting

Creative is more important than ever on Meta—which makes creative testing a requirement, not a nice-to-have. Many struggling accounts try to scale ads that never proved they can carry performance. If the creative doesn’t stop the scroll or clearly signal who it’s for, increasing the budget often just buys more expensive data.

Creative testing doesn’t mean tiny tweaks, though. Swapping a headline or changing a background color rarely produces insights you can scale. What works better is testing meaningfully different creative, such as:

  • A variety of formats, from UGC to product demos to testimonials
  • New hooks and angles
  • Creators or spokespeople
  • Storylines designed for specific personas

This matters even more as Meta leans into new AI systems. In the past, the platform’s ads retrieval systems were only able to apply limited personalization. But now, its machine learning-powered systems like Andromeda and Generative Ads Model (GEM) enable hyper-personalization to deliver the right message to the right person at the right time.

Here’s what effective creative testing looks like:

  • Test distinct concepts with clearly different hooks/POVs.
  • Build distinct personas (e.g., recent grads vs. retirees vs. new parents), while keeping Advantage+ targeting on.
  • Let the creative define the audience, and let Meta find the people who respond to each persona.
  • Track your learnings and refresh creative to prevent ad fatigue.

For example, you might find influencer-led creative works well for recent grads and retirees, but less well for new parents. If that’s what the data shows, you can double down by:

  • Bringing in new creators who match the winning personas
  • Updating briefs based on what performed
  • Promoting the top-performing ads while you iterate on the next batch

The new Meta standard

The old Facebook playbook of tight targeting plus small creative tweaks no longer works. To succeed on Meta today, your account needs clean data, an automation-friendly structure, and creative that does the targeting.

These three pieces reinforce each other. Strong creative can’t overcome broken measurement, and clean tracking won’t matter if your structure prevents the algorithm from learning.

Want to know whether your account is set up to succeed—or stuck in an outdated approach? Primer can audit your tracking, structure, and creative system. Book a growth call to assess your Meta foundation and build a creative testing engine that scales.