What’s costing advertisers money on Meta

We’ve carefully analyzed thousands of ads and hundreds of Meta ad accounts, and we can now say this with confidence: most advertisers aren’t losing money because of “bad creatives,” but because they lack an effective overall system.

Meta’s new algorithm now rewards creative volume and variety. The more material you give it to test, the faster it can identify the right audiences on its own.

So the advertisers who win aren’t necessarily the ones with the best ideas, but the ones with the best system for producing, testing, and iterating on them.

Here are the five budget leaks we see across almost every advertiser, and how to plug them.

1. Manual competitor research

Your competitors have already paid to test hundreds of ads. Meta’s Ad Library gives you access to those tests for free. Yet most advertisers barely scratch the surface of what it has to offer.

The scenario is always the same. Someone opens the Ad Library, scrolls through the ads of one or two competitors, takes a few screenshots, and drops them into Slack or a shared folder. Three weeks later, nobody remembers where they are or why those ads were saved in the first place.

Here’s what this actually costs you:

• Missed signals. An ad that has been running for three months at a competitor with sufficient reach is likely profitable. Without tracking it over time, it’s impossible to distinguish a winner from a failed test.
• Incomplete research. You look at the two competitors you already know, rarely the dozens of adjacent players testing angles you could adapt to your own business.
• Wasted hours. Everyone repeats the same research from scratch because nothing is centralized or tagged.
• Creatives reinvented from scratch. Without a shared reference library, the team starts from a blank page instead of building on what’s already working in the market.

The solution: centralize your competitor research in a single tool. Save ads, categorize them by angle, format, hook, and competitor, and track how long they stay live. Your research then becomes a shared knowledge base that the entire team can consult before every creative brief.

2. A disorganized creative production process

Creating an ad often involves three or four people: the media buyer who defines the need, the creative strategist who writes the brief, the designer or video editor who produces the asset, and the manager who approves it...

A fragmented workflow. Briefs live in Google Docs, feedback gets buried in emails, files sit in Drive, and approvals happen on WhatsApp (yes, we’ve seen it). Days are wasted searching for the latest version or waiting for feedback. Meanwhile, live creatives start to lose their effectiveness, and cost per acquisition rises.

Unnecessary artistic perfectionism. On Meta, an ad gets judged in less than two seconds, in a feed, on a phone screen. Spending a full day tweaking a gradient won’t change the results. A video shot on a smartphone can regularly outperform a polished production because it feels more authentic. Time spent polishing is time taken away from creative volume.

Variations that are too superficial. Changing a button color or the background of a visual doesn’t make it a new ad. Under Meta’s Andromeda system, overly similar creatives are grouped together and treated as one. You think you’re testing ten ads, but the algorithm sees only one. A meaningful variation changes something significant: the hook, format, person on screen, angle, or promise.

The solution: build a single workflow where briefs, production, feedback, and approvals all happen in one place. Set a “good enough” standard instead of chasing perfection, and think of creative variations as fundamentally different approaches, not just cosmetic changes.

3. No structured testing pocess

Many advertisers claim to “test,” but few have a genuine testing process. They launch a batch of creatives, see which one performs best, turn off the others, and start over. The problem? At the end, they know which ad won, but not why.

Without a method, every win is a lucky break they can’t replicate. And every failure leaves them with no real learnings.

An effective testing process separates several levels:

1. Concept: the core idea behind the ad (customer testimonial, product demonstration, before-and-after comparison, behind-the-scenes brand content).
2. Angle: the reason to buy that the ad emphasizes (saving time, saving money, status, fear of missing out, solving a specific pain point).
3. Hook: the first two to three seconds that stop someone from scrolling.
4. Format: UGC video, static image, carousel, motion graphics.

Where relevant, test each of these by persona too (and it often is relevant).

Document every test with a clear hypothesis (“the time-saving angle converts better than the cost-saving angle among freelancers”), one isolated variable, and a recorded result. After a few months, you’ll have a brand-specific library of learnings that makes every new brief more effective.

4. A tedious ad publishing process

Once the creatives are ready, there’s still the task of getting them live. And that’s often where everything slows down.

The media buyer downloads files from Drive, renames them, opens Ads Manager, recreates the ad sets, pastes copy from another document, adds UTM parameters from a spreadsheet, and checks formats for each placement. Multiply that by thirty creatives and several ad accounts, and the entire day is gone.

Constantly switching between tools has three consequences:

• Costly mistakes. A broken link, a missing UTM parameter, or a creative published in the wrong ad set can distort your data or waste budget for days.
• A slower launch cadence. When publishing is a pain, you publish less often. Yet creative refresh frequency is precisely what helps limit ad fatigue.
• An underutilized media buyer. Instead of analyzing results and refining strategy, your media buyer spends their time on manual data entry.

The goal is to minimize back-and-forth between platforms. Strict naming conventions, reusable campaign templates, and a tool that publishes directly from the workspace where creatives are approved can turn ad launches from a matter of hours into a matter of minutes.

5. A fragmented approach to AI

Almost every team is already using AI. The copywriter generates hooks in a chatbot, the designer creates visuals with an image-generation tool, the video editor adds automatic subtitles in a third tool, and the media buyer analyzes exports in a fourth. Everyone works in their own silo.

This fragmented approach severely limits the gains:

• No shared context. Every tool starts from scratch. It knows nothing about your brand, your personas, the competitor ads you’ve analyzed, or the results of your latest tests.
• Generic output. Without that context, AI produces interchangeable copy and visuals that look like everyone else’s.
• Subscription overload. Multiple paid tools with overlapping use cases, and no clear view of their return on investment.
• Endless copy-pasting. The time saved by AI generation is lost again transferring outputs from one tool to another.

AI delivers real value when it’s connected across the entire workflow: it draws on your competitor research to suggest angles, generates genuinely distinct creative variations, incorporates learnings from your tests, and prepares ads for publication.

The solution: bring your AI workflows together around a single source of truth: your brand, your research, and your results. AI then stops being an individual productivity gimmick and becomes a force multiplier for the entire team.

Conclusion: the problem isn’t Meta. It’s your system.

These five budget leaks have one thing in common: none of them shows up in Ads Manager. They don’t appear as a line item in your spending, but they affect every result. Fewer creatives tested, slower launches, lost learnings, and teams burning out on low-value tasks.

Advertisers who succeed on Meta understand that performance is built upstream. They have centralized competitor research, a smooth production workflow, a rigorous testing methodology, fast publishing, and AI that works for the entire team.

The good news? Every one of these leaks can be fixed. Start with the one costing you the most time today, and you’ll quickly see the difference in your customer acquisition costs.

‍