Ad Targeting Is Broken—Here’s How to Fix It

User acquisition is getting more expensive.

CPMs are rising. Competition is fierce. Most brands are still overpaying for low-quality users.

For years, digital advertising promised precision targeting:

• The right person

• The right time

• The right message

But here’s the reality: Ad targeting isn’t as precise as you think.

Platforms push AI-driven bidding and hyper-targeted audiences, but many of these tactics are inefficient, expensive, and often misleading. Marketers keep optimizing for the wrong signals, leading to wasted spend.

If you want to refine your UA strategy, optimize your media mix, and scale profitably, it’s time for a new approach.

The Illusion of Precision in Ad Targeting

Marketers love to believe that digital advertising is more sophisticated than traditional media. After all, billboards can’t optimize in real time. TV spots can’t bid on high-intent users.

But most digital ad targeting is just as inefficient—if not worse.

The industry’s obsession with granular targeting (“People who searched for a mountain bike last week, are 25-34, live in LA, and recently bought workout gear”) creates false confidence.

Why Hyper-Targeting Fails

• Many “in-market” users aren’t actually in-market. They may have already made a purchase or weren’t serious buyers.

• Audience data is often outdated, misclassified, or missing critical signals. Platforms rely on fragmented tracking.

• The cost of targeting premium audiences often outweighs the benefit. Doubling your CPM to reach a refined audience does not always translate to better conversions.

The irony is that the more complex targeting becomes, the less effective it often is.

Why Broad Targeting Beats Granular Audiences

Platforms like Meta and Google promote audience precision. They offer hyper-specific targeting options and AI-driven smart bidding.

But in many cases, these tools are designed to maximize platform revenue—not your performance.

Key Advantages of Broad Targeting

• Lower audience costs – Many brands pay two to three times more in audience costs for every dollar of actual media spend just to target premium users.

• Wider reach, same performance – AI finds in-market users within broad pools, often without the need for hyper-targeting.

• More efficient machine learning – Platforms optimize better with more data. Broad targeting allows their algorithms to work properly.

A Simple Test to Compare Performance

1. Run an ad set targeting a hyper-specific audience.

2. Run another with only age, gender, and location targeting.

You will often see little to no difference in CPA, but the broad audience will scale more efficiently at lower CPMs.

The Real Problem With AI-Driven Targeting

Platforms push conversion-based targeting, where AI finds high-value users based on conversion signals. Sounds smart in theory.

But if the data you’re feeding AI is flawed, the entire optimization collapses.

Why AI-Driven Targeting Often Fails

• Bad signals in, bad results out – Many conversions come from bot traffic, misclicks, or low-intent users.

• AI may prioritize cheap clicks over real buyers – Platforms optimize for volume, not value. If your signal is weak, expect weak results.

• Marketers lose control over where their ads appear – AI-driven bidding often places ads on low-quality apps and websites to maximize attributed conversions.

Ever wonder why your ads show up on obscure mobile apps? That’s AI-driven bidding at work—maximizing conversions without considering brand safety or real audience quality.

Fixing the Root Issue: Feeding AI the Right Data

AI-driven optimization isn’t the problem—it’s the data you give it.

Most advertisers rely on platform-reported conversions (Google, Meta, TikTok) without validating them against a source of truth like an MMP or first-party data. This is where things go wrong.

How to Fix It

1. Send real, high-quality signals to ad platforms

• Use server-side MMP data (AppsFlyer, Adjust, Singular) to send only real, authenticated conversions.

• Avoid relying on platform-reported conversions without external validation.

2. Pass back post-install engagement, not just installs

• Feed ad platforms data on retention, repeat purchases, and LTV to improve AI decision-making.

• If you only optimize for installs, AI may prioritize low-quality users who never engage.

3. Monitor platform-reported ROAS vs. actual revenue

• Compare platform-attributed conversions to your backend revenue data.

• If there is a mismatch, AI is likely optimizing for the wrong audience.

4. Reduce reliance on platform black-box algorithms

• Use incrementality testing and Marketing Mix Modeling (MMM) to validate actual lift.

• Don’t assume high ROAS equals real business impact.

How to Fix Your Targeting Strategy

If hyper-targeting doesn’t work and AI-driven bidding is unreliable, what should marketers do?

1. Simplify Audience Targeting

• Test broad audiences against hyper-targeted segments.

• Measure incrementality, not just attributed ROAS.

2. Use Inclusion Lists, Not Exclusion Lists

• In programmatic, focus on high-quality publishers instead of blocking bad ones.

• Avoid platforms that optimize for cheap, low-quality traffic.

3. Optimize Creative Over Targeting

• Good ads outperform bad targeting.

• If your creative speaks to the right audience, they will self-select.

4. Rethink Mid-Funnel Targeting

• Many “in-market” users are actually post-purchase and no longer relevant.

• Stop optimizing for low-value signals like clicks and page views.

Try This Now

• Run a broad vs. targeted campaign test this week—see if your CPA changes.

• Check your MMP vs. platform-reported conversions—are they matching?

• Audit your site and app placements—are you paying for traffic from low-quality sources?

The Future of Ad Targeting

Digital marketing is not about hitting the bullseye every time.

It is about giving platforms the right data and creative so they can optimize effectively.

Where Future UA Performance Will Come From

The biggest wins in ad performance today do not come from better targeting. They come from:

• Stronger creatives that make users self-select.

• Smarter media mix modeling to measure real impact.

• Less reliance on platform algorithms that prioritize their own profits.

The future of UA is not about more precise targeting.

It is about smarter, broader, and more efficient advertising.

Key Takeaways

• Stop overpaying for over-segmented audiences.

• Test broad vs. targeted campaigns and measure incrementality.

• Let creative do the work of filtering the right users.

• Use AI cautiously—only trust it if the input data is accurate.

• Ensure platforms receive clean, validated signals from an MMP or first-party data.

Ad targeting is not broken.

Marketers just need to stop overcomplicating it and start feeding platforms better data.

Want weekly insights to refine your UA strategy, optimize your media mix, and scale profitably?

Sign up for the newsletter.