Stop Overthinking Your Meta Targeting. Let Your Creative Do the Work.

Here is something counterintuitive that practitioners are seeing across Meta accounts right now: as the algorithm has gotten smarter, more specificity in targeting often leads to worse performance, not better.

If you have spent the last few years building layered interest stacks, testing lookalike percentages, and fine-tuning demographic overlaps, that statement probably feels wrong. All of that work is supposed to improve performance. But account after account, across verticals, the pattern is the same. The advertisers going broader are outperforming the ones going narrower.

This is not a fluke. It is the new reality of how Meta’s ad platform works. And it changes what you should be spending your time on.

---

The Old Playbook Is Working Against You

The old Meta targeting playbook was straightforward: find the perfect interest stack, layer demographics on top, narrow the audience until you were only reaching people who looked exactly like your best customers. Build lookalikes from purchasers, test 1% vs. 3% vs. 5%. Exclude anyone who had already bought. Exclusions nested three layers deep.

This worked when Meta’s algorithm was less sophisticated. You were essentially doing the machine’s job for it - hand-selecting who should see your ads because the system could not figure it out on its own.

That era is over. Meta’s Advantage+ and broad targeting options now outperform hand-picked audiences in most cases because the algorithm has more room to optimize. When you constrain it to a narrow interest stack, you are not focusing it. You are limiting it. You are telling a system trained on billions of conversion events across millions of advertisers to ignore most of what it knows and only consider a small subset of users.

Interest targeting does not focus the algorithm. It hobbles it.

---

The Three-Layer Framework That Actually Works

If you are not supposed to obsess over interest targeting, what should your account structure look like? Here is the framework we use across accounts.

1. Prospecting: Go broad with demographic targeting only

Age range, gender, geo - that is it. No interest layers. No lookalikes. Let Meta’s algorithm find the buyers. The more you constrain it, the more you limit its ability to optimize delivery and find high-intent users at the lowest cost.

This feels reckless if you are used to the old playbook. It is not. You are not removing intelligence from the system - you are letting the system use its own intelligence instead of overriding it with yours.

2. Retargeting: Use custom audiences

Website visitors, email lists, engaged social followers. This is where precision matters - you are reaching people who already know you. Build your retargeting layers around first-party data: site visitors (segmented by recency and depth), add-to-cart abandoners, email subscribers, and people who have engaged with your content.

This is not interest targeting. This is behavioral targeting with data you own. Keep it.

3. Let your creative qualify your audience

This is the key insight, and it is the one most advertisers miss. Instead of using targeting settings to find the right people, use creative that would only resonate with people who are interested in your product or service.

If you sell premium skincare, show premium skincare. The wrong audience will scroll past. The right audience will stop. Your creative is your targeting.

---

Why Broad Targeting Wins Now

The mechanics here are simple. Meta’s algorithm optimizes for the objective you set. If you set purchase conversions as the objective and give it broad reach plus strong creative, it learns who converts and finds more people like them - faster than you could with manual interest targeting.

Every purchase event, every add-to-cart, every engagement signal feeds back into the delivery model. The algorithm builds a conversion profile that is far more nuanced than any interest category you could select. It is looking at behavioral patterns you do not have access to: scroll velocity, content interaction sequences, cross-platform signals, time-of-day patterns, device usage habits. Your “women 25–34 interested in yoga” audience is a crude approximation of what the algorithm already knows.

There is also an auction dynamic at play. When you constrain your audience to a narrow interest segment, you are competing in a smaller, more expensive auction alongside every other advertiser who selected the same interests. Go broad and the algorithm can find the same buyers at lower cost, because it is pulling from a pool where those buyers are not being fought over by every competitor running the same targeting playbook.

---

Your Creative Is Your Targeting

If broad targeting is the structure, creative strategy is the substance. This is where your time and energy should go - not into audience settings.

The principle is straightforward: your ad creative should pre-qualify your audience. If your ad resonates with everyone, it converts no one. Specificity in your creative replaces specificity in your targeting settings.

Think about what a tight interest audience used to do. It pre-filtered your audience to people who were plausibly in your market. Broad targeting starts with everyone, which means your creative has to do that filtering work. The best broad-targeting creative leads with a sharp, specific message that naturally self-selects. On TikTok, this principle becomes even more extreme - the algorithm ignores targeting parameters entirely and distributes based purely on engagement signals, making creative the only audience-selection mechanism available.

• Weak: “Premium trail running shoes. Shop now.” (No one self-selects. Everyone scrolls.)
• Strong: “If you’ve ever rolled an ankle on a rocky descent because your shoes weren’t built for it - this is for you.” (The right person stops. The wrong person scrolls. The algorithm learns.)

Every click, every purchase, every engagement signal trains the broad algorithm to find more people who respond the same way. Your click-through rate becomes a quality signal, not just a performance metric - it tells the algorithm who your buyer is.

This is why creative velocity matters more in a broad targeting environment. More variants give the algorithm more signal to work with. Test different hooks, different formats, different problem statements. The algorithm needs diverse creative inputs to find diverse pockets of buyers. Use the Noble Growth Ad Calculator to model how creative performance compounds into ROAS over time.

---

When This Doesn’t Work

Broad targeting is not universally correct. There are real situations where it underperforms.

Very early campaigns with no pixel data. The algorithm needs conversion history to optimize. If you are launching a brand-new account with zero purchase events, broad targeting is essentially guessing. In this case, use interest targeting or lookalikes as training wheels until you have built up enough conversion signal - then test broad.

Extremely niche B2B products. If your total addressable market on Meta is genuinely tiny - specialized industrial equipment, niche professional services - the algorithm may never see enough conversions to build an effective model. Interest targeting can act as useful guardrails here.

Very small budgets. The algorithm needs data to learn. If your daily budget is so low that you are getting a handful of conversions per week, there is not enough signal for broad to work. You need a minimum threshold of conversion volume for the machine learning to do its job.

For everyone else - consumer products, DTC, SaaS, local services, anyone with a reasonably broad potential market and enough budget to generate consistent conversions - broad prospecting paired with strong creative is the move.

---

The Real Optimization

The brands spending hours building the perfect interest stack are optimizing the wrong thing. They are tweaking audience settings when the algorithm has already surpassed their ability to manually select who should see an ad.

The brands winning on Meta in 2026 are the ones who have shifted that time and energy into creative testing. New hooks every week. New formats. New angles on the same core message. Testing whether UGC outperforms polished video. Testing whether a problem-first hook outperforms a product-first hook. That is the work that actually moves performance now.

The structure is simple: go broad for prospecting, use custom audiences for retargeting, and let your creative do the qualifying. The sophistication is not in the audience settings. It is in the creative strategy. Get that right, and the algorithm will do the rest.

---