Instagram Ads Targeting: Advanced Strategies Beyond Basic Interests

Why Interest Targeting Doesn't Work Anymore

Let me be blunt: if you're still building Instagram ad audiences by stacking interests like "Yoga + Organic Food + Sustainability," you're wasting money.

I ran an experiment last quarter. Same ad creative, same budget, three different targeting methods:

Interest-based (the old-school way)

Custom audiences + lookalikes

Advantage+ (Meta's automated targeting)

The interest-based campaign had a cost per acquisition 2.3x higher than the others.

Here's why interest targeting is broken:

Meta's algorithm has gotten too good. It doesn't need your help figuring out who to show ads to. When you add interests, you're basically telling Meta "I know better than your billion-dollar machine learning system." Spoiler: you don't. Privacy changes killed accuracy. iOS 14.5 gutted Meta's tracking. The "interest" data Meta has on users is way less reliable than it was in 2020. People lie. Someone who liked a yoga page five years ago isn't necessarily your customer today. Meta's interest data is stale and cluttered.

Don't get me wrong — interest targeting isn't completely dead. But it's not where you should start.

Custom Audiences That Actually Perform

Custom audiences are where the magic happens.

These are people who've already interacted with your business: website visitors, email subscribers, app users, etc. They're warm traffic, and warm traffic converts.

Here are the custom audiences I use most:

Website Visitors (with Nuance)

Don't just create a generic "all website visitors" audience. Get specific.

I build separate audiences for:

• Homepage visitors (casual browsers)
• Product page viewers (interested but not committed)
• Add-to-cart abandoners (close to converting)
• Checkout abandoners (VERY close to converting)

Then I run different ads to each group. Someone who abandoned their cart needs a discount or urgency. Someone who just hit your homepage needs education.

Last month I ran a campaign for a fitness app. We retargeted checkout abandoners with a "Complete Your Order" ad offering 15% off. CPA was $12. Our cold traffic CPA was $64.

Email List Uploads

If you have an email list, upload it to Meta. Even if it's small.

I worked with a client who had 2,400 email subscribers. We uploaded the list, created a lookalike, and ran ads to the lookalike audience.

Result: 34% lower CPA than interest-based targeting.

Pro tip: segment your list. Don't just upload everyone. Upload your best customers (high LTV, repeat buyers) separately from inactive subscribers. The quality of your seed audience matters.

Video Viewers

This one's underrated.

If you're running video ads, create custom audiences of people who watched 50%, 75%, or 95% of your video. Then retarget them with a conversion-focused ad.

Someone who watched 95% of your video is way more interested than someone who watched 3 seconds.

I use this for lead gen campaigns. First ad: educational video. Second ad: "Ready to try it?" offer to people who watched most of the video.

Engagement Audiences (I'll cover this more below)

People who engaged with your Instagram or Facebook content. Huge opportunity here that most people ignore.

Offline Events

If you have a physical store or attend events, upload customer data from in-person interactions. Meta can match emails and phone numbers to profiles.

I did this for a local gym. Uploaded their member list, created a lookalike, and ran ads. Brought in 47 new members in 30 days.

Real talk: I used to ignore custom audiences because I thought "my list is too small." Turns out even a 500-person list can power a lookalike that performs better than interests.

Lookalike Audience Strategies

Lookalike audiences are Meta's way of finding new people who "look like" your existing customers.

They're one of the most powerful targeting tools you have — if you use them right.

Start with a Quality Seed Audience

Your lookalike is only as good as your seed audience.

I see people create lookalikes from "all website visitors" or "everyone who liked my page." That's garbage data. You're telling Meta "find me more random people."

Instead, seed your lookalike with:

• Purchasers (best option)
• High-value customers (even better)
• Add-to-cart users (if you don't have enough purchasers)
• Email subscribers who've opened 3+ emails

The more specific and high-quality your seed, the better your lookalike will perform.

Test Different Percentages

Meta lets you create lookalikes from 1% to 10%.

• 1% = most similar to your seed (smaller audience)
• 10% = least similar (huge audience)

I always start with 1%. It's the most accurate. If 1% works and you need to scale, test 2-3%.

I rarely go above 5%. At that point, you're basically just targeting "people in this country," which isn't much better than broad targeting.

Layer Lookalikes with Exclusions

Here's a trick most guides don't mention: exclude audiences from your lookalikes.

For example, create a 1% lookalike of purchasers, then exclude:

• Existing customers (no point selling to them again)
• Recent website visitors (they'll see retargeting ads instead)
• Low-quality traffic sources (if you track this)

This keeps your lookalike focused on NEW, high-quality prospects.

Use Value-Based Lookalikes (If You Have Conversion Data)

If you have a pixel tracking purchases with values, Meta can create a "value-based lookalike."

Instead of finding people similar to all buyers, it finds people similar to your BEST buyers (highest spending).

I used this for a SaaS client. Regular lookalike CPA: $78. Value-based lookalike CPA: $54.

The difference? We were targeting people more likely to buy the premium plan.

Refresh Your Seed Audiences

Don't create a lookalike once and let it run forever.

I refresh mine every 60-90 days. Customer behavior changes. Your seed audience should reflect recent high-quality customers, not people who bought from you in 2022.

Want better targeting insights? AdsMAA analyzes your audience performance and tells you exactly which segments are driving conversions.

Advantage+ Targeting (When to Use It)

Advantage+ is Meta's way of saying "let us handle targeting."

You give minimal input (age, location, maybe gender), and Meta's algorithm decides who to show ads to.

I was skeptical at first. I like control. But after testing it on 20+ campaigns, I'm a convert — with caveats.

When Advantage+ Works

You have a proven offer. If your creative and offer already convert, Advantage+ will find more people like your converters. You have pixel data. Meta needs at least 50 conversions per week to optimize well. If you're getting 10 conversions a month, Advantage+ will struggle. You're scaling. Once you've maxed out your custom audiences and lookalikes, Advantage+ helps you reach new people without manually testing interests.

I ran Advantage+ for an e-commerce store selling home decor. They were already profitable with manual targeting. Advantage+ found new pockets of buyers we hadn't reached and lowered CPA by 18%.

When Advantage+ Fails

You have no conversion data. If Meta doesn't know who converts, it'll just show ads to anyone. You'll waste money. Your offer is niche. If you sell something ultra-specific (like left-handed ergonomic keyboards for gamers), Meta's broad targeting will miss the mark. Stick with manual audiences. You need tight control. If you have strict brand guidelines or can't afford to waste budget on testing, Advantage+ is risky.

How I Use Advantage+

I don't rely on it exclusively. I run it alongside manual campaigns.

Typical setup:

• Campaign 1: Retargeting (custom audiences)
• Campaign 2: 1% lookalike (manual)
• Campaign 3: Advantage+ (automated)

Then I compare performance after spending $200-300 on each. I keep the winners, kill the losers.

I also set exclusions in Advantage+. I exclude existing customers and people who visited the site in the last 7 days. No point wasting impressions on them.

Engagement Audiences Nobody Talks About

Here's a goldmine most advertisers ignore: engagement audiences.

These are people who interacted with your Instagram or Facebook content — liked, commented, shared, saved, or viewed your profile.

Why does this matter?

Because these people already know who you are. They're not cold traffic. They're lukewarm.

Instagram Engagers

Create audiences of people who:

• Engaged with your Instagram account (any action in the last 365 days)
• Engaged in the last 90 days (more recent = warmer)
• Sent you a DM
• Saved your posts (high-intent action)

I ran an ad campaign targeting people who saved our posts in the last 90 days. CTR was 4.2% — 3x higher than cold traffic.

Why? Because saving a post means "I want to come back to this later." These people are interested.

Video Viewers (Deep Engagement)

I mentioned this earlier, but it's worth repeating.

Create audiences based on video watch percentage:

• 25% (casual interest)
• 50% (moderate interest)
• 75% (high interest)
• 95% (very high interest)

Then run different ads to each group.

75%+ watchers get direct offers. 25-50% watchers get more educational content.

Lead Form Openers

If you run lead gen campaigns, create an audience of people who opened your lead form but didn't submit.

They were interested enough to click but didn't convert. Maybe they got distracted. Maybe the form was too long.

Retarget them with a simpler offer or social proof.

Page Engagers on Facebook

Don't forget Facebook. If you have a Facebook page, create audiences of people who engaged with it.

I know, Facebook feels old. But the targeting works. I've had clients get 20-30% lower CPAs by retargeting Facebook page engagers on Instagram.

Audience Layering Tactics

Okay, here's where it gets advanced.

Instead of using one audience type, you layer multiple signals to create super-targeted groups.

Tactic 1: Lookalike + Interest (Narrow)

Take a 1% lookalike audience and add a narrow interest layer.

Example: 1% lookalike of purchasers + "Must also be interested in Fitness."

This works when you want to scale a lookalike but keep it focused.

I used this for a supplement brand. Straight 1% lookalike was performing well but started to plateau. Added "Health & Wellness" as a narrow filter. CPA dropped by 12%.

Tactic 2: Custom Audience Stacking

Combine multiple custom audiences into one.

Example: Create an audience of people who:

• Visited your website in the last 30 days
• AND engaged with your Instagram in the last 90 days
• AND watched 75% of a video ad

This is ultra-warm traffic. These people have interacted with you multiple times across multiple channels.

I call this "audience stacking," and it's one of my favorite retargeting strategies.

Tactic 3: Exclusion Layering

Sometimes the best targeting is knowing who NOT to show ads to.

I exclude:

• Recent purchasers (unless it's a repeat-purchase product)
• Low-quality traffic (high bounce rate pages, accidental clicks)
• Competitors' employees (yes, I've done this)

One client was wasting budget showing ads to people who clicked but immediately bounced (under 5 seconds on site). We excluded them. CPA dropped by 22%.

Tactic 4: Geo + Demo + Custom Audience

Layer location and demographic filters on top of custom audiences.

Example: Retarget website visitors who live in California and are aged 25-34.

This works for local businesses or products with specific regional appeal.

I ran this for a regional food brand. Retargeted website visitors in their delivery zone. Conversion rate went from 2.1% to 4.8%.

Tactic 5: Sequential Audiences (Funnel Approach)

Build a sequence where people move through progressively warmer audiences.

Example:

• Ad 1: Target cold traffic (1% lookalike) with educational content
• Ad 2: Retarget video viewers (75%+) with product demo
• Ad 3: Retarget website visitors with discount offer

This mimics a traditional marketing funnel but uses Instagram ads.

I use this for high-ticket products. You can't sell a $2K course to cold traffic. But you can educate them, warm them up, then close them.

Here's the truth: Most advertisers overcomplicate targeting.

They stack 10 interests, layer in weird demographics, and wonder why their ads don't work.

The best targeting strategies are simple:

Start with custom audiences (warm traffic)

Build lookalikes from high-quality seeds

Test Advantage+ if you have data

Use engagement audiences for retargeting

Layer smartly, but don't go overboard

The goal isn't to get fancy. It's to show ads to people who are most likely to convert.

Stop guessing. Start testing. And let the data tell you what works.

Ready to optimize your targeting? Get a free audit from AdsMAA and see which audiences are actually driving your results.