Instagram Ads Guide 2025: Everything You Need to Know

Why Instagram Still Matters in 2025

Let me be honest — Instagram ads aren't the shiny new toy anymore. TikTok gets all the hype. But here's what nobody talks about: Instagram still converts like crazy if you know what you're doing.

I manage about $180K/month in Instagram ad spend across e-commerce, SaaS, and lead gen clients. Our average ROAS is 3.8x. That's not because we're geniuses — it's because we stopped listening to what Meta tells us to do and started testing what actually works.

The platform has changed massively since iOS 14.5 killed tracking. If you're still running Instagram ads like it's 2020, you're burning money. But if you adapt to how the algorithm actually works now, there's still gold here.

Real talk: Most Instagram ads guides are written by people who haven't spent a dollar on the platform. This one isn't. I'm going to tell you what works, what doesn't, and what Meta pushes that you should ignore.

Campaign Objectives (And Which Ones Are BS)

Meta gives you six campaign objectives now. You need two of them. Maybe three. Here's the breakdown:

Sales (formerly Conversions) — This is your bread and butter for e-commerce and lead gen. The algorithm optimizes for actual purchases or form fills. Use this 80% of the time. Traffic — Useful if you're driving to a blog post or long-form content where conversion tracking is tricky. Otherwise, skip it. Traffic campaigns optimize for clicks, not quality. I've seen CPCs drop 40% switching from Traffic to Sales for the same landing page. Engagement — Only use this if you're building social proof on a new account or trying to grow followers. Don't use it thinking engaged users will magically convert later. They won't. Leads — The on-platform lead forms can work for high-volume, low-quality lead gen (think insurance quotes). But if you need qualified leads, send them to your actual landing page with a Sales campaign. Lead quality from instant forms is usually garbage. App promotion — If you have an app, use this. Otherwise, irrelevant. Awareness — Honestly, skip this unless you have $50K+ monthly budgets and you're doing brand building. For 99% of advertisers, this is a waste. "Awareness" doesn't pay the bills.

I spent six months testing every objective across 40+ accounts. Sales campaigns won every single time for performance. Don't overthink this.

Every Ad Format Explained

Instagram gives you way too many format options. Here's what you actually need to know:

Reels Ads (Full-Screen Video)

This is the king right now. Our Reels ads get 60-70% lower CPMs than feed ads and 40% better CTRs. Why? Because Instagram desperately wants to compete with TikTok, so they're giving Reels massive reach advantages.

Specs: 9:16 vertical video, 15-90 seconds (but keep it under 30), file size under 4GB.

What works: Native-looking content. If your ad looks like an ad, you're dead. The best performing Reels we run look like normal creator content until the CTA at the end.

Stories Ads

Still solid, especially for retargeting. Stories users are in a browsing mindset, so impulse purchases work well here.

Specs: 9:16 vertical, image or video, 15 seconds max for video.

What works: Bold text overlays, single product focus, clear swipe-up CTA. Stories move fast. Don't be subtle.

Feed Ads (Image or Video)

Feed ads aren't dead, but they're expensive. We still run them for high-value products or when creative works better in square/horizontal format.

Specs: Square (1:1) or landscape (1.91:1) recommended. Carousel ads work here too (2-10 cards).

What works: Stop-the-scroll imagery. Carousels for showcasing multiple products or telling a story. Video still outperforms static images by about 30% in my tests.

Explore Ads

Here's where I disagree with most people — I turn Explore off 90% of the time. The placement costs more and converts worse than Reels or Stories. Meta will tell you Explore users are "in discovery mode" which sounds great, but in practice, those clicks are expensive and low-intent.

Only exception: Brand awareness campaigns with beautiful creative.

A word on carousels: Everyone sleeps on carousel ads. We've had some of our best performers with 5-card carousels showing different product angles or use cases. Engagement Rate averages 1.8% vs 1.1% for single-image ads. People like to swipe.

Placements: What to Turn Off Immediately

By default, Meta wants you to run "Advantage+ placements" (formerly Automatic Placements) which spreads your ads across Instagram, Facebook, Messenger, and Audience Network. This is terrible advice.

Here's my standard placement strategy:

Always on:

• Instagram Reels
• Instagram Stories
• Instagram Feed

Usually on:

• Facebook Feed (if audience is 35+)
• Facebook Reels (testing phase, results mixed)

Always off:

• Audience Network (this is random app traffic, mostly accidental clicks)
• Facebook Right Column (do people even see this?)
• Messenger (ultra low quality)
• Instagram Explore (too expensive unless specific reason)

I tested Advantage+ placements vs manual for three months across 23 campaigns. Manual placement selection beat Advantage+ on ROAS every single time. Don't be lazy here.

The reason is simple — Meta optimizes for cheapest results, not best results. Audience Network clicks are dirt cheap and worthless. The algorithm will dump budget there if you let it.

Want to audit your current placement performance? Try AdsMAA's free campaign analysis — it breaks down spend and conversions by placement so you can see what's actually working.

Creative Guidelines That Actually Matter

Meta has a million creative "best practices." Most are cargo cult advice that doesn't move the needle. Here's what actually matters:

The 3-Second Rule

If your ad doesn't hook attention in the first 3 seconds, it's dead. Instagram users scroll FAST. We A/B test hooks relentlessly — same ad, different opening 3 seconds. CTR swings by 200-300%.

Pattern interrupts that work: bold text, fast cuts, people looking directly at camera, unexpected movement.

Text Overlay Limits

Instagram doesn't have the old 20% text rule Facebook had, but ads with tons of text still perform worse. Keep text overlays to 2-3 short lines max. Let the visual do the talking.

Sound-On Creative

80% of Instagram users watch with sound on (vs only 15% on Facebook). Use this. Ads with voiceovers, music, or sound effects outperform silent videos by about 35% in our tests.

Don't just slap a music track on. The audio should add value — either explaining the product, creating emotion, or giving context the visuals alone can't.

UGC Crushes Polished Creative

User-generated content looks authentic. Polished brand content looks like an ad. Guess which one performs better?

We had an e-commerce client spending $40K/month on a professional videographer for ads. ROAS was 2.1x. We ran a test with iPhone-shot customer testimonials. ROAS jumped to 4.3x. The "ugly" ads won by a landslide.

Authenticity beats production quality every time on Instagram in 2025.

The CTA Mistake Everyone Makes

Don't bury the call-to-action. I see ads all the time where the CTA is a tiny button at the end or just assumed. Make it obvious.

"Shop Now," "Get 20% Off," "Try Free for 30 Days" — tell people exactly what to do. Our CTR increases 40-60% when we add a strong verbal CTA in the last 5 seconds of video ads.

Targeting in the iOS 14.5+ Era

Here's the uncomfortable truth: targeting doesn't matter as much as it used to. iOS privacy changes killed detailed audience insights. Meta's algorithm now does most of the heavy lifting through Advantage+ Audience.

But there are still levers that matter:

Start Broad

I know this feels wrong, but broad targeting outperforms detailed targeting in 2025. We run most campaigns with just age, gender, and country. That's it.

The algorithm finds your customers better than you can manually. I've tested this hundreds of times. Detailed interest targeting used to win. Now it loses.

Exception: If you have a tiny niche (like "left-handed golf clubs for seniors"), some targeting helps. But for most products, go broad.

Lookalike Audiences Still Work (Kind of)

Lookalikes aren't dead, but they're weaker. We still see 20-30% better performance with 1% lookalikes of purchasers vs cold broad audiences. Not the 2-3x advantage they had in 2019, but still worth running.

Upload your best customer list (email or phone). 1% lookalike, your country only. 3-5% lookalikes are too diluted now.

Retargeting Is Your Best Friend

This hasn't changed. Retargeting still crushes. We segment by:

• Website visitors, last 30 days (broad awareness)
• Add-to-cart, last 14 days (high intent, abandoned cart)
• Video viewers 50%+ (engaged prospects)
• Instagram engagers (people who liked/commented/saved)

Retargeting campaigns run at 5-8x ROAS vs 2-4x for cold traffic. Allocate 30-40% of your budget here.

Custom Audiences from Your Data

If you have an email list, use it. Upload it, exclude existing customers (unless you're doing retention), and target those users plus a 1% lookalike.

Tools like AdsMAA can help sync your customer data with Meta automatically so you're always targeting the right segments without manual CSV uploads every week.

Budget and Bidding Strategy

Everyone asks "how much should I spend?" Wrong question. The right question is "how much do I need to spend for the algorithm to learn?"

Minimum Budget Reality

You need at least 50 optimization events per week for the algorithm to work properly. If your product is $50 with a 2% conversion rate, that means:

• 50 purchases needed
• 2,500 landing page visitors required
• At $1 CPC, that's $2,500/week minimum
• About $350/day

If you can't spend that much, either lower your conversion event (optimize for Add-to-Cart instead of Purchase) or extend your learning phase expectations.

I see people trying to run conversion campaigns at $10/day and wondering why results suck. You're not giving the system enough data to optimize.

CBO vs ABO (Campaign Budget Optimization)

Use CBO. I've tested both extensively. CBO lets Meta move budget to the best-performing ad sets automatically. You set one campaign budget, Meta distributes it.

ABO (Ad Set Budget Optimization) forces you to guess which audience will perform best upfront. You'll guess wrong and waste money on poor performers while under-funding winners.

Bid Strategies That Actually Work

Ignore most of Meta's bidding options. Here's what I use:

Highest volume (no cap) — Default choice. Let the algorithm spend your budget to get maximum conversions at whatever cost. Use this 80% of the time, especially during learning phase. Cost cap — Only use this once you know your target CPA and have strong historical data. Tell Meta "don't spend more than $X per conversion." Useful for scaling proven campaigns while controlling CPA. Bid cap — Almost never use this. Too restrictive. You'll under-deliver on budget.

The algorithm is better at bidding than you are. Don't overthink it.

Testing and Optimization Framework

Here's the testing framework I use for every new Instagram campaign:

Week 1-2: Creative Testing

Launch with 3-5 different video creatives, same targeting, same ad copy. Let them run. You're testing hooks, formats, and angles.

Kill anything with CTR under 1% after 3 days. Let the rest run for 7 days minimum before making decisions. We've had ads that looked terrible on day 2 become top performers by day 7.

Week 3-4: Scaling Winners

Take your best 1-2 creatives and scale budget by 20-30% every 3 days. Don't double overnight — that resets the learning phase and tanks performance.

Also test new variations of winning creatives. If a blue background ad wins, test red background. If a testimonial wins, test a different customer. Iterate on what works.

Ongoing: Refresh Creative Monthly

Instagram ad fatigue is real. Once your frequency hits 3-4 (people seeing the same ad 3-4 times), performance drops. We refresh creative every 3-4 weeks even for winners.

You don't need to reinvent the wheel. Take the same concept, new footage. Same hook, different product angle. Freshness matters more than novelty.

What to Actually Track

Forget vanity metrics. Track:

• ROAS or CPA (depending on business model)
• CTR (proxy for creative quality)
• Landing page conversion rate (don't blame Meta for your bad landing page)
• Cost per landing page view (measures traffic quality)

That's it. Ignore engagement rate, video views, and all the other stuff Meta tries to show you.

The Instagram + Facebook Question

Should you run Instagram only or combine with Facebook?

Depends on your audience age:

Under 35 target audience: Instagram-only campaigns often outperform combined. Facebook placements just dilute performance. 35-55 target audience: Combined Instagram + Facebook usually wins. These people use both platforms actively. 55+ target audience: Facebook-heavy (but still include Instagram Reels — everyone watches Reels).

I run them as separate campaigns when possible so I can control budget split and optimize independently. Combined campaigns tend to dump budget wherever it's cheapest, which isn't always where it converts best.

Common Mistakes I See All the Time

Switching off ads after 2 days — The algorithm needs 5-7 days minimum to stabilize. I see people panic and kill campaigns on day 2 because CPA is high. Give it time. Optimizing for engagement — Unless you sell engagement, optimize for sales. I've had clients come to me with thousands of likes and zero revenue. Likes don't pay rent. Running the same creative for months — Refresh your ads. Ad fatigue is real and expensive. Not excluding existing customers — If you're not doing retention marketing, exclude purchasers from your audience. Why pay to advertise to people who already bought? Ignoring mobile landing page experience — 90%+ of Instagram traffic is mobile. If your landing page sucks on mobile, your ads will fail no matter how good they are. Over-targeting — Detailed targeting is mostly dead. The narrower your audience, the higher your costs and the longer your learning phase. Go broader than feels comfortable.

My Actual Setup Checklist

When I launch a new Instagram campaign, here's my exact process:

Campaign objective: Sales

Campaign type: CBO (Campaign Budget Optimization)

Budget: Minimum $30-50/day, ideally $100+

Ad set: Broad targeting (age + gender + country only)

Placements: Manual — Instagram Reels, Stories, Feed only

3-5 video creatives in one ad set (different hooks)

Ad copy: 2-3 lines max, clear CTA

UTM parameters set up for tracking

Exclude existing customers (unless retention campaign)

Set it, let it run for 7 days, then optimize

That's it. No fancy audience stacking. No complicated testing matrices. Simple beats complex every time.

Final Thoughts

Instagram ads in 2025 are fundamentally different than they were even two years ago. The platform prioritizes Reels. Targeting is automated. Creative quality matters more than ever.

The good news: if you stop trying to outsmart the algorithm and work with it, Instagram ads still drive absurd ROI. My best performing campaigns right now are also my simplest campaigns.

Stop overthinking it. Test creative aggressively, let the algorithm do its job, and scale what works.

Want to see how your current Instagram campaigns stack up? Get a free audit from AdsMAA — it takes 2 minutes and you'll get a breakdown of what's working and what's wasting money.

Now go run some ads.