Every week, I audit at least 5-10 Meta ad accounts. And without fail, I see the same costly mistakes being repeated over and over again — by small businesses, by growing D2C brands, and even by some agencies that should know better. These aren't subtle mistakes. They're account-killers that silently drain budgets while delivering mediocre results.
If you're running Meta ads and wondering why your ROAS is stuck, chances are you're making at least 3-4 of these mistakes right now. Let's fix that.
Too Many Ad Sets — Ad Set Fatigue Is Real
Running 12-15 separate ad sets with ₹200/day budgets each is one of the most common mistakes I see. When your budget is spread too thin, no single ad set has enough data to exit Meta's learning phase (which requires 50 conversions per ad set per week). The algorithm literally cannot optimise properly, and you end up with inflated CPMs, zero learning, and wasted spend.
The 2025 rule: fewer ad sets, larger budgets per set. Meta's algorithm rewards consolidation. One well-funded ad set with 5 creative variants inside it will almost always outperform 10 separate ad sets with different audiences and tiny budgets.
Consolidate to 2-3 core ad sets per campaign. Use Campaign Budget Optimisation (CBO) and let Meta allocate spend dynamically. Target minimum ₹1,500-2,000/day per ad set to give the algorithm enough fuel to learn.
Resisting Broad Targeting in 2025
Interest targeting was Meta ads 2018. In 2025, manually stacking audiences like "interested in yoga + health food + fitness + meditation" almost always underperforms compared to broad targeting. Meta's AI has seen billions of conversions. It knows who buys products like yours far better than any interest cluster you can manually construct.
Advertisers who cling to hyper-specific interest stacks are paying a premium for less reach, worse algorithm training, and lower ROAS. Broad targeting gives Meta's system the freedom it needs to find your actual buyers — not just people who clicked an interest category once three years ago.
Switch to Advantage+ audiences or broad targeting (age, gender, location only). Upload your existing customer list as a seed audience. Test broad vs. interest for 2 weeks with identical creative — you'll almost certainly see broad win on cost per conversion.
Not Testing Enough Creative Variants
Running a single ad or two static images is not a creative strategy — it's a prayer. Creative fatigue sets in fast, especially if your campaign is spending ₹3,000-5,000+ per day. Frequency rises, CPMs increase, ROAS collapses, and most advertisers assume "Meta stopped working" — when really, they just ran out of fresh creative.
The winning advertisers on Meta in 2025 are treating their creative pipeline like a manufacturing operation: constantly producing, testing, iterating, and scaling winners. They're testing hooks, formats, durations, messaging angles, and offers simultaneously.
Commit to launching 3-5 new creative variants every week. Test different hooks in the first 3 seconds of each video. Track CTR (link click-through rate) to identify which hooks win, then test variations of winners. Never run fewer than 5 active creative variants per campaign.
Wrong Campaign Objective for Your Goal
Choosing "Reach" or "Awareness" when you want sales is like hiring a delivery driver to design your logo. Meta optimises for exactly what you tell it to — if you tell it to maximise video views, it will find people who watch videos, not people who buy things. This is an expensive lesson most advertisers learn only after burning ₹30,000-50,000 on the wrong objective.
Similarly, jumping straight to conversion campaigns before your pixel has data is counterproductive. The pixel needs minimum 50 conversions per week to optimise properly — without that data, even a purchase-objective campaign will underperform.
Use Sales (formerly Conversions) objective for e-commerce and direct response. Use Lead Generation for service businesses. If your pixel is new, start with Add to Cart as your conversion event until you have 50+ weekly events, then switch to Purchase.
No Retargeting Funnel — Leaving 70% of Revenue on the Table
Studies consistently show that 97-98% of first-time website visitors don't buy. If you're not retargeting those visitors, you're paying to acquire traffic and then letting 97% of it walk away forever. A three-layer retargeting funnel — warm engagers, cart abandoners, checkout abandoners — can add 30-50% to your total Meta revenue without increasing your prospecting budget at all.
Retargeting audiences convert at dramatically higher rates (often 5-10x) compared to cold traffic, which means retargeting ads generate the highest ROAS of any campaign type — yet most small advertisers skip this entirely.
Build a retargeting campaign immediately. Layer 1: Website visitors 1-30 days (social proof + offer). Layer 2: Add-to-cart abandoners 1-7 days (urgency + free shipping). Layer 3: Checkout abandoners 0-48 hours (time-sensitive discount). Allocate 30-40% of total budget here.
Want Us to Audit Your Meta Ad Account?
We'll find every rupee being wasted in your account and give you a 30-minute action plan to fix it — completely free.
Book Your Free Audit →Not Excluding Existing Customers from Prospecting
Running your new-customer acquisition campaigns to people who already bought from you last month is a wasteful practice that inflates CPMs and skews your ROAS data. When an existing customer converts through a prospecting campaign, it looks like acquisition success — but you've simply re-sold to someone who was already going to buy. Meanwhile, you're paying premium CPMs for this "discovery."
Beyond budget waste, this creates a poor customer experience — existing loyal customers seeing ads pitching them as if they've never heard of you is jarring and can actually damage brand perception.
Create a custom audience of all purchasers from the last 90-180 days. Add this as an exclusion to all prospecting ad sets. Separately, create a specific re-purchase or cross-sell campaign targeting existing customers with relevant follow-up offers — this typically has excellent ROAS.
Ignoring Mobile Optimisation
Over 85% of Meta ad impressions are served on mobile devices in India. Yet the majority of small business ad accounts I audit are still running landscape images designed for desktop, linking to non-mobile-optimised landing pages that take 6+ seconds to load. Every one of these friction points kills conversions and destroys ROAS.
Mobile optimisation isn't just about screen size — it's about the entire mobile user experience: vertical creative formats, fast page loads, thumb-friendly CTAs, and checkout flows designed for one-thumb navigation.
Design all creatives in 9:16 vertical format first, then adapt for other placements. Test your landing page load speed on mobile using Google PageSpeed Insights — target under 2.5 seconds. Ensure your checkout is simplified and mobile-native. Use mobile-specific ad placements (Stories, Reels).
Daily Budget Too Low for the Learning Phase
Meta's learning phase requires your ad set to generate 50 conversion events within a 7-day period. If your daily budget is ₹300-500 and your cost per purchase is ₹800, you will mathematically never exit the learning phase. Your campaign will be stuck in perpetual "learning limited" status, which means poor delivery, inflated CPMs, and inconsistent results.
This is one of the most common rookie mistakes, and it's understandable — advertisers want to "test with small budgets." But underfunding a campaign doesn't reduce risk; it guarantees failure. A properly funded campaign that exits the learning phase almost always delivers dramatically better results.
Budget minimum ₹1,000-1,500 per day per ad set if your CPA target is ₹500-800. For higher-cost conversions, scale accordingly. If budget is truly limited, use the "Add to Cart" event instead of "Purchase" — it fires more frequently and allows proper learning at lower spend levels.
Not Using Meta Pixel Properly (or at All)
The Meta Pixel is the brain of your ad account — without it, you're flying completely blind. But equally harmful is a poorly configured pixel that's firing incorrectly, tracking the wrong events, or missing 40-60% of conversions due to iOS privacy changes. Many advertisers assume their pixel is "working" if the green dot is on, without ever verifying that events are firing correctly and being attributed properly.
Without clean pixel data, Meta's algorithm can't identify your buyers, can't build lookalike audiences, and can't optimise for your desired outcomes. The result is wasted spend on the wrong audiences and misleading ROAS numbers that don't reflect reality.
Set up Conversions API (CAPI) via your e-commerce platform (Shopify, WooCommerce) or through a middleware like Stape.io. Verify pixel events in Events Manager — check that Purchase, Add to Cart, and View Content events are firing correctly. Enable Advanced Matching to improve attribution accuracy.
Stopping Campaigns Too Early
The most expensive mistake of all: killing campaigns before they've had time to prove themselves. Most advertisers look at Day 1 or Day 3 results and panic — CPAs are high, ROAS is low, "Meta doesn't work for my business." But this is completely normal during the learning phase. The algorithm is still identifying who responds to your ads and optimising delivery accordingly.
Stopping campaigns prematurely and restarting them costs you the learning phase data you've already paid for and resets the clock entirely. It's the equivalent of planting a seed, checking it after 2 days, deciding it "won't grow," ripping it out, and starting again.
Give every new campaign a minimum 7-14 day evaluation window before making major changes. During this period, only make small adjustments (never change budget by more than 20% at once — this resets learning). Judge performance only after 50+ conversion events. If results are still poor after proper learning, then systematically test creative and audience changes.
"The brands that win with Meta ads aren't necessarily spending more — they're making fewer mistakes. Fix these 10, and you'll outperform 80% of advertisers in your category."
Run through this list as a checklist for your own ad account. If you're making even 3-4 of these mistakes, fixing them alone could double your ROAS without adding a single rupee to your budget. The opportunity is there — it's just hiding behind bad habits and outdated strategies.