Everyone claims they can run cheap Facebook ads. Few can prove it with real numbers. Here is the exact targeting methodology, creative framework, and bidding strategy we used to bring a major e-commerce brand's cost-per-click from $2.50 down to $0.19 — and how you can replicate it for your brand.
The Brief: A Broken Ad Account
When the client came to us, they were spending ৳85,000 per month on Facebook ads and getting roughly 34,000 clicks. That is about $2.28 per click at the exchange rate at the time. For an e-commerce brand in Bangladesh where average order values sit between ৳1,200–৳2,500, that math simply does not work.
Their existing setup had all the classic symptoms: broad targeting, carousel ads with product catalogue images, optimising for link clicks instead of purchases, and a single ad set blasting everyone from 18 to 65 in Bangladesh. No segmentation. No funnel structure. Just spend and hope.
Our goal was clear: get the CPC below $0.50 within 30 days and below $0.25 within 60 days. We ended up hitting $0.19 by day 47.
Step 1 — Audience Architecture (The 3-Layer System)
The biggest mistake most advertisers make is treating their audience as one blob of people. We build every account with three distinct layers, each with a separate budget, creative, and objective.
- Layer 1 — Cold Traffic (Top of Funnel): Detailed interest targeting layered with lookalike audiences. We built 1%, 2%, and 3% lookalikes from the client's purchase list from the past 90 days. These people behaviorally resemble existing buyers but have never heard of the brand.
- Layer 2 — Warm Traffic (Middle of Funnel): Website visitors (all pages, last 30 days), Instagram engagers (last 60 days), and video viewers (50%+ watched). These people know who the brand is. We just need to bring them back.
- Layer 3 — Hot Traffic (Bottom of Funnel): Add-to-cart abandoners, product page viewers, and past purchasers for repeat purchase campaigns. This layer always has the lowest CPC and highest ROAS.
We allocated budget as 50% / 30% / 20% across cold / warm / hot. Most brands reverse this or ignore the funnel entirely.
"The audience is never the problem. The structure is the problem. Segment properly and the algorithm will reward you every single time."
— Azad Hossain, Head of PerformanceStep 2 — The Creative Framework That Drove Down CPC
Facebook's algorithm rewards ads that get high engagement relative to their reach. A high CTR tells the platform "people like this ad" and it lowers your CPM as a reward. So reducing CPC is really about increasing CTR — and CTR is really about creative quality.
We tested three creative formats against each other in the first two weeks:
- UGC-Style Video: 15–30 second videos shot on a phone, showing real products in real homes or hands. No studio lighting. No branded overlays. The winner by a huge margin — CTR of 4.8% vs 1.2% for polished studio shots.
- Problem-Solution Static: Single image ad with a bold Bengali text headline calling out a specific pain point, followed by the product as the solution. CTR of 3.1%.
- Catalogue Carousels: Standard product catalogue ads. CTR of 0.9%. We killed these by day 10.
We also discovered a copy formula that worked consistently: Open with a relatable situation → Identify the problem → Present the product as the solution → Urgency close. In Bengali for the cold traffic layer, in English for the warm and hot layers (these users tend to be more digitally engaged).
Step 3 — Bidding Strategy
Most beginners use Lowest Cost bidding and let Meta spend freely. This works for testing but bleeds budget once you scale. We switched to a hybrid approach:
- Cold traffic: Lowest Cost with a daily budget cap per ad set (not a campaign budget). This gives the algorithm freedom to find the right audience while controlling spend velocity.
- Warm traffic: Cost Cap set at 2× the target CPA. This tells Meta "don't spend more than this per result" — it slows delivery but keeps costs predictable.
- Hot traffic: Lowest Cost with a bid cap at 1.5× our maximum acceptable CPC. Small audiences respond well to this without inflating costs.
Pro tip: Never run Advantage+ Campaign Budget (formerly CBO) when you're still in the learning phase. Let each ad set learn independently first. Consolidate budgets only after you have winning ad sets identified — usually after 50+ conversion events per ad set.
What Did NOT Work (Equally Important)
We tried a few approaches that failed and cost us two weeks of learning budget:
- Targeting by income or profession in Bangladesh — Meta's income data for BD is notoriously inaccurate. We got better results with behavioural signals (online shoppers, frequent travellers) than income brackets.
- Boosting posts — absolutely do not do this. You lose all control over objective, placement, and bidding. It exists to generate revenue for Meta, not results for advertisers.
- Running ads 24/7 without dayparting — we found CPC was 37% higher between 2am–6am. Scheduling ads off during those hours dropped our average CPC further.
Can You Replicate This?
Yes — with caveats. The $0.19 CPC was achievable because of the specific product category (everyday household goods with broad appeal), a well-built pixel with 18 months of purchase data, and a client willing to commit to a 90-day testing budget rather than demanding instant ROAS on day one.
If you have a new pixel, no purchase history, and expect results in week one — this strategy will not produce $0.19 CPC. It will, however, get you substantially lower CPCs than the industry average of $0.90–$1.50 in the Bangladesh market.
The fundamentals hold regardless of budget or category: 3-layer audience structure, UGC-first creative, and conversion-optimised bidding. Start there.