Fixzoo Facebook Ads Case Study — E-Commerce Retargeting & Dynamic Product Ads

When Fixzoo approached us in early 2025, they were already operating a well-established e-commerce platform in Bangladesh specialising in mobile phone accessories, repair tools, and gadget parts. Their product catalogue spanned over 800 SKUs — from tempered glass screen protectors and phone cases to charging cables, batteries, display assemblies, and precision repair toolkits. They had a functional Shopify store, a modest social media following, and were running Facebook ads with varying degrees of success.

But there was a fundamental problem: their Facebook advertising was treating every user the same. Whether someone had just landed on the homepage for the first time or had added three items to their cart and abandoned the checkout, they saw the same generic ad creative. There was no personalisation, no behavioural segmentation, and — critically — no systematic abandoned cart recovery strategy. Their cost per purchase (CPP) was climbing steadily month over month, and their retargeting efforts, such as they existed, were limited to manually created Custom Audiences that quickly became stale.

This case study documents how we rebuilt Fixzoo’s Facebook Ads infrastructure from the ground up — implementing a complete Dynamic Product Ad (DPA) ecosystem, building multi-layered retargeting funnels, and deploying an automated abandoned cart recovery system that transformed their e-commerce performance. The result: a 5.1x Retargeting ROAS, a 35% reduction in Cost Per Purchase, and a sustainable growth framework for scaling their online sales.

If you run an e-commerce brand in Bangladesh — whether you sell electronics, fashion, home goods, or any product category with a substantial catalogue — the DPA and retargeting strategies outlined here are directly applicable to your business.

Project Snapshot

Metric	Detail
Client	Fixzoo — E-Commerce Platform (Mobile Accessories & Gadget Parts)
Industry	E-Commerce / Retail — Electronics Accessories & Repair Parts
Target Market	Bangladesh (Dhaka, Chattogram, nationwide via e-commerce)
Platform	Facebook Ads (Meta Ads Manager) + Meta Product Catalog
Campaign Type	Dynamic Product Ads (DPA), Multi-Layer Retargeting Funnels, Abandoned Cart Recovery
Monthly Ad Spend	BDT 85,000–1,10,000
Products in Catalog	800+ SKUs across 40+ categories
Retargeting ROAS Achieved	5.1x
Cost Per Purchase Reduction	−35%
Cart Recovery Rate via DPA	14.7%
Timeline	February 2025 – May 2025 (120 days)
Expert	Kanok Miah — Digital Marketing & Facebook Ads Strategist

The Problem: A One-Size-Fits-All Approach to a Multi-SKU Catalogue

Fixzoo’s Facebook advertising before our engagement had several structural weaknesses that were silently draining budget and limiting growth. These are problems we see repeatedly among Bangladeshi e-commerce brands that have crossed the initial product-market fit stage but have not yet built the advertising infrastructure to scale efficiently.

1. No Product Catalogue Integration

Fixzoo was running standard image and video ads — manually uploading individual product photos to Facebook Ads Manager. With over 800 SKUs, this meant that the vast majority of their catalogue was never advertised. Worse, every ad featured the same offer and the same primary product image, regardless of who was viewing it. A customer who had browsed iPhone 16 screen protectors saw the same ad as someone who had browsed Samsung charging cables. There was zero personalisation.

Meta’s Product Catalog — the foundational tool for Dynamic Product Ads — was not set up at all. This meant Fixzoo was unable to leverage Facebook’s most powerful e-commerce feature: the ability to automatically show each user the products most relevant to them, based on their browsing behaviour, purchase history, and predicted intent.

2. No Structured Retargeting Funnel

Retargeting existed in the most basic form: a single Custom Audience of “All Website Visitors (last 30 days)” with a single ad set showing a generic “Come back and shop” message. There was no differentiation between:

A user who viewed one product page for 3 seconds
A user who browsed 15 products across 3 sessions
A user who added items to their cart but didn’t check out
A past purchaser who might be interested in complementary accessories

This lack of segmentation meant the retargeting campaigns suffered from audience saturation (the same people seeing the same ad repeatedly), ad fatigue, and ultimately, declining click-through rates and rising costs.

3. Abandoned Cart Recovery: Entirely Manual

Fixzoo was experiencing a cart abandonment rate of approximately 72% — typical for e-commerce in Bangladesh, where Cash on Delivery (COD) is the dominant payment method and customers often add items to cart as a “wishlist” rather than with immediate purchase intent. However, the brand had no automated recovery mechanism. The only follow-up was a manual WhatsApp message sent (when staff had time) to customers who had provided their phone number during checkout. This recovered perhaps 3–4 carts per week at best.

There was no Facebook-based abandoned cart retargeting, no dynamic ad showing the exact products left behind, no time-based sequencing of recovery messages, and no discount incentive strategy for recovering at-risk carts.

4. Rising Costs from Audience Fatigue

Because the same audiences were being retargeted with the same creative for weeks on end, frequency — the average number of times a user saw an ad — was climbing. At the time of our audit, the retargeting ad set had a frequency of 7.4 over 30 days. This is far beyond the optimal 2.0–3.0 range for e-commerce. High frequency leads to ad blindness, higher CPMs (as Facebook penalises ads with declining engagement), and a poor brand experience for the customer.

5. No Cross-Sell or Upsell to Past Purchasers

Fixzoo had a database of approximately 3,200 past customers, but this audience was being entirely ignored by the advertising strategy. There were no cross-sell campaigns (“You bought a phone case — here’s a tempered glass protector for the same model”), no upsell campaigns (“Upgrade to our premium leather case”), and no re-engagement campaigns for customers who had not purchased in 90+ days. This was the single biggest missed opportunity.

The Solution: A Dynamic Product Ad Ecosystem with Multi-Layer Retargeting

Our strategy for Fixzoo was built on five core pillars: Product Catalog Infrastructure, Dynamic Product Ad Campaign Architecture, Multi-Layer Retargeting Funnels, Automated Abandoned Cart Recovery via DPA, and Creative & Offer Strategy Aligned to Catalogue Depth.

Pillar 1: Product Catalog Infrastructure

Everything in DPA begins with the product catalogue. We needed to create a complete, optimised, and continuously synced product feed that Facebook could use to serve personalised ads at scale.

Catalogue Setup via Shopify Channel Integration

Fixzoo’s Shopify store was already well-structured with clean product data, but the feed required several optimisations for Meta’s Commerce platform:

Product Titles: Restructured for search clarity and visual skimming. Format: “Brand | Model | Product Type | Key Feature” — e.g., “Samsung | Galaxy S25 Ultra | 9H Tempered Glass | Anti-Blue Light”. Titles included Bengali keywords alongside English for bilingual browsing behaviour. Example: “আইফোন ১৬ প্রো ম্যাক্স | iPhone 16 Pro Max | Premium Silicone Case | MagSafe Compatible”.
Google Product Category (GPF) Mapping: Every SKU was mapped to the correct GPF taxonomy — crucial for Meta’s algorithm to understand product relationships and serve relevant cross-sell recommendations. For example, “iPhone 16 Screen Protector” was categorised under “Electronics > Communications > Telephony > Mobile Phone Accessories > Screen Protectors” with a parent-child relationship to “iPhone 16 Cases”.
Custom Labels for Campaign Segmentation: We added 5 custom label fields to categorise products by: (a) Price Tier — Budget (under BDT 300), Mid (BDT 300–1,000), Premium (BDT 1,000+), (b) Category — Screen Protection, Cases, Charging, Audio, Tools, (c) Seasonality — Ramadan Offers, Monsoon Gear, New Arrivals, (d) Margin Tier — High Margin Promotable, Low Margin, (e) Bestseller Status — Top 100, Mid-Range, Slow Mover.
Inventory Sync: Real-time inventory tracking was enabled so that out-of-stock products were automatically paused in ad delivery — preventing wasted spend on products that couldn’t be fulfilled.
Image Optimisation: All product images were standardised to 1:1 aspect ratio at 1,080×1,080 pixels with clean white backgrounds. We added secondary lifestyle images showing the product in use (e.g., a phone with the case installed) for the additional image slots in the catalogue.
Sale Price & Discount Fields: The price and sale_price fields were populated with campaign-specific discount data. This allowed DPA ads to display strikethrough pricing with dynamic discount badges — a feature that significantly improved CTR.

The catalogue was set up to sync automatically via Shopify’s native Meta channel integration, with updates propagating within 4–6 hours of any product change. For urgent inventory updates (flash sales, low-stock alerts), we used Meta’s scheduled feed upload as a fallback.

Pixel & Event Foundation

Before launching any DPA campaigns, we audited and rebuilt Fixzoo’s Meta pixel infrastructure:

Standard Event Audit: We discovered that the ViewContent event was firing correctly on product pages, but AddToCart was inconsistent (especially on quick-view modal add-to-cart interactions), and Purchase was delayed by up to 24 hours for COD orders because the event was only firing on Shopify order fulfilment rather than order placement.
Conversions API (CAPI) Setup: We configured Meta’s Conversions API via the Shopify-Meta plugin, sending server-side Purchase, AddToCart, and ViewContent events directly from Shopify’s backend. This gave Meta real-time purchase signals regardless of browser tracking limitations (iOS 14.5+ ATT, browser cookie restrictions, ad blockers). CAPI increased tracked purchase volume by 31% within the first week.
Event Deduplication: We implemented Meta’s recommended deduplication protocol using eventID and event_source_url parameters, ensuring that when both the browser pixel and CAPI fired for the same conversion, Meta counted it only once.
Micro-Event Tracking for DPA Optimisation: Beyond standard events, we implemented custom micro-events for DPA-specific optimisation: AddToWishlist, Subscribe (for back-in-stock alerts), and Lead (for service inquiries — Fixzoo also offers phone repair services alongside parts sales). These micro-events provided richer behavioural signals for Meta’s algorithm to personalise product recommendations.

Pillar 2: Dynamic Product Ad Campaign Architecture

With the catalogue live and the pixel foundation solid, we launched Fixzoo’s DPA campaign architecture. This consisted of four campaign tiers, each with a distinct role in the customer journey:

Tier 1 — Dynamic Prospecting (Catalogue Sales — Broad)

This campaign introduced Fixzoo’s products to users who had never visited the website. Instead of showing a single hero product to everyone, Meta’s algorithm dynamically selected which product to show to which person based on their demographic profile, device usage patterns, and real-time browsing behaviour across Facebook and Instagram.

Targeting: Bangladesh-wide. Gender: All. Age: 16–55. No detailed targeting interests — pure broad targeting, letting Facebook’s algorithm determine relevance based on catalogue performance data.

Product Set: All 800+ active SKUs with the following filters applied: in-stock = true, price between BDT 50 and BDT 15,000. Low-margin items under BDT 50 were excluded because the cost per purchase would exceed the product value.

Creative Format: Single image DPA (primary image from catalogue) + Carousel DPA (up to 10 products displayed in a swipeable carousel). The carousel format outperformed single image by 27% in CTR for this tier.

Budget: 50% of total monthly spend was allocated to this prospecting tier.

Tier 2 — Dynamic Retargeting (Catalogue Sales — Custom Audiences)

This was the core of our DPA retargeting strategy — showing users the exact products they had interacted with, in a time-based sequence that mirrored intent level.

Ad Set 2A — ViewContent (1–14 days): Users who viewed a product page but did not add to cart. The DPA showed the precise product(s) they viewed, with a “Still thinking about it?” message and social proof overlay — “27 people bought this today”. Frequency cap: 3 per 7 days. Budget: 15% of total.

Ad Set 2B — AddToCart (1–7 days): Users who added a product to cart but did not complete purchase. The DPA showed the cart-abandoned product(s) front and centre, with urgency messaging — “Low stock — only 5 left!” and “Your cart is waiting”. This ad set used the ATC event for customisation, showing related accessories alongside the abandoned product. Frequency cap: 4 per 7 days. Budget: 18% of total.

Ad Set 2C — InitiateCheckout (1–3 days): Users who reached the checkout page but didn’t complete. Highest-intent segment. DPA showed the abandoned items with a limited-time offer overlay — “Free shipping on orders over BDT 500” or “5% off your order — code FIXZOO5”. Frequency cap: 5 per 7 days (higher frequency justified by higher intent). Budget: 10% of total.

Tier 3 — Cross-Sell & Upsell to Past Purchasers

Past customers are the highest-value audience in any e-commerce business. We created a dedicated DPA campaign targeting anyone who had purchased from Fixzoo in the last 180 days, with product sets dynamically tailored to their purchase history:

Complementary Accessories: A customer who bought an iPhone 16 case would see ads for iPhone 16 screen protectors, charging cables with the same colour finish, and MagSafe-compatible accessories. Meta’s algorithm determined the best cross-sell based on product catalogue relationships.
Upgrade Path: Customers who purchased budget-range products (under BDT 300) were shown premium alternatives (BDT 800–1,500) for their next purchase, with comparison messaging — “Upgrade to our Premium Series”.
Category Expansion: A customer who only bought phone cases would be introduced to other categories — audio (wireless earbuds), charging (power banks), or tools (repair kits).
Re-engagement (90–180 days): Customers who hadn’t purchased in 90+ days were shown a “We miss you” sequence with new arrivals and a special returning-customer discount code.

Budget: 12% of total monthly spend. Despite the relatively low budget share, this tier delivered the highest ROAS (7.8x) of any campaign.

Tier 4 — Lookalike Audience Expansion for Catalogue

With clean purchase data flowing through the pixel (now capturing 80–100 purchase events daily), we built Lookalike audiences specifically for the product catalogue:

1% Lookalike (Purchase-Based): Narrowest, highest-intent audience, approximating Fixzoo’s best customers. Used for the top 20% of products by margin and AOV.
3% Lookalike (Purchase-Based): Broader reach with strong purchase propensity. Used as the primary DPA prospecting audience alongside broad targeting.
Value-Based Lookalike (Top 25% by LTV): An audience modelled on Fixzoo’s most valuable customers (those with highest lifetime value). This Lookalike consistently delivered 25–30% higher AOV than the standard 1% Lookalike.
Category-Specific Lookalikes: We created separate Lookalikes for high-volume categories — “Screen Protector Purchasers”, “Case Purchasers”, “Charging Accessory Purchasers”. This allowed us to serve category-specific DPA creative to audiences that were pre-disposed to buy in that category.

Budget: 5% of total. Highly efficient with strong ROAS but limited scale ceiling.

Pillar 3: Multi-Layer Retargeting Funnels

Beyond the standard DPA audience tiers, we built a sophisticated retargeting funnel that sequenced touchpoints based on user behaviour, time elapsed, and creative format.

Layer 1 — Dynamic Product Showcase (Hours 0–24)

Within hours of a user’s first interaction with Fixzoo’s website, the DPA system would serve them a personalised ad showing the products they had browsed. This “warm lead” touchpoint used single-image DPA format with a clean, product-focused creative — no discounts, just product presentation. The goal at this stage was to reinforce product interest without undervaluing the brand with aggressive discounting.

Layer 2 — Urgency & Social Proof (Days 1–3)

If the user did not return within 24 hours, the next ad layer introduced scarcity and social proof elements. Creative format shifted to a carousel DPA with the following overlay elements dynamically applied:

Stock count: “Only X left in stock” (pulled from the inventory field in the catalogue)
Recent purchase count: “X people bought this in the last 48 hours”
Rating stars: Average product rating (from catalogue review data)
Delivery promise: “Free delivery across Bangladesh”

Layer 3 — Offer-Led Retargeting (Days 3–7)

For users who still hadn’t converted after 3 days, we introduced a limited-time offer layer. This used DPA with the sale_price field activated, showing a strikethrough original price and the discounted price alongside a time-bound offer: “Special price ends in 48 hours.” The offer was unique to the retargeting audience — not available to new prospects — creating a sense of exclusivity.

Layer 4 — Video Re-engagement (Days 7–14)

Users who had not converted after seeing the offer layer were moved into a video re-engagement sequence. This was not a DPA format — we used manual video creative showing product demos, unboxing footage, and customer testimonials for Fixzoo. The goal of this layer was not immediate conversion but rebuilding engagement signals so that the user was “warm” again for future DPA targeting. Once a user watched 50%+ of a video, they were re-entered into the Layer 1 DPA sequence.

Pillar 4: Automated Abandoned Cart Recovery via DPA

The abandoned cart recovery system was Fixzoo’s single biggest source of incremental revenue. Here is exactly how we built it:

Stage 1 — Real-Time AddToCart + InitiateCheckout Signal Capture

With CAPI and the browser pixel both firing AddToCart and InitiateCheckout events in real time, Meta had near-instantaneous data on cart abandonment. The moment a user left the website with items in their cart, the DPA system received the signal.

Stage 2 — Time-Based DPA Sequence (Hours 1–48)

We created a timed DPA sequence for abandoned carts:

1 Hour Post-Abandonment: First DPA touchpoint showing the exact abandoned product(s) with the headline “You left these behind” and a subtle CTA “Pick up where you left off”. No discount at this stage. Format: Single image DPA.
6 Hours Post-Abandonment: Second DPA touchpoint, now showing abandoned products alongside complementary accessories (cross-sell). Example: user abandoned an iPhone 16 case — the ad now showed the case + a matching screen protector + a MagSafe wallet. The headline: “Complete your setup.” Format: Carousel DPA.
24 Hours Post-Abandonment: Third DPA touchpoint introducing a limited-time cart recovery incentive. The ad automatically applied a 5% discount code (FIXCART5) and showed the abandoned products with a countdown: “Your cart items are reserved for 24 hours — complete your order with 5% off.” Format: Single image DPA with urgency text overlay.
48 Hours Post-Abandonment: Final touchpoint with a stronger offer: “Free shipping on your order — plus 5% off. Don’t miss out.” If the user had not returned by 72 hours, they were sequenced out of the cart recovery funnel and into the standard retargeting layers.

Stage 3 — Combined WhatsApp + Facebook Recovery

For users who provided their phone number during the checkout process, we combined the DPA sequence with a WhatsApp follow-up (sent via Shopify’s automated notification system). The WhatsApp message arrived 2 hours after abandonment, referencing the Facebook ad the user would have seen — creating a multi-channel recovery touchpoint. This combined approach recovered 14.7% of all abandoned carts, with 62% of recoveries attributed to the Facebook DPA sequence and 38% to the WhatsApp follow-up.

Pillar 5: Creative & Offer Strategy Aligned to Catalogue Depth

DPA ads are inherently creative-light because the product image is pulled dynamically from the catalogue. However, the creative wrapper around the product — the ad copy, headline, description, CTA, and any overlay elements — required meticulous strategy.

Headline & Primary Text Strategy

We developed 12 headline templates and 18 primary text templates, systematically A/B tested over the first 30 days. The winning combinations for each funnel stage were:

Prospecting (Broad): “Premium [Product Category] — Starting at BDT [Price]” | Primary text focused on free delivery and 7-day returns — the two trust factors that Bangladeshi e-commerce shoppers care about most.
ViewContent Retargeting: “Still exploring? Here’s a closer look at [Product Name]” | Primary text included a brief product benefit statement + social proof (ratings).
AddToCart Retargeting: “Your cart misses [Product Name] — complete your order” | Primary text highlighted limited stock and delivery timeline.
InitiateCheckout Retargeting: “Don’t lose your items — complete checkout now with free shipping” | Primary text included the discount code.
Cross-Sell / Past Customers: “You bought [Past Product Category] — you’ll love [Related Product Category]” | Primary text positioned as a personalised recommendation.

Language & Cultural Localisation

All ad copy was written in Bengali (primary) with English keywords included for search visibility. We tested purely Bengali copy against bilingual copy and found that bilingual headlines (e.g., “আপনার iPhone এর জন্য Premium Tempered Glass — всего BDT 250 থেকে”) delivered 18% higher CTR than Bengali-only or English-only variants. The mix signals relevance to both language segments in the Bangladeshi market.

Dynamic Creative Optimisation

We enabled Meta’s Dynamic Creative (DC) feature on the prospecting DPA campaigns. DC automatically tested different combinations of creative elements — primary text, headline, description, image (from catalogue), and CTA button — within a single ad. After the learning phase (approximately 50 conversions per ad set), Meta’s algorithm allocated 80% of impressions to the best-performing combination. This improved the prospecting CPA by 19% compared to manual creative uploads.

The Results: 120-Day Transformation

The impact of the DPA and retargeting infrastructure was visible within the first 14 days and compounded significantly over the 120-day engagement period.

Metric	Before (Jan 2025)	After 120 Days (May 2025)	Change
Retargeting ROAS	1.8x	5.1x	+183%
Overall ROAS (Blended)	2.1x	4.3x	+105%
Cost Per Purchase (CPP)	BDT 420	BDT 273	−35%
Cart Recovery Rate	~2% (manual)	14.7%	+635%
Retargeting Frequency (30-day avg)	7.4	2.9	−61%
Average Order Value (AOV)	BDT 680	BDT 940	+38%
Monthly Tracked Purchase Events	~480	~2,400	+400%
Returning Customer Rate	12%	31%	+158%
Monthly Revenue (FB-Driven)	BDT ~4.2 Lakh	BDT ~14.8 Lakh	+252%
Impressions per Ad Set (Weekly)	~85,000	~3,20,000	+276%

Performance Breakdown by Campaign Tier (May 2025)

Dynamic Prospecting (Broad — Tier 1): 38% of total sales. Blended ROAS: 3.2x. The carousel DPA format significantly outperformed single image for prospecting (4.1x vs 2.8x ROAS). The top-selling product categories through prospecting were tempered glass screen protectors (28% of revenue), followed by silicone phone cases (22%) and fast-charging cables (18%). Broad targeting with no interest restrictions consistently delivered the lowest CPM (BDT 98) and widest reach.

Dynamic Retargeting (Tier 2): 34% of total sales. Blended ROAS: 5.1x. The InitiateCheckout ad set delivered the highest ROAS at 7.8x, followed by AddToCart at 6.2x and ViewContent at 3.8x. The timed sequencing (Layers 1–4) was critical: Layer 2 (urgency/social proof) converted 23% more users than Layer 1 (product showcase alone), and Layer 3 (offer-led) converted a further 18% who had not responded to Layers 1 or 2.

Cross-Sell / Past Customers (Tier 3): 18% of total sales. Blended ROAS: 7.8x — the highest ROAS of any campaign tier. Category expansion ads (introducing past customers to new product categories) drove 43% of this tier’s revenue. The “upgrade path” ads (selling premium versions of previously purchased products) delivered the highest AOV at BDT 1,450.

Lookalike Expansion (Tier 4): 10% of total sales. Blended ROAS: 5.6x. The Value-Based Lookalike (top 25% LTV) outperformed the standard 1% LAL by 22% in ROAS and 35% in AOV. Category-specific Lookalikes (e.g., “Screen Protector Lookalike”) delivered higher CTR than general purchase Lookalikes but had limited audience size (typically 80,000–2,00,000 people vs 15,00,000–30,00,000 for general Lookalikes).

Abandoned Cart Recovery: Deep Dive

The automated cart recovery system was transformative for Fixzoo’s bottom line. Key statistics from the 120-day period:

Total Abandoned Carts Tracked: 12,847 (across all payment methods including COD, bKash, and card)
Carts Recovered via Facebook DPA Sequence: 1,156 (9.0% recovery rate)
Carts Recovered via WhatsApp Follow-Up: 734 (5.7% recovery rate)
Total Recovery Rate: 14.7% (1,890 carts)
Revenue Recovered: BDT 12.6 Lakh (approximately $10,500 USD)
Cost Per Recovered Cart (Facebook DPA only): BDT 62 — dramatically lower than the standard CPP of BDT 273
Best-Performing Recovery Touchpoint: The 24-hour DPA touchpoint with the 5% discount code converted 38% of all recovered carts — the most effective single touchpoint in the sequence.
Optimal Recovery Window: 68% of all cart recoveries occurred within the first 24 hours of abandonment. Recovery rate dropped sharply after 48 hours (only 12% of recoveries happened after the 48-hour mark).

The data clearly shows that speed is the most important variable in cart recovery. The first 24 hours account for over two-thirds of all recoveries, and the optimal strategy is a multi-touch sequence that escalates from product reminder (0–6 hours) to social proof (6–24 hours) to discount incentive (24–48 hours).

Key Takeaways for E-Commerce Brands Scaling with Facebook Ads in Bangladesh

Dynamic Product Ads Are Not Optional for Multi-SKU E-Commerce. If your catalogue exceeds even 20 SKUs, DPA is no longer a “nice to have” — it is the most efficient way to advertise. Manual ad creation for each product does not scale. DPA automates product selection, personalisation, and delivery at a level that no manual campaign structure can match. Fixzoo’s 800+ SKU catalogue would have required 800+ individual ad sets under the old approach. DPA handled it with a single campaign.
Segmented Retargeting Funnels Dramatically Outperform Single-Audience Retargeting. The difference between a single “All Website Visitors” retargeting campaign and a segmented DPA funnel is the difference between a 1.8x ROAS and a 5.1x ROAS. The behavioural segmentation — ViewContent vs AddToCart vs InitiateCheckout — allows you to tailor messaging, offer strength, and frequency to intent level. Higher intent deserves stronger offers and higher frequency.
Abandoned Cart Recovery via DPA Is the Highest-ROI Investment in E-Commerce Advertising. Fixzoo’s cart recovery DPA sequence delivered a return of BDT 62 cost per recovered cart versus a BDT 273 standard CPP. That is over 4x efficiency compared to prospecting. The investment required is purely strategic — setting up the DPA sequence costs nothing in additional tools or software — but the revenue impact is immediate and compounding.
The First 24 Hours Are Everything for Cart Recovery. Two-thirds of all cart recoveries happen within the first 24 hours of abandonment. Your DPA retargeting sequence must be triggered in near real-time (within 1 hour of abandonment, ideally sooner). Delay reduces recovery rates exponentially. If your pixel or CAPI setup has latency beyond 2–3 hours, you are losing recoverable revenue.
Past Customers Are Your Most Profitable Audience — Treat Them That Way. Fixzoo’s cross-sell and upsell campaigns to past purchasers delivered a 7.8x ROAS — the highest of any campaign tier. Yet most e-commerce brands in Bangladesh spend zero ad budget on past-customer retention. A dedicated cross-sell DPA campaign, even at 10–15% of total budget, will outperform most prospecting campaigns on ROAS. The key is using purchase history data to inform product recommendations — not just blasting “new arrivals” to everyone.
Product Feed Quality Determines DPA Success. The single biggest technical success factor for DPA is the quality of your product feed. Products with high-resolution images, accurate titles, correct Google Product Category mapping, and populated custom labels will be prioritised by Meta’s algorithm and will deliver better CTR and conversion rates. Invest heavily in feed optimisation before launching DPA campaigns — it is the foundation everything else rests on.
Bengali + Bilingual Copy Wins in Bangladesh. Our A/B testing consistently showed that bilingual ad copy (Bengali primary text with English product names and keywords) outperformed Bengali-only and English-only variants. The Bangladeshi Facebook user’s brain switches between languages fluidly while scrolling. Meet them in that bilingual space for maximum relevance.
CAPI + Pixel Dual Setup Is Non-Negotiable for E-Commerce Tracking. With Apple’s iOS 14.5+ App Tracking Transparency (ATT) and ongoing browser privacy changes, the browser pixel alone is no longer sufficient for reliable conversion tracking. Meta’s Conversions API (CAPI) provides a server-side backup that captures events the browser misses. Fixzoo saw a 31% increase in tracked purchase events immediately after CAPI installation — events that would otherwise have been invisible to Meta’s optimisation algorithm.
Offer Sequencing Matters More Than Offer Size. The best-performing cart recovery sequence did not lead with a discount. The first touchpoint (1 hour post-abandonment) was a clean product reminder with no offer. The discount was introduced only at the third touchpoint (24 hours). This sequencing — reminder, then social proof, then discount, then urgency — respects the user’s decision-making process and preserves margin by converting users at higher price points before resorting to discount incentives.
Scale Is Governed by Pixel Maturity, Not Budget. The single biggest lesson from Fixzoo’s 120-day transformation: you cannot buy your way out of a bad pixel. Before our engagement, Fixzoo was spending BDT 85,000+ per month with a broken tracking foundation. After fixing the pixel and CAPI, we were able to scale budget by 30% while simultaneously reducing CPP by 35%. The tracking infrastructure must be solid before you increase spend. More budget on a broken pixel just means you waste more money faster.

Related Case Studies

If you found this case study valuable, you may also be interested in our other Facebook Ads and digital marketing case studies:

See more Facebook Ads case studies — Browse our complete portfolio of Facebook Ads, SEO, and digital marketing success stories.
Iqra Cadet Madrasha Facebook Ads Case Study — How we drove admissions through Facebook lead generation campaigns for a religious educational institution.
Dohar Malabis Facebook Ads Case Study — Islamic clothing brand scaling from BDT 2.0 Lakh to BDT 7.5 Lakh monthly sales through conversion-focused campaigns.
Ansar VDP Sadar Tangail Facebook Ads Case Study — Recruitment marketing via Facebook Ads for a government security force unit in Tangail.

Conclusion

The transformation of Fixzoo‘s Facebook advertising from a generic, one-size-fits-all approach to a sophisticated Dynamic Product Ad ecosystem with multi-layer retargeting and automated abandoned cart recovery is a textbook example of what is possible when e-commerce advertising infrastructure is built correctly.

The results speak for themselves: a 5.1x Retargeting ROAS, a 35% reduction in Cost Per Purchase, a 14.7% cart recovery rate recovering over BDT 12.6 Lakh in otherwise lost revenue, and a sustainable, scalable advertising engine that continues to improve as Meta’s algorithm accumulates more data from the clean pixel and CAPI events.

None of these results came from a single magic bullet. They came from systematically building the right infrastructure: a complete product catalogue with optimised feed data, a pixel and CAPI foundation capturing every relevant event with zero data loss, a DPA campaign architecture that respects user intent and purchase stage, an abandoned cart recovery sequence timed to the user’s behavioural window, and a creative and offer strategy rooted in the realities of the Bangladeshi e-commerce market — bilingual copy, culturally authentic messaging, and trust-building social proof.

For any e-commerce brand in Bangladesh with a catalogue of 50+ SKUs, the question is not whether to implement Dynamic Product Ads and strategic retargeting — it is how quickly you can build the infrastructure and start compounding the learning. Every day without DPA is a day your competitors are using Facebook’s most powerful e-commerce tool to take market share.

Contact Kanok Miah to discuss how we can apply the same DPA retargeting and abandoned cart recovery methodology to your e-commerce brand.