Most people who start selling online in Bangladesh choose their first product on a hunch. A cousin's page is doing well with three-pieces. A gadget video went viral. A wholesaler in Chawkbazar offered a good price. So they buy 200 pieces, boost a post for ৳2,000 and wait. Sometimes it works. More often the boxes sit under the bed for months.
There's a better way to choose, and it costs nothing but an hour of reading. Every ad that runs on Facebook and Instagram in Bangladesh is public. An ad that keeps running week after week is one the seller keeps paying for, and nobody keeps paying for an ad that brings no orders. Read the ads the right way and they tell you which products are selling.
This guide shows you how: what to look for in the ads, how to check the seller behind a product, and how to do the maths before you buy stock. The numbers come from BanglaRadar, which records the Facebook and Instagram ads running in Bangladesh. On 30 September 2026 that was about 189,000 ads from 35,352 online shops.
What a "winning product" really means
A winning product isn't one that sold well once. It's one that keeps selling, at a price that still leaves you money after paying for:
- the courier, inside and outside Dhaka,
- parcels that come back because the buyer refused the cash-on-delivery order,
- the ad money it took to get each order,
- packaging, the courier's cash-collection fee and your own time.
That's why the same product can make one seller good money and sink another. The product matters, but so do the price, the courier maths and the ad. Keep all four in mind through every step below.
Step 1: Find ads that have been running for months
This is the strongest free signal there is, because most ads die young. Of the shop ads running on 30 September:
- 37% had started in the last 7 days,
- 36% had been running for 8 to 30 days,
- 17% for 31 to 90 days,
- and only 10% for more than 90 days.
The typical running ad was just 13 days old. Sellers test constantly and switch off whatever doesn't sell, so an ad that has survived three, six or twelve months is one of the few that paid for itself. Only about one shop in three has even a single ad that has passed 30 days.
To find them in BanglaRadar, search for a product word in Bangla or English, set Status: Active and sort by Longest running. In Meta's own Ad Library you can do the same by hand: open each ad and read its "Started running on" date.
Searching "স্কুল ব্যাগ" (school bag), active ads, longest running first. Two different shops have kept waterproof school bag ads running for 300 to 414 days.
Look at what this tells you. It isn't one shop getting lucky with one ad. Two separate sellers have advertised almost the same waterproof school bag, in almost the same words, for more than a year. The demand is steady, and so is the competition. Both are worth knowing before you order stock.
A few long runners we found
These products all had ads running for more than 60 days on 30 September. They were picked at random from ads whose copy the seller had launched at least three times:
- a foldable plastic stool, sold as "a big solution for a small space": 196 days, copy launched 55 times,
- a herbal baby massage oil: 243 days, copy launched 57 times,
- waterproof school bags at "50% off": 173 to 414 days, from more than one shop,
- a henna and indigo pack for grey hair and beards: 238 days,
- a pocket blood pressure monitor: 194 days,
- ready-to-wear sarees for young girls, with no pins or pleats to fix: 216 days,
- a small hand saw for branches and bamboo: 213 days,
- cooling gel patches for fever and headache: 652 days.
Notice the pattern. Nearly every one solves a small, specific problem you can show in a 30-second video: nowhere to sit, a bag that gets soaked in the rain, grey hair, a saree that's hard to put on a child. That's more useful than any "top 10 trending products" list, because you can open the actual ad that sells each one and see how it's done.
Step 2: Look for ad copy the seller keeps relaunching
The second signal is quieter but just as telling. When an ad works, sellers don't only leave it running. They launch it again and again: the same words in a new ad, sometimes with a new video, to reach new people or try a bigger budget. Every relaunch is the seller putting more money behind that product.
On 30 September, 38% of running shop ads used copy their seller had launched at least 3 times, and 14% used copy launched 10 times or more. The record was one piece of ad copy launched 334 times.
Sorted by "Most variants": the running ads whose copy has been launched the most. The blue number on each card is how many times.
One finding surprised us. Ads launched in the last week carried more relaunches on average (10.8) than ads that had been running for over a month (5.1). Sellers who are scaling don't just leave one ad running. They keep relaunching their best copy as fresh ads. So don't dismiss an ad because it's two days old: if its copy has been launched 300 times, it's a proven winner in a new wrapper.
The strongest sign of all is an ad that is both old and heavily relaunched.
Step 3: Study the seller, not just the ad
Once a product looks promising, open the advertiser behind it. You want to know how long they've been at it, how hard they're pushing right now and how they take orders.
The seller of the foldable stool: 229 ads running, the oldest for 305 days, and 30 new ads in the last 30 days, 50% more than the month before. "Testing intensity 25x" means each piece of copy has been launched 25 times on average.
What to read on a page like this:
- The longest run. Months of steady advertising means the business has survived its own mistakes. A page that appeared last month and already runs 200 ads is either very well funded or about to burn out.
- Launches this month against last month. Rising means they're scaling. Falling can mean the product is getting crowded.
- Few scripts, many versions. This seller runs 229 ads from only 29 different scripts. That's what a seller with a handful of proven products looks like.
- Where the ads send people. A website, Messenger or WhatsApp. It shows how they take orders, and what you'd need to set up to compete.
Then open one of their ads and read it properly.
One of the stool ads: 55 versions of the same copy, the same video and text used in 25 ads, and the website it sends people to, registered two years ago.
Visit the landing page too. Write down the price, the delivery charge, whether they ask for advance payment and what they promise about returns. That's the offer you'll be competing with, and you'll need to beat it or match it.
If you don't know where to start, look at the biggest shops first. Our Top 50 e-commerce companies in Bangladesh ranks online shops by how many ads they're running right now.
The live ranking: ads running, new this week, each shop's longest-running ad, and whether it sells through a website or Messenger.
Step 4: Learn how Bangladeshi shops actually sell
Before you write your own ad, it helps to know what normal looks like. Across the shop ads running on 30 September:
- Where ads send buyers: a website 49%, Messenger 37%, WhatsApp 9%. Chat is still almost half of all shop ads, so if you don't have a website yet, you're not behind.
- The button: "Order now" on 38.5%, "Send message" on 36%, "Shop now" on 9% and "WhatsApp" on 9%.
- Video: about two ads in three use video. Among ads running for 30+ days it's 70%, against 66% for ads launched this week. Video helps, but a clear photo with a clear offer still works.
- Offers: 35% mention an offer, a discount or a percentage off.
- Cash on delivery: only 12.5% spell it out, because buyers take it for granted. About 6.5% promise free delivery, about 6% promise delivery anywhere in the country, and just 4% mention returns or exchange.
- Places: 14% of ads name a district. Dhaka appears in 11%, far ahead of Chattogram, Sylhet, Khulna and Rajshahi.
The last two points leave room for you. Few ads promise easy returns, and few speak to buyers outside Dhaka. Either one could set your ad apart.
What prices sell?
One ad in five writes its price in the copy. Taking the highest price each of those ads mentions:
- the middle price is ৳850, and half of all prices fall between ৳450 and ৳1,250,
- 53% are between ৳500 and ৳1,500, 29% are under ৳500, and only 18% are above ৳1,500,
- the most common price of all is ৳999, then ৳990,
- 55% end in 90, 99, 50 or 49,
- ads running for 30+ days had a lower middle price (৳750) than ads launched this week (৳899).
The ৳500 to ৳1,500 range is where most shop ads live: cheap enough for a buyer to order from a single ad, cash on delivery, without a long think. Above ৳2,000, expect more questions on Messenger and more time before people decide.
Which products last longest?
We also matched product words in the ad copy. It's a rough method, so read these as directions rather than exact counts, but the pattern is clear. Beauty and health products are more common among ads that have run for a month than among new ones, and clothing is the other way round:
- Beauty and skin care: 12.4% of ads running 30+ days, 9.9% of ads launched this week,
- Health products: 8.0% against 5.6%,
- Clothing: 9.7% against 14.3%.
Clothing sellers launch all the time because designs change every season, so a single clothing ad rarely lasts. Beauty and health products are bought again and again, so one good ad can run for months. If you'd rather make one strong video and run it for a long time, look at products people buy repeatedly.
Step 5: Do the COD maths before you buy stock
This is where most new sellers lose money. They compare the buying price with the selling price and stop there. Take a product you can buy for ৳350 and sell for ৳990, with the buyer paying the delivery charge. It looks like ৳640 profit per order. Now count properly for 100 orders confirmed on Messenger. These are example numbers, so put in your own:
- 80 delivered, 20 returned. Assume one parcel in five comes back, refused at the door or never collected. Some pages see more.
- Money in: 80 × ৳990 = ৳79,200.
- Stock sold: 80 × ৳350 = ৳28,000. Returned pieces go back on the shelf if they're undamaged.
- Returned parcels: you still pay the courier and collect nothing. At about ৳120 each, 20 returns cost ৳2,400.
- Cash-collection fee: couriers typically take about 1% of the cash they collect, so ৳792. Check your courier's rates.
- Ads: at ৳150 of ad spend per confirmed order, ৳15,000.
- Packaging: ৳15 × 100 = ৳1,500.
Profit: ৳31,508, or about ৳394 per delivered order, not ৳640. Still good. But if the ads cost ৳300 per order (common while you're learning) and 30% of parcels come back, the same product makes about ৳9,000 on 100 orders, around ৳129 per delivered order. One bad week and it's a loss.
A rule of thumb many sellers use: if you can't sell at about three times what the product costs you, the margin probably won't survive returns and ad costs. It also helps when the product is light and doesn't break in a courier bag. Delivery is cheaper and fewer parcels come back damaged.
Step 6: Test small, then scale what works
- Pick two or three products that passed the steps above. Don't bet everything on one.
- Make one honest video for each. Show the problem, the product solving it, the price and the delivery promise. Study how the long-running ads use their first three seconds and their first line of text, but don't copy their video or their brand. A copied video can get your ad reported and taken down.
- Start with a small daily budget you can afford to lose, for three or four days per product. Track the cost per message or per confirmed order, not likes.
- Keep the product whose cost per confirmed order still leaves a profit after the Step 5 maths. Switch the rest off without regret.
- Scale the way the winners do: relaunch the ad that works, try new videos with the same message, and raise the budget step by step.
Step 7: Keep watching after you launch
Winning products don't stay winners forever. New sellers join, prices drop, buyers move on. Two habits keep you ahead:
- Watch your competitors. Add the five or ten sellers you compete with to a tracker and see when they launch new ads, stop old ones or suddenly scale up. When a competitor launches 20 ads in a week, something new is working for them.
- Save your product searches. Save the search from Step 1 and get an email when new ads match it. When new sellers start pushing the same product, you'll know within the hour, not when your orders drop.
The Brand tracker: every ad a tracked brand launches or stops, with an email as it happens, once a day or once a week.
Saving the school bag search with email alerts on. New matching ads arrive in your inbox along with your other alerts.
Mistakes that burn money
- Arriving at the end of the run. If ten sellers started the same product last month and prices are dropping, you're late. Look at the start dates: lots of brand-new ads and only a few very old ones usually means a crowded product.
- Heavy or fragile products. Glass, large electronics and heavy items cost more to ship and come back broken. Light, sturdy products are kinder to your margin.
- Health promises you can't keep. Plenty of long-running ads promise to cure diabetes, gastric problems or joint pain. Meta restricts ads that make health claims and can take them down or limit your page. Medicines, including Ayurvedic, Unani and herbal ones, must be registered with the Directorate General of Drug Administration (DGDA), and many foods and cosmetics need BSTI certification. Don't build a shop on a claim that can get it shut down.
- Fake branded goods. Copies of well-known brands sold as originals invite reports, page bans and legal trouble.
- Not reading your inbox. The questions buyers ask on Messenger (Which size? Is the colour the same as the photo? Will it fit a six-year-old?) tell you exactly what to put in your next ad.
A seasonal calendar for Bangladeshi sellers
- Ramadan and Eid-ul-Fitr: panjabis, three-pieces, sarees, attar, dates and gift boxes. Start before Ramadan, because many people shop early.
- Eid-ul-Adha: knives, choppers and kitchen tools, spices, storage and clothing.
- Falgun and Valentine's Day (mid-February), Pohela Boishakh (14 April): yellow, red and white clothing, flowers, jewellery and gifts.
- Winter (December to February): blankets, jackets, shawls, lotions, lip care and heaters.
- Monsoon (June to September): umbrellas, raincoats, waterproof bags, mosquito nets and repellents.
- New school year (January): school bags, stationery, lunch boxes and water bottles.
Eid moves about 11 days earlier every year, so check the date each year, and look back at which ads ran longest in the same season last year.
Frequently asked questions
What sells best online in Bangladesh?
Going by the ads running on 30 September 2026, the biggest groups are clothing, beauty and skin care, baby and kids' products, bags, shoes and watches, health products, and home and kitchen items. But the best product for you is one where you can beat the offers already out there on price, quality, delivery or trust.
How do I see my competitors' Facebook ads?
Meta's Ad Library shows every active ad: choose Bangladesh and search for a page. BanglaRadar records the same ads and adds Bangla and English search, filters for days running, relaunches and where ads send buyers, plus brand tracking and email alerts. Here's how it compares with Meta's Ad Library.
How long must an ad run before I call the product a winner?
There's no fixed rule. But an ad still running after 30 days is already in the top quarter of shop ads, and one past 90 days is in the top 10%. Give the most weight to ads that are both old and relaunched many times.
Do I have to offer cash on delivery?
For a new shop, almost always. Buyers trust it, which is why so few ads even mention it. What you can control is the return rate: confirm each order by phone, show the real product in your video, and state the size, colour and delivery time clearly.
How much money do I need to test a product?
Enough for a small batch of two or three products and three or four days of ads for each. Where you can, buy small first orders from local wholesalers before importing in bulk, so a product that doesn't sell costs you little.
Finding a winning product isn't luck, and it isn't a secret list. The sellers who've already found one are showing you in public, every day. Read their long-running ads, check the seller, do your own COD maths and test small. When you're ready to search every Facebook and Instagram ad running in Bangladesh, in Bangla or English, you can create a free BanglaRadar account.