How to sell online in Indonesia

"Ready stock, Kak?"

You get that message forty times a day and you answer it forty times a day, and somewhere in there is an actual order you nearly missed.

Selling online in Indonesia needs four things and none of them is a website. A page with prices and stock status a buyer can read without asking. An account or number they can send money to, on a rail they already use. A courier and an honest ongkir. And a rule for cash on delivery that you decide once, in advance, rather than per order at 11pm. A BI-FAST transfer between banks clears in seconds and your bank can charge you at most Rp 2,500 for it under Bank Indonesia's cap, which is the cheapest way money will ever reach you. Cards will cost you far more than that, if you can get them at all, and we'll come to why that sentence has an "if" in it.

The deal happens in the chat. The page exists to shorten the chat

Nearly every small Indonesian shop runs the same loop. You post on Instagram or TikTok. Someone comments or DMs. You answer stock, price, size, ongkir to their kota. They ask for a discount. You agree. They transfer or they pick COD. You ship.

That loop works and you should not try to replace it. The seller who says "just buy it on my website" is asking an Indonesian buyer to skip the part where they decide whether you're real.

What you can do is cut the chat in half. Every question you answer more than twice a day belongs on a page: sizes, variants, what's ready stock and what's PO, how long PO takes, which couriers you use, ongkir by zone or at least by island, and whether you do COD. Then your DM time goes on the questions that genuinely need you, and those are the ones that close sales anyway. "Is this one thick enough for a bayi?" "Can you send before Jumat?"

If most of your orders arrive as chat messages, how to take orders on WhatsApp covers the mechanics of not losing them.

The rails your buyers already have

Indonesian buyers pay in four broad ways, and only two of them are things a small shop can run without a company behind it.

Bank transfer. BI-FAST has been running since December 2021 and moves money between banks in seconds, any hour, any day. Bank Indonesia caps what a bank may charge the sender at Rp 2,500 per transfer, so it's near enough free to the buyer. This is the workhorse. Your buyer opens their banking app, sends, screenshots. You check your own balance.

QRIS. One QR standard that every payment app in the country can scan, mandated by Bank Indonesia from January 2020. Warungs use it. Ojek drivers use it. If you have a QRIS code through your bank or a payment provider, buyers will use it happily.

E-wallets. GoPay, OVO, DANA, ShopeePay. Balance sitting in an app, usually topped up for rides and food.

Cash on delivery. The rider knocks, the buyer pays, the rider remits to the courier, the courier remits to you.

Now the honest part, and it matters before you plan anything. Sailo's complete list of payment rails is card, WhatsApp, Telegram, Instagram, email, phone, bank transfer and cash on delivery. There is no QRIS rail and no e-wallet rail. Not "coming soon", not hidden in a setting. It doesn't exist.

What sellers do instead is run those methods manually. You put your QRIS instructions or your DANA number into the bank transfer instructions field, which is free text and takes whatever you want to write, then you confirm the payment yourself from your own app. That works, and a lot of people do exactly that. Be clear about what it is, though. Sailo is holding your catalogue and your order record. Your phone is holding the money. Nobody is settling QRIS on your behalf, and if that's what you need, this is not the tool for it.

Cash on delivery is the default, not the fallback

If you've read Western e-commerce advice, COD sounds like something you offer reluctantly. In Indonesia it's the thing a lot of buyers assume is available, and a shop that doesn't offer it looks like a shop that might not exist.

It converts, and it converts for a reason that has nothing to do with cards. The buyer keeps hold of the only thing they can enforce, their own money, until the box is in their hand. That's the whole logic, and why cash on delivery is popular in some markets and not others works through it properly.

The cost lands in three places. Refused parcels, which you pay freight on twice. The float, because your money sits with the courier for days after delivery. And the handling fee on collection.

Refusals are the one that hurts. The clearest published numbers on this come from India rather than Indonesia. GoKwik, an Indian checkout company, reports COD orders running near a 26% return-to-origin rate against under 2% on prepaid, with the industry average around 20 to 25% and fashion reaching 40% (their figures, read August 2026). Indonesian couriers don't publish an equivalent, so don't take that as your number. Take it as the shape of the problem, then ask your own courier for your own RTO rate, because they have it and most sellers never ask.

Two things bring refusals down more than anything else. Confirm the order by message before dispatch, with the total including ongkir written out in full. And put a rule on who gets COD. First-time buyer in your own city, fine. First-time buyer on another island for a Rp 400,000 order, ask for a transfer or a deposit.

Write your COD note in Indonesian even if the rest of your page is in English. The person who opens the door and reads it is often not the person who placed the order.

Ongkir decides more sales than price does

This is the part sellers underprice, in both senses.

A Rp 42,000 jar with Rp 30,000 ongkir to Makassar is a Rp 72,000 jar, and the buyer does that arithmetic in their head before you've finished typing. Hiding the ongkir until checkout doesn't help. It just moves the moment they leave.

Publish it. Zone by zone if you can, island by island at minimum. Jawa this much, luar Jawa this much, Papua ask us. Buyers respect a number they can see and resent a number that appears late.

The couriers most small sellers rotate between are JNE, J&T, SiCepat, Ninja Xpress and Anteraja, plus instant options through the ride apps inside a city. They're not interchangeable. Pick two, learn their actual pickup times and their actual failure modes, and stop switching every week for a Rp 3,000 saving that costs you an afternoon of tracking.

One structural thing worth knowing. You get charged on volumetric weight, which means a big light box costs more to send than a small heavy one, because you're paying for space in the van. A bulky Rp 60,000 item can carry Rp 35,000 of ongkir to Kalimantan and there is nothing you can do about it except pack smaller.

Worked example: Dewi, sambal, Yogyakarta

Dewi makes three chilli sambals in Yogyakarta. Rp 42,000 for a 200ml jar, Rp 115,000 for a box of three. She sells around 130 jars a month, mostly through Instagram and a TikTok account that occasionally does something unexpected.

Her setup after eighteen months of adjusting it:

• One link in bio to a page with all three sambals, both sizes, spice level written as a number out of five, and a line at the top saying "ready stock, kirim Senin sampai Jumat, cut-off jam 2 siang".
• Payment: bank transfer as the default, with her account name and number on the page and "tulis nomor order di berita transfer" in the instructions. Her DANA number is in the same instructions box for buyers who prefer it.
• COD: offered for Yogyakarta and Solo only, on orders up to Rp 200,000. Everything outside that is transfer first.
• Ongkir: Jawa Rp 12,000 to Rp 20,000, luar Jawa quoted per order, both written on the page.

Her month looks like roughly Rp 5,000,000 of goods. Her cost of taking payment is zero, because no processor is involved on transfers or COD. Her cost of confirming payment is about eight minutes a day of opening her banking app and matching credits to orders, and that's the real bill nobody prints on a pricing page.

The mistake she made in month five: she offered COD nationally after a video did well, took nineteen COD orders in a week from all over the country, and had six come back. She paid freight both ways on all six, two jars arrived home leaking, and the courier held the money from the successful thirteen for another eleven days. She now caps COD by city and by value, and her refusal rate went from something she was embarrassed to calculate to roughly one order in twenty.

Trust gets built before the money moves

Indonesian buyers check. Not aggressively, just steadily, and mostly for the same handful of signals.

• The account name on the transfer details matching the shop they think they're paying. A mismatch stops more transfers halfway than any other single thing.
• Photos of the actual product in an actual hand or an actual room, not a supplier catalogue image. Reverse image search takes four seconds and buyers do it.
• Testimoni. Screenshots of previous buyers saying it arrived. Unglamorous, slightly ugly on a page, and they work.
• A reply. Speed of first reply does more for conversion in DM-led markets than almost anything you can change about the product.
• A price written down. A shop that answers "how much" with "DM ya" reads as a shop pricing by how rich you look.

The full checklist is in proving you are a real seller, and it's worth an hour before you spend anything on ads.

Your own link, or Shopee and Tokopedia

Both, usually, and for different jobs.

Marketplaces bring you traffic you didn't earn and buyers who already trust the escrow. They also take a commission on every order, own the customer, and put you in a grid next to eleven people selling something similar for less. TikTok Shop sits inside Tokopedia now, after ByteDance took a controlling stake in Tokopedia in December 2023 and folded its Indonesian shop operations in, which is why live selling and the marketplace feel like one thing.

Your own link does the opposite job. No marketplace commission, full control of the page, the customer's phone number is yours, and nobody undercuts you in the next tile. But you have to bring the traffic yourself.

The split most sellers land on is marketplace for discovery and your own link for everyone who already follows you. Your bio link should go straight to products, not to a menu of other links. That argument is short and it's in why your link in bio is costing you sales.

Where Sailo fits, and where it stops

Sailo gives you the page. sailo.store/yourname, live the moment you sign up, with products, options, prices and stock, and an order button that opens WhatsApp with the item, the options, the address and the total already typed out. Bank transfer takes bank name, account name, account number, IBAN, SWIFT or BIC and a free-text instructions box. Cash on delivery takes a free-text delivery notes field. Sailo takes nothing at all on bank transfer, cash on delivery or chat orders, because on those it never touches the money.

Two limits you should know before you build around it.

The card button may simply not be available to you. Sailo's card rail runs on Stripe and only Stripe. On Stripe's own global availability page, read in August 2026, Indonesia is listed as "preview" rather than launched. Preview means invite, not sign-up. So no amount of upgrading to the Business plan at $49 a month will necessarily produce a card button for an Indonesian seller, and you should plan on transfer, QRIS-by-instructions and COD, which is what your buyers wanted anyway.

Nothing confirms your payments but you. Sailo can't see your bank account and can't see your DANA. The order sits at pending until you mark it paid. That eight minutes a day is yours, permanently, and it is the price of a manual rail, which takes no commission.

The free plan is $0 and caps at 10 products. For Dewi's three sambals in two sizes that's irrelevant. For someone listing 60 variants of hijab colours it's a wall you'll hit in week two.

What to do next

Tonight, write your ongkir as a list of real numbers by zone and put it where a buyer can read it without messaging you. Then write your COD rule in one sentence, in Indonesian, and stop making that decision per order.

Tomorrow, message your courier and ask for your return-to-origin rate for the last three months. Whatever number comes back is the one your COD policy should be built on, not a number from an article.

Then take the last ten DMs where someone asked a question and didn't buy, and answer every one of those questions on the page instead. If you want the wider comparison of what each payment rail costs and what it leaves on your desk afterwards, how to take payment as a small seller lays it out.