- For a brand with regular customers, the low-risk path is to sell in both places: let Shopee and Lazada find new buyers, and let your website look after the people who buy again.
- Your own site gives you a customer list you can contact yourself and the freedom to set your own prices and promotions. It gives you no passing traffic, though. Customers have to know the brand before they visit.
- Start when you see at least three of the six signals, such as repeat buyers, people typing your brand name into search, or shops messaging you on LINE to order in bulk.
- Open the site once stock matches across every channel. A regular customer whose web order gets cancelled because Shopee sold the last unit first won't be quick to try a second time.
What a marketplace gives you, and what the platform keeps
Selling on Shopee or Lazada is a bit like renting a booth in a big shopping mall. Crowds walk past all day, and the mall runs the campaigns, takes payment for you and arranges delivery. But the mall sets the rules and decides which booth gets the good spot. Your brand's website is your own shophouse. Far fewer people walk by, but you know every regular, and you set the rules for the whole shop.
What the platform really gives you is buyers who don't know you yet. They type "sunscreen," find your product next to other shops, and feel safe buying because they trust the platform's payments and returns.
What the platform keeps is the relationship with the customer. You see only as much buyer data as the platform chooses to share, search ranking depends on systems you can't control, and the platform's fees and commissions change with its policies. This starts to hurt when the same customer comes back for a third purchase and you still have to pay for ads so they see your product before a competitor's.
What your own website gives you, and what doesn't come with it
A brand website gives you things that build into an asset for the shop. The first is customer data: who bought what, when, and how many weeks until they buy again. That data sits in your own system under the consent your customers gave, so you can message people who are nearly due to reorder through LINE or email yourself, instead of paying for ads to reach the same people again.
The second is control of the brand. Product pages, gift sets and member prices work the way you design them. Sales on the website don't share a commission with a platform, though you still pay payment fees and shipping as before. And on the day your marketplace shop loses visibility or campaign terms change, your regulars can still find you.
What doesn't come with it is passing traffic. Customers have to come from channels you build yourself, such as a LINE Official Account (LINE OA), a Facebook page, people searching your brand name on Google, or ads you pay for. Work the platform used to do for you, like answering chats, confirming payments and issuing refunds, comes back to the shop. What the website costs depends on what it has to connect to, and the factors that move the price are in what decides the cost of a website or app.
| Area | On a marketplace | On your brand website | To prepare if you build a site |
|---|---|---|---|
| New customers | Come from platform search and campaigns | You bring them in from LINE, Facebook, Google or ads | Channels where people already follow you |
| Customer data | You see what the platform shares | Held in your system, used under consent | A consent page and a privacy policy |
| Prices and promotions | Follow the platform's rules and campaigns | You set member prices, bundles and coupons | Someone who sets the discount ceiling |
| Cost per order | Platform fees and commission | Payment fees, system costs and the cost of bringing visitors in | Per-channel cost figures you can compare |
Six readiness signals before you open a store on your own domain
Answer these six with reports your shop already has, one point each, and look back at least six months. Numbers from one big campaign month are easy to misread.
Six readiness signals
- Signal 1: repeat buyers. Download the order reports from every channel's seller center and count the buyers whose account name appears more than once. Products that get used up, like creams or coffee, usually show this signal first.
- Signal 2: people search for your brand. They type your brand name rather than the product category. Check the platform's search-term report if it has one, or Google Search Console.
- Signal 3: people message you directly. Your LINE OA or Facebook page regularly gets questions about products or requests to order direct. These people are the website's first customers.
- Signal 4: you want to do things the platform doesn't allow. Member prices, brand-only bundles, or a delivery that arrives every month.
- Signal 5: wholesale or business orders. Shops or spas order in bulk over LINE and want volume pricing or tax invoices. Orders like these don't sit comfortably on a marketplace.
- Signal 6: stock across channels is slipping out of control. You oversell during campaigns, or someone spends every evening correcting stock numbers in several seller centers. This one says you need a single stock record first, whether or not you build a website.
A score of 0 to 2 means keep selling on marketplaces and use the time to build LINE followers and collect reviews. With 3 to 4, open a small web store, starting with best sellers and customers who already know you. With 5 to 6, make the website the brand's main home and use marketplaces to find new customers.
Before you add up the score, ask one more question: who will answer chats and pack web orders as fast as marketplace orders? A perfect six won't help if the site opens with nobody minding it. Find that person before you look for a developer.
Sell in both places, and shift in stages
Selling in two places at once works when each has a clear job. The marketplace finds new customers and the website looks after people who have already bought. You move the weight gradually as the signals change, without closing either shop.
Stage one: keep selling on marketplaces and build channels of your own. Open a LINE OA where customers can ask how to use your products. Be careful here, because some platforms forbid steering buyers to purchase off-platform, both in chat and inside the parcel. Read each platform's terms before you put a QR code or a website card in the box.
Stage two: open a web store on your own domain, connected to Shopee and Lazada. Start with a handful of best sellers, a membership system and one stock record shared with every channel. Prices on the website don't need to be lower. Offer member benefits instead, such as points or a coupon for the next order, so your two channels aren't undercutting each other. Before you talk to developers, go through what the store needs with the 12-part checklist for hiring an e-commerce developer.
Stage three: make the website the brand's main home. Add an ordering page for wholesale customers, and product pages Google can find when people search for the product name. Marketplaces keep bringing in new customers. Two numbers tell you when to move up a stage: the share of returning customers' orders that have moved to the website, and each channel's cost per order, side by side.
Hypothetical case
A herbal soap and shampoo brand from Chiang Mai (name and details invented) has sold on Shopee and Lazada for three years, and its margin per item gets thinner every year. The owner counts the six signals in the last six months of reports. About a quarter of buyers bought more than once, mostly shampoo refills. The LINE OA gets questions about delivery dates every day, three spas order several cartons at a time over LINE, and during the year-end campaign the shop oversold twice. That makes five. The one that fails is brand search, which is still low.
The owner opens a small website with refills, gift sets and an ordering page for spas, sharing one stock record with both platforms, and tells people who message on LINE that refills ordered on the website earn points. For the first three months, almost all new customers still come from marketplaces, while refill orders from existing customers move to the website a little more each month. The surprise is chat on the website. Existing customers expect replies as fast as on LINE, so one admin now handles web orders and nothing else.
Which shops shouldn't build an online store yet
Waiting can be a good decision too. If your shop matches any of the points below, do the missing work first.
- The brand is new, almost every customer buys once, and nobody searches for the brand name yet.
- Your products compete on price alone, and customers open the platform to find the cheapest shop.
- You expect the website to find new customers on its own, with no ad budget and nobody making content.
- Nobody on the team can look after chats, orders and products on the website.
- Stock is still counted by hand, and nobody can say how much is really in the warehouse.
While they wait, these shops can build a following and get their stock into a proper system. When the score gets there, both are ready to use straight away.
For the development team · Stock and orders across channels
Keep stock in one place as the source of truth, and have every channel read from it and deduct against it. Connect to each marketplace through its official open API, receive orders by webhook or by polling, whichever the platform supports, and respect each platform's rate limits. Decide clearly whether stock is reserved when the order is created or when it is paid, and release it when an order times out unpaid. Use the platform's order number to avoid importing the same order twice, hold safety stock for campaigns, and reconcile daily against each platform's reports. Buyer data from a marketplace may be used only under that platform's terms, so store it separately from the members who sign up on the website.
Open a web store without leaving Shopee and Lazada
The people who know how your shop works are your own team: the owner who sets prices, the admin who answers chats, the packers, and the accountant who closes the day's books. DNA Maker starts by talking with them, then traces one real order from each channel: where the money lands, when stock is deducted, and who confirms what. We write that down as a workflow, stock-deduction rules and a screen for each person. If your six signals aren't there yet, we'll tell you plainly to wait.
For brands that are ready, our e-commerce website development service builds a store on your own domain that takes cards, PromptPay, mobile banking and installments. It sends orders to the courier you already use with tracking numbers generated automatically, shares one stock record with Shopee, Lazada and TikTok Shop, and lets your marketing team message past buyers through LINE or email themselves. All the work is done by DNA Maker's own team in Bangkok, with no subcontracting. For other kinds of websites, see our website development page, or count your six signals from the shop's reports first and bring the score to talk it through with us.
SOFTWARE ENGINEERING GLOSSARY
Software engineering glossary
| Term | What it is | A simple example | What executives should ask the development team |
|---|---|---|---|
| Omnichannel | Selling across several channels where orders, stock and customers connect into one picture | You sell on Shopee, Lazada and your website, and the team sees every channel's sales on one screen | Which channel still needs data keyed from one system into another by hand? |
| First-party Data | Customer data the brand collects directly, with the customer's consent | Members who sign up on the website and tick the box to get news on LINE | Where is our customer data stored, and who can access it? |
| SKU | The code for one sellable item, split by size, color or scent | A pump bottle and a refill of the same shampoo are different SKUs | Does every channel use the same set of SKU codes? |
| Source of Truth | The single place that holds the real version of the data, which other systems follow | The stock count in the central system that Shopee, Lazada and the website all deduct from | If two systems disagree, which number do we trust? |
| CRM | A system that keeps customer history and contacts, used to pick groups of customers to message | You can see who last bought a refill, and when | Can the marketing team pick customer groups themselves, without asking anyone for a file? |