[{"data":1,"prerenderedAt":49},["ShallowReactive",2],{"article-en-dont-build-you-own-ec-platform-team":3},{"article":4,"related":18,"prevArticle":38,"nextArticle":42,"availableLocales":46},{"id":5,"slug":5,"title":6,"excerpt":7,"content":8,"image":9,"date":10,"published":11,"status":12,"category":13,"keywords":14,"meta_description":15,"og_image":16,"custom_css":16,"created_at":10,"updated_at":17},"dont-build-you-own-ec-platform-team","Why we don't recommend building or outsourcing your own e-commerce site","Building your own e-commerce site looks cheaper than handing a cut of every sale to a marketplace. This article explains why we still advise against it.","\u003Cp>The risks of building and running your own e-commerce site are probably much bigger than they look on the surface.\u003C/p>\u003Cp>\u003Cstrong>TL;DR;\u003C/strong>\u003C/p>\u003Cp>Lately quite a few companies that built their own e-commerce sites have asked us to assess and run acceptance testing on the whole platform, so that after outsourcing the work they end up with a site they can actually run reliably. The trouble is that by then the contract is often already in its acceptance phase, and many features and design decisions can no longer change. When we test in depth against the requirements, we find plenty of places where the system design simply does not match what was needed. Fixing them at that stage costs extra, which puts us in an awkward position. We can hardly tell a client that a lot of the money spent over the past months was wasted.\u003C/p>\u003Ch2>Outsourced quality is very hard to judge in advance\u003C/h2>\u003Cp>Put plainly, it is a vicious circle.\u003C/p>\u003Cp>Because clients usually cannot write a professional specification, contractors tend to add project management as part of the service.\u003C/p>\u003Cp>The contractor interviews the client, writes up the requirements and specification documents, has the client sign off, and then starts development. That is a perfectly reasonable service. The problem is that most clients find it very hard to tell how well a specification actually matches their needs.\u003C/p>\u003Cp>Even the boundaries of the requirements are hard to pin down. That is why so many systems hit one problem after another at go-live acceptance, and those problems are not in the spec, so fixing them means paying more for development and maintenance, or even giving up on the system altogether.\u003C/p>\u003Cp>For a project of some size, you can invite several companies to bid, or even have them help write the requirements document, much like the government procurement method in Taiwan known as \"restricted tendering after public selection or public solicitation\".\u003C/p>\u003Cp>This is a good approach, especially when the person running the procurement can tell which parts of a requirement description lock you in and which are open. The downside is that if that person is less familiar with the field, they may not spot the more obscure (or restrictive) requirements in the spec, and the project ends up being one only a particular vendor can compete for. That is not necessarily bad, but the risk is a bit higher.\u003C/p>\u003Cp>Sometimes a project goes badly not out of bad faith, but because the vendor was overconfident in its own system design ability.\u003C/p>\u003Ch2>Owning your own e-commerce site?\u003C/h2>\u003Cp>Back to e-commerce. In the early days of a business, the person responsible for the brand or sales channel usually does not care much about the details. Any platform will do, as long as things sell.\u003C/p>\u003Cp>As sales keep growing, you look at the financial statements and quickly notice a very steady take rate in your cost of sales that goes to the e-commerce platform. Platforms want a share of your revenue to cover their development costs. Because it is not a fixed cost, it hurts more the better you do. That is usually when you start wondering whether you could switch to a system with a fixed fee, so that higher sales do not mean higher costs.\u003C/p>\u003Cp>But unless you have deep pockets and fully understand the risks, do not jump in based on the surface difference. Neither a vendor's promises nor an eloquent technical expert will be tied to your sales.\u003C/p>\u003Cp>Almost no outsourcing vendor will put compensation for lost business in its SLA. If none of them will, how exactly are you buying peace of mind?\u003C/p>\u003Cp>So below we start with costs, then risks, and finally, if you really do want your own e-commerce site, the details to watch for when choosing a vendor.\u003C/p>\u003Ch2>What does a standalone e-commerce site actually cost?\u003C/h2>\u003Cp>The figures below exclude the cost of internal communication and launch. If you run the business, I suggest estimating that at 15~20% of the total project build cost. The real number may be a little lower, but not by much.\u003C/p>\u003Ch3>\u003Cstrong>Build costs\u003C/strong>\u003C/h3>\u003Cp>This is the cost everyone sees first. It falls into two categories: building your own tech team, or outsourcing development. Let's look at them separately.\u003C/p>\u003Ch3>Building your own tech team\u003C/h3>\u003Cp>This is usually the owner's favourite option, because the people are your own and it feels easier to control direction and progress.\u003C/p>\u003Cp>First you need a PM who understands basic e-commerce systems, ideally someone who has built products rather than run projects. \"Understands\" covers a wide range here. If you are hiring only one, I suggest not looking at anyone below about 1.5 million a year. Paying more does not guarantee a good result, but below that level your chances of finding the right PM are very low, and you are more likely to pay for your own headaches.\u003C/p>\u003Cp>Once you have a reliable PM, you need one or two fairly senior backend engineers, because most of an e-commerce system is backend work. Backend engineers who can design a reliable e-commerce system are hard to find in Taiwan. Below 1.8 million a year it is very difficult to find the right person, and better candidates generally cost 2.2–2.4 million. Even then they may not know e-commerce, only have the ability to design it. If you want development to move faster, hiring two is safer.\u003C/p>\u003Cp>Then there is the storefront. Someone has to build all those user interfaces.\u003C/p>\u003Cp>So we also need one or two frontend engineers. The spread in frontend skill is no smaller than in backend. If you want someone who can carry the frontend of the whole site, including design details, effects and tracking, a frontend engineer below 1.2 million a year is a matter of luck.\u003C/p>\u003Cp>In this situation I would usually suggest outsourcing the visual design rather than hiring your own designer. But if you want a decent user experience, it is best to find a studio or agency that can also handle UX. This is deep water, especially since e-commerce has many details and user paths, and it is not something a typical graphic designer can take on at the same time.\u003C/p>\u003Cp>Add it up, and before you have seen any results you have already spent about 8 million.\u003C/p>\u003Cp>Yes, and that is assuming you hired the right people. What owners run into more often is that they simply cannot tell who is right. So we will skip this option for now. If you really do have deep pockets and want to build your own team, come and talk to us; our Q4 coffee chats are about to start again.\u003C/p>\u003Ch3>Outsourcing development\u003C/h3>\u003Cp>Surely hiring a vendor is much simpler than hiring people yourself?\u003C/p>\u003Cp>Honestly, it is less hassle. Most vendors in the Taiwanese market genuinely do their best and are willing to think from the client's point of view.\u003C/p>\u003Cp>In terms of business model, you will meet two kinds of vendor.\u003C/p>\u003Cp>The first has never built an e-commerce system. They see that most e-commerce systems on the market take a cut of sales and want to get into that market too, so you get a lower quote. Their goal is simply not to lose money; they will then adapt the system they built for you and sell it to the next client.\u003C/p>\u003Cp>The second already has an e-commerce system. They only need to adjust it to your requirements rather than build from scratch.\u003C/p>\u003Cp>They sound similar, so why separate them?\u003C/p>\u003Cp>With the first kind, be aware that e-commerce systems are not simple, at least not as simple as they look. If you have never built one, it is hard to piece together how to do it. So most of the time, hiring this kind of vendor means paying to train their staff, and you become the first bite of the e-commerce platform pie the vendor has promised its employees. There are exceptions: a team that has designed high-traffic systems is well suited to make the switch. But in Taiwan, apart from exchanges of various kinds and ad tech companies, it is hard to find systems that handle more traffic than e-commerce. This kind of vendor will charge anywhere from 1.5 million to 6 million depending on your requirements. I usually suggest estimating from revenue: if you expect online revenue of under 2 million / 2–20 million / over 20 million, the first round of development will cost roughly 1.5–1.8 million / 2–4 million / over 3.5 million. These are only the basic development fees.\u003C/p>\u003Cp>The second kind of vendor has already been through a few client launches, so it will certainly be better than the first. But a system of this size, after several rounds of changes, usually has fairly high coupling. In plain terms, it is hard to change. Not impossible, but each change takes longer and is more likely to break something. That said, if this vendor's earlier clients were in the same line of business as you, you can be more at ease. The price will not be lower than the first kind, though, and is usually a bit higher. The early clients were training rounds; now that things are more stable, the vendor will naturally want to earn more.\u003C/p>\u003Cp>Does that feel much cheaper than building your own team? Don't decide yet. There are costs we have not counted.\u003C/p>\u003Ch3>\u003Cstrong>Maintenance costs\u003C/strong>\u003C/h3>\u003Cp>If you hire your own people, they can generally handle both maintenance and new features, so the second year costs about the same. But you will have to give raises. A small company that does not pay a bit more will see people leave for somewhere more attractive. Let's be conservative and call it 16 million for the first two years.\u003C/p>\u003Cp>If you outsource, maintenance is usually charged at 15-25% of the original scope. But e-commerce needs to keep up with and integrate quite a few new systems, and every new feature is paid development. That makes 1.5–2.5 million a year a very basic expense, so the first two years will also cost over 6 million.\u003C/p>\u003Ch3>\u003Cstrong>Operating costs\u003C/strong>\u003C/h3>\u003Cp>An e-commerce platform has to run on some machines or cloud servers. 20,000–30,000 a month is a very frugal setup, and hundreds of thousands a month is not unusual. When a system is poorly designed, you can spend a lot every month just on running it. That money goes out quietly, you cannot avoid spending it, and it is hard for anyone outside to review or audit. This is deep water too.\u003C/p>\u003Ch3>\u003Cstrong>Transaction costs\u003C/strong>\u003C/h3>\u003Cp>This does not mean the fees paid to payment channels, but the vendor's cut of sales. Vendors that build these e-commerce systems usually want to recover their development staff costs through a percentage of transactions. If they do not take a cut, expect to add another 30-50% of the original development fee on top of the vendor quote above. So spending over 8 million in the first two years is basically unavoidable.\u003C/p>\u003Ch3>Why so expensive? There are cheaper options on the market!\u003C/h3>\u003Cp>Cheap does not mean bad, but you may not be able to tell the difference. Leaving aside whether a cheap option meets your goals, the risk is almost certainly higher. So let's move on to the risks.\u003C/p>\u003Ch2>The risks of building or outsourcing an e-commerce site\u003C/h2>\u003Ch3>\u003Cstrong>Technical risks\u003C/strong>\u003C/h3>\u003Cp>On the surface an e-commerce system is just product listings, a shopping cart and checkout, but behind that is a highly complex piece of engineering. When it is handled badly, it shows up directly in revenue and user experience.\u003C/p>\u003Cul>\u003Cli>\u003Cp>\u003Cstrong>Traffic spikes: \u003C/strong>Taiwanese online shoppers are long used to flash sales and limited-time promotions. Without mature load balancing and caching, a site can easily go down at peak times and lose a large batch of orders.\u003C/p>\u003C/li>\u003Cli>\u003Cp>\u003Cstrong>System coupling: \u003C/strong>Self-built or customised systems often pile up technical debt right after the first version goes live. Adding a feature then touches everything, and the cost of change multiplies.\u003C/p>\u003C/li>\u003Cli>\u003Cp>\u003Cstrong>Third-party integrations: \u003C/strong>Payment, logistics, customer service and ad-tracking APIs are constantly updated. A platform without long-term operations and the ability to fix things quickly can easily break at a critical moment.\u003C/p>\u003C/li>\u003Cli>\u003Cp>\u003Cstrong>Testing and quality control: \u003C/strong>An e-commerce site has to cover hundreds of transaction and exception flows. Without rigorous testing, any small change can cause wrong orders or failed checkouts.\u003C/p>\u003C/li>\u003C/ul>\u003Cp>For a brand, the question with these problems is usually not \"can it be solved\" but \u003Cstrong>\"how many people and how much time will it take to solve\"\u003C/strong>. The hardest part is having the budget and still not finding the right people.\u003C/p>\u003Ch3>\u003Cstrong>Security risks\u003C/strong>\u003C/h3>\u003Cp>The most sensitive asset in running an e-commerce site is \u003Cstrong>customer data\u003C/strong>: member personal information, order details and payment records. If this data leaks, you not only violate Taiwan's Personal Data Protection Act, you also seriously damage trust in your brand, a reputational loss you cannot undo.\u003C/p>\u003Cul>\u003Cli>\u003Cp>\u003Cstrong>Payment security: \u003C/strong> Credit card transactions must comply with international standards such as PCI DSS. Self-built sites often lack a proper audit, and any small hole can be exploited by attackers. But don't worry, most developers will tell you the site is secure.\u003C/p>\u003C/li>\u003Cli>\u003Cp>\u003Cstrong>Data protection: \u003C/strong> Member login, password management and purchase history need strong encryption and access control. Otherwise, once data is stolen, you may face a class action.\u003C/p>\u003C/li>\u003Cli>\u003Cp>\u003Cstrong>Continuous monitoring: \u003C/strong> Large platforms have dedicated security teams monitoring unusual traffic and attacks 24 hours a day. A brand's own site usually does not have those resources, and by the time a problem is found it is often too late.\u003C/p>\u003C/li>\u003Cli>\u003Cp>\u003Cstrong>Who carries the liability: \u003C/strong> Most outsourcing vendors only guarantee that the system works. They will not cover business losses caused by security holes. In other words, the risk ends up with the brand.\u003C/p>\u003C/li>\u003C/ul>\u003Cp>By comparison, running transactions on a mature platform or using a SaaS tool with international certifications spreads and reduces security risk more effectively than a self-built site. Also, ISO 27001 is only the entry level. Having it does not mean you are more secure, but not having it certainly means more risk.\u003C/p>\u003Ch3>\u003Cstrong>Operational risks\u003C/strong>\u003C/h3>\u003Cp>Even with stable technology and solid security, operations can still be the biggest variable. Many brands only discover after launch that the cost of running the site is much higher than they expected.\u003C/p>\u003Cul>\u003Cli>\u003Cp>\u003Cstrong>Dependence on marketing: \u003C/strong>A self-built site has to bring in its own traffic. Organic traffic needs planning and ad spend to stay steady, and without an advertising budget or social media work, sales are hard to grow.\u003C/p>\u003C/li>\u003Cli>\u003Cp>\u003Cstrong>Payments and returns: \u003C/strong> Without mature reconciliation and refund systems, customer disputes and financial risk follow easily. This is usually the weak spot of an e-commerce system; the longer it has been running and the more pitfalls it has hit, the more mature it gets.\u003C/p>\u003C/li>\u003Cli>\u003Cp>\u003Cstrong>Customer service and after-sales: \u003C/strong> As order volume rises, the load on customer service and warehousing multiplies. If the processes are badly designed, the overall experience drops fast. You really do not want to know what it is like when the system flow was not thought through and the only fix is a room full of people covering the gaps.\u003C/p>\u003C/li>\u003Cli>\u003Cp>\u003Cstrong>Regulation and tax: \u003C/strong> A self-built system has to meet requirements such as Taiwan's e-invoice rules, personal data protection and cross-border trade regulations. These are hidden costs that are hard to estimate.\u003C/p>\u003C/li>\u003C/ul>\u003Cp>The core of operational risk is scaling and sustained investment. Many brands run smoothly at first, but once sales grow and orders multiply, the original processes and tools become the constraint.\u003C/p>\u003Ch2>Conclusion: you can build it yourself, but know what game you are playing\u003C/h2>\u003Cp>Building your own e-commerce site is not hard, but you need to understand the cost and responsibility involved.\u003C/p>\u003Cp>When you have to carry the technical, security and operational risks yourself, the cost is no longer just \"website development fees\". It becomes a long-term commitment of investment and management.\u003C/p>\u003Cp>If you want your brand to focus on products and marketing rather than chasing bugs or patching security holes, a mature SaaS e-commerce platform or cloud architecture service is the more stable and sustainable choice.\u003C/p>\u003Cp>\u003Cstrong>Finally, one reminder for those of you who read this far: if you really do need to build your own e-commerce system, bring in a third party other than the developer for joint acceptance testing. Not because developers are unreliable, but because there are always blind spots. Once acceptance is signed off, any further change costs real money.\u003C/strong>\u003C/p>","https://cdn.sqa.tw/articles/dont-build-you-own-ec-platform.png","2025-10-30T06:44:00+00:00",true,"published","industry","build your own e-commerce site, e-commerce development, outsourced development risks, e-commerce setup costs, e-commerce security risks, SaaS e-commerce platform, outsourced web development, e-commerce maintenance costs, brand website development","Building your own e-commerce site is more than a tech project. From our acceptance testing work, the cost, security and operational traps companies overlook.",null,"2025-11-13T18:12:13.321+00:00",[19,26,32],{"id":20,"slug":20,"title":21,"excerpt":22,"image":23,"date":24,"category":25},"ai-development-sdlc-release-verification","When AI builds it in two hours but verification still takes three days","Once AI speeds up development, spec sign-off, testing and release do not automatically speed up with it. Using a real case from a booking system SaaS, we describe how we adjusted the SDLC to cut full release verification from two or three days to under two hours, and how the SQA role changed along the way.","https://cdn.sqa.tw/articles/ai-sdlc-verification-bottleneck-cover.jpg","2026-11-05T12:00:00+08:00","perspective",{"id":27,"slug":27,"title":28,"excerpt":29,"image":30,"date":31,"category":25},"sme-guide-to-choosing-ai-agent","AI agents are everywhere. How should a small business choose one?","We were recently asked to run a full post-deployment review of an AI agent, and found it had quietly caused five times too much raw material to be ordered from customer order emails, and nearly created a liquidity risk on the finance side. The problem with agents is rarely that they are not smart enough. It is that nobody asked the right questions.","https://cdn.sqa.tw/articles/sme-agent-black-box-inspection.jpg","2026-09-01T15:00:00+00:00",{"id":33,"slug":33,"title":34,"excerpt":35,"image":36,"date":37,"category":25},"ai-coding-team-collaboration-three-patterns","Three patterns we see once AI coding reaches the team","After working with more than ten development teams over the past six months, we keep seeing the same thing: individuals getting faster with AI does not mean the team is ready to get faster together.","/images/blog/ai-coding-teams-editorial.jpg","2026-08-20T08:00:00+00:00",{"slug":39,"title":40,"date":41},"simply-check-product-quality","A simple way to check the quality of a project or product","2025-09-20T09:18:00+00:00",{"slug":43,"title":44,"date":45},"ai-or-not-the-developers-survival-line","Is using AI the line between development teams that survive and those that don't?","2025-10-30T16:49:00+00:00",[47,48],"en","zh",1794970113066]