Skip to main content
WANFENG TECHNOLOGY
HomeServicesBenchmark SolutionsInsightsContact
☀☾Dark
▾
  1. Home
  2. /Insights
  3. /Export site build vs buy: the first-year quote …
←Back to WANFENG Insights

Engineering Delivery

WANFENG INSIGHT — Traceable narrative, engineering receipts.

Nov 21, 2026

Export site build vs buy: the first-year quote lies

Build-vs-buy for export sites usually comes down to the first-year quote. A three-year TCO table, three tests to run, and when custom is the wrong call.

Picking an export site: the first-year quote shouldn't be the deciding number

One quote reads 'full site build, ¥38,000, eight pages included'. Another reads 'priced per scope, item list in attachment one'. Put both on the same approval form, compare first-year totals, and the first almost always wins.

They aren't selling the same thing. One sells a package; the other sells hours against a defined scope. Equal totals don't mean equal content, and a higher total doesn't mean bad value.

The October 2026 round-up of export-site vendors scored on price bands, launch speed and who each vendor suits. Useful as context, useless as a decision rule. A site isn't a car you buy once; it's a system someone will keep changing.

What each camp is actually selling

Templates plus a buyout licence: pay the licence or build fee once, then annually for hosting, plugins and theme upgrades. Launch is measured in weeks, and marketers swap images without a developer. For trading companies with a stable catalogue and enquiries mostly from trade shows and organic search, that covers the job, and costs are predictable.

Scope-based custom: data sits in your own database, integrations are written against your ERP and payment provider, service levels go into contract clauses, changes come through a signed change order instead of a platform roadmap. The cost is higher upfront spend, a longer timeline, and someone internal who keeps raising requirements.

Neither side is wrong. They assume different companies: one assumes stable requirements, the other assumes the business will move.

Why first-year totals don't compare

Different granularity. A package quote bundles page caps, product caps and language caps, with overages priced by tier in a corner of the contract annex. A scope-based quote splits page types, template count, field structure, integrations and migrated content items. Comparing the totals is comparing a wholesale price with an hourly rate.

Different structure after year one. Buyout sounds like a single payment; every year you still pay hosting or cloud, CDN, SSL, domain renewal, theme licence, feature subscriptions, gateway fees. Package quotes bury this as 'third-party fees not included'. Custom quotes often omit it too.

Labour is hidden. Both options need someone for product photos, multilingual copy, enquiries and analytics. That spend doesn't depend on the option chosen, but it's the line most often missing from the budget sheet.

Switching costs are asymmetric. Template lock-in shows up late: plugins bind you to a platform, exports come out incomplete, changing vendor is close to a rebuild. Custom locks you into code and data structures, but with source code and accounts in your hands another team can take over.

Filling in a three-year TCO

TCO isn't the annual fee times three. Run this table for three years and again for five, fill both vendors with the same rows, and write 'not included' rather than estimating zero.

| Cost line | Template + buyout | Scope-based custom | | Build and licence | Package fee, tier overage | Hours against the scope list | | Fixed annual | Hosting, CDN, SSL, domain, theme licence | Server or cloud, domain, certificates | | Plugins and extensions | Per-feature subscriptions, priced by site or order volume | Inside the dev scope, later changes billed | | Transactions | Platform commission plus gateway fees | Gateway fees, FX loss, chargebacks | | Integration | Via plugins, quality depends on the plugin author | Billed by interface hours, licences extra | | Content and localisation | Internal staff or outsourced translation | Same | | Migration and downtime | Export and rebuild hours when switching platform | Handover hours when switching developer | | Internal owner | Marketing doubles up | Someone must own requirements and acceptance |

The two rows most often underestimated are plugins and integrations. On the template side, whether a feature works depends on whether anyone built a plugin, whether it's still maintained, and whether it's billed by traffic or order volume. On the custom side, it depends on whether the other system exposes an API, how licensing is charged, and how far field definitions diverge — frequently discovered after signing.

Three years and five years can produce opposite conclusions. Custom's upfront spend amortises in year three; template annual fees compound into a real number by year five. Put both figures side by side.

Downtime and SEO recovery is almost never filled in. Changing platform or vendor means 301 mapping, re-indexing and lost backlinks take months to settle. No quote carries that loss, and it hits B2B sites that live on organic search.

Test one: is the requirement standard?

Standard means requirements cluster around the catalogue, certifications and enquiry forms, with changes mostly to content.

A cheap screen: the last twelve months of product and page change records, or marketing's email thread if no log exists. If most changes are swapping photos, adding models, editing spec tables, updating certificates, adjusting prices, the requirement is standard, and a template handles them at close to zero marginal cost.

If these keep recurring, it isn't standard: prices by customer tier, different certifications or model names per market, the same product with different specs per region, enquiries routed by product line, live stock on product pages. Those are data structure problems, not content problems. Templates stack plugins layer by layer, and by the third layer one change breaks another.

No percentage is needed. Look for whether one class of structural requirement recurs.

Test two: ERP, payments and logistics

Answer this with 'is anyone moving data by hand', not 'do we want it'.

If someone copies marketplace orders into the ERP daily, edits stock levels onto the site by hand, or pastes tracking numbers one by one, that labour is your quantified baseline. Convert it to an annual cost and set it against the integration hours in the custom quote.

Usual integration points: product and stock sync, order write-back, customers and enquiries into CRM, tracking visibility, multi-currency and tax, local payment coverage.

One precondition gets treated as a deliverable: whether the other system has a usable API. Some ERPs charge licence or per-call fees, some only offer scheduled export files, some have good docs and a flaky sandbox. Do the feasibility study before quoting; its conclusion decides whether integration is work or uncontrollable risk. [TO VERIFY: API licensing and billing terms of the target ERP vendor — this drives the integration hour estimate directly.]

Payments sit in the same category. Cross-border collection is charged as a percentage, so fee structure, FX loss, chargebacks and settlement cycle belong in the transaction row. Platform payments are convenient, but rates and supported countries are the platform's call; third-party gateways are flexible, and you handle chargebacks and fraud yourself. That difference lands on each order, not on the build quote.

Test three: will the business keep changing?

Evolution is readable from signals you have now: SKU count over three years, number of target markets, B2B and B2C running in parallel, member tiers and quote approvals.

Four or more languages, wholesale and retail side by side, pricing by tier — the data structure will change. On a template that means plugin sprawl and page logic fighting itself; on custom it means paying hours for every change.

A more useful line than 'does the page look good': content changes, template; data structure changes, custom.

Counter-signal: if the main category might shift within a year, the three-year premise fails. Then the priority is minimising migration cost — data that exports, accounts in your name, URLs not welded to one theme's routing — not rushing into custom.

Combining the three tests

Test one true, two and three false: template plus buyout, with the budget going into product photography, multilingual copy, technical SEO and content. The extra money custom costs buys flexibility you won't use.

Test two true: custom. Not because pages look better, but because data definitions and integrations need writing correctly once. Manual data movement scales with order volume; an integration is a one-time spend.

Test three true: custom, with two things accepted upfront — an internal owner and an annual change budget. Without both, custom becomes a project that costs more every year.

Two and three true, budget short: an overlooked middle path. Template front end for content and enquiries, integration layer built as a separate middle service on your own server. Swap the front end once direction is clear; the integration layer survives.

When custom is the wrong call

Scope-based custom assumes the scope can be written down. Three situations break that: nobody can keep raising requirements, requirements aren't thought through, and there are two months to launch.

In those cases the project drifts into the vendor's interpretation, the two sides define 'done' differently at acceptance, and rework never ends. A template is steadier there, because the platform sets the boundary.

So the test isn't 'custom beats template'. It's 'can your requirements be written as a list?'

Scope-based pricing is only comparable at line-item level

'Product listing page' isn't a scope. 'Three-level categories, filter by certification, currency switch by market, 24 per page, URL pattern /products/, single column on mobile' is. 'Connect ERP' isn't either; name the interface, direction, fields, frequency and retry behaviour.

If the list is too coarse the quotes can't be compared: one vendor under-counted three templates, another over-counted two interfaces, and the gap you're staring at is fiction.

Practical move: have both options quote the same list, the template vendor adding an included/not-included table and the custom vendor an hour breakdown. Put the three documents side by side.

Four questions for the negotiation

Template side: how overages are billed, whether theme and plugin licences travel with the site, how many content items migration covers, and what a second language costs in year two.

Custom side: how many scope hours are integration, whether a change order is priced by hour or function point, which repositories the code handover includes, and how handover works when maintenance ends.

Last thing

One action: spend two days writing the scope list before you talk price. Once it's written, much of the argument dissolves, because you'll find you need fewer features than you assumed.

A question I can't answer: if the company might be acquired within a year, or the main category might change, test three stops working. Then the question isn't which option to pick, it's 'what do I need to walk away with in three years' — data, domain, content, URL history. That answer usually decides the choice better than the TCO table.

Sharing & attribution

Originally from WANFENG TECHNOLOGY, published in WANFENG Insights. Feel free to share—please link back to this page and credit WANFENG TECHNOLOGY.

You may also like

  • Engineering Delivery · Nov 17, 2026

    Cross-border site selection: nail six contract clauses

    Read →

  • Technical Practices · Nov 14, 2026

    Custom Mini Program Storefront: Check Data Before Features

    Read →

  • Technical Practices · Nov 10, 2026

    AI integration: pick the embedding layer before the model

    Read →

Unsure what you need?

Talk 1:1 with a senior advisor. We will clarify your goals and return a plan with a ballpark quote—free of charge.

Book a free scoping call

No hard selling. Pure technical and business discussion.

WANFENG TECHNOLOGY

Custom development, technical SEO, UI/UX motion, and private LLM programs in one delivery chain. 10+ years of engineering and growth for domestic and global scenarios.

On this page

  • Home
  • Insights
  • Benchmark Solutions
  • Contact
  • RSS feed

Business inquiries

Reach us by email, WeCom, or your usual business channel.

© 2026 WANFENG TECHNOLOGY. All rights reserved.

粤ICP备2020119225号