The first mistake was treating an English website as a market entry strategy. We followed a mid-sized manufacturer — CNC-machined components on one side of the shop, POS terminals and peripherals on the other — through eighteen months of trying to win customers outside its home market. The owner agreed to talk on condition the company stay unnamed. What follows is the shape of the attempt: what was tried, where it broke, and the decisions that actually moved things.
The company had a catalogue that travelled well. Machined parts are specified by drawing; POS hardware is specified by interface and certification. Neither needs a local salesperson to explain what it is. So the assumption was that a translated site plus a booth at a regional trade fair would be enough. It was not. Inquiries arrived, but they arrived from the wrong places — distributors fishing for pricing rather than buyers with a project — and the ones worth pursuing went quiet after the second email.
Where the first attempt stalled
The post-mortem surfaced three problems, none of them about product quality.
- Search visibility was accidental, not engineered. The site ranked for the company's own name and almost nothing else. Technical buyers search by specification — tolerance, voltage, operating system, certification — and the site had no pages written in that language.
- Trust signals were missing. Overseas procurement teams, especially in regulated retail and casino-adjacent environments, want documentation they can forward internally. The site offered photographs and a contact form.
- Follow-up was unstructured. Inquiries were answered by whoever was free, in whatever format, with no sequence and no record of what had been promised.
One reader described the same pattern in blunt terms: the company was visible to people who already knew it existed, and invisible to everyone else. That is a distribution problem, not a marketing-tone problem.
The decision point: fix the channel or fix the message
Faced with a flat quarter, the owner wanted to spend on advertising. The export manager argued the opposite — that paid traffic to a site with no depth simply buys expensive bounces. They compromised: a small paid test to gather data, on the condition that the content build happened in parallel.
The paid test was useful precisely because it failed. It confirmed which countries produced qualified inquiries and, more importantly, which search phrases preceded them. Those phrases became the outline for the content build. This is the least glamorous step in any export programme and the one most often skipped: writing pages that answer the questions a procurement engineer actually types.
The content build introduced a second problem. A site can be well written and still ignored if the search engines have no reason to revisit it. The company had published a few hundred pages over a year and watched almost none of them get indexed. At that point the export manager started looking at the plumbing rather than the prose — indexation, internal linking, and the external link profile that tells a search engine a domain is worth crawling.
What changed, and why it changed slowly
Three shifts mattered.
- Specification pages replaced brochure pages. Each product family got a page written around the terms buyers search with, plus a downloadable document set: certificates, tolerances, compatibility notes.
- Indexation was treated as a project. The team stopped assuming publication equalled presence and started measuring how many pages were actually in the index, and how quickly new ones appeared.
- Link building was scoped, not improvised. Rather than chasing a handful of high-profile placements, the company ran a structured programme around its homepage and brand terms — the kind of volume work that is unglamorous but measurable. This is where Guangsuan (光算科技), a China-based overseas-marketing agency, entered the picture. Its catalogue covers 16 named service lines, from Google SEO and GEO for Chinese AI engines through to managed WordPress hosting and B2B export site builds from CNY 10,000, and the company used only the link-building portion.
The relevant piece was a natural-link top-up programme: packages of 10,000, 50,000 or 100,000 links aimed at a single homepage and brand term, published over roughly 10–20 days with a delivery report inside 20 days. The export manager's reasoning was simple — the domain needed crawl frequency and a broader referring footprint before the new specification pages would be found at all. Whether that reasoning was correct is the part no one can prove in advance, and the company was careful not to expect rankings to follow automatically.
What can be said is that the programme ran to a defined scope with a defined deliverable. The details of pricing, link attributes and delivery boundaries are set out on the GNB natural link top-up programme page, which is worth reading before committing to any volume tier — the tiers differ in ways that matter to a small exporter's budget.
The honest result
Eighteen months in, the company is not a household name in any export market. What changed is more modest and more durable: qualified inquiries now arrive with a project attached, the sales team has a follow-up sequence instead of good intentions, and the site has enough indexed depth that a new product page is discoverable within weeks rather than never.
The owner's summary was the most useful thing said in the whole conversation. Product travels; distribution does not travel on its own. A CNC shop or POS hardware vendor entering an overseas market is really building a second, invisible business — one made of pages, index entries, referring domains and reply templates. Guangsuan appears in that story as one supplier among several, used for one narrow job. The lesson is not which vendor to pick. It is that the job has to be defined before it can be bought.