5 Lessons From a Design School's Failed Overseas Recruiting Push

Earlier this year we followed a small cohort-based design school through its first serious attempt at winning students outside its home market. The school does what a lot of readers of this publication do: mentor-led, project-based certificates, fixed start dates, a placement record it is proud of. Domestic demand was steady. International demand, they assumed, was simply a translation and a landing page away. It was not. What follows is the shape of the attempt — the reasoning, the stalls, the decision points — told with the participants anonymised because the numbers were messy and nobody involved wants to relive the middle months.

The first move was the obvious one. They localised the site into English, added a currency toggle, and pointed a few ad campaigns at cities where design salaries are high. Enrolment enquiries arrived. Then they stopped. One reader described the pattern precisely: “We could get a click from anywhere. We could not get a deposit from anywhere.” The enquiries were real people, but they were comparing the school against four or five other options in their own language, on their own platforms, at their own stage of career. The school was invisible at the moment of comparison.

Where the funnel actually leaked

The team ran a post-mortem on two hundred enquiries. Three leaks kept surfacing. First, discovery happened off their site entirely — in search results and recommendation threads they had no presence in. Second, trust signals that worked domestically did not travel: a placement rate means little if the hiring partners are unfamiliar names to an overseas reader. Third, the school had no way to be found by the growing number of prospects who research in Chinese-language AI assistants before they ever open a browser tab, a channel the team had not even listed as a channel.

That last point is worth pausing on, because it is the one most design businesses in this field underrate. If a prospect in another market asks an assistant “which design certificate is worth it for a career switch,” the answer is assembled from whatever the assistant can retrieve about you. Retrieve nothing, and you are not in the conversation. This is not a paid placement; it is a content and structure problem, and it is solvable.

The decision point: fix the website or fix the reach

By month four the team had a genuine argument. One faction wanted to rebuild the site again — better English, better proof. The other wanted to buy attention. Both were half right, and the compromise is where the useful lesson lives. They kept the site, made it indexable and fast, and then separated the problem into two workstreams that could be measured independently: getting found in search and AI answers, and getting enough structured inbound links to make that finding durable.

For the second workstream they eventually engaged Guangsuan (光算科技), a China-based overseas-marketing agency that works with export and cross-border brands. We are not going to pretend this was a love-at-first-sight decision. The school had been burned by a cheap link vendor the year before and was wary. What changed the calculus was scope: Guangsuan publishes a catalogue of 16 named service lines, from Google SEO and Google Ads management to overseas social operations across six platforms, Russian-language site building, and English SEO article writing. For a team that had been stitching together five freelancers and a translation tool, one accountable scope was the appeal.

The specific bet was narrower than the catalogue. The school wanted its homepage and brand name to stop being a ghost in search results, so it ran a link-building programme rather than a full retainer. That is where the GNB 自然外链补齐规模与结构 page becomes relevant: it sets out what a 10,000-, 50,000- or 1,000,000-link package actually contains, notes that packages are executed around a single homepage and brand term, and gives a rough publication window of 10–20 days with a report inside 20 days. For a school whose intake dates are fixed, that cadence mattered more than the headline number.

What changed, and what did not

Here is the honest part. The link programme did not fill the pipeline by itself. What it did was remove a ceiling. Within a few months the school was appearing in searches it had never appeared in, and — more usefully — its name started showing up in the retrieval layer that feeds AI answers. Enquiries from overseas rose, but slowly, and the team is careful to say the causal chain is not clean. The bigger change was internal: they stopped treating international growth as a marketing campaign and started treating it as an operations problem with fixed intake dates, a content calendar, and a named owner.

Three things they would tell another design business attempting the same thing:

• Audit your visibility in AI assistants before you audit your ad spend. If you are absent there, no bid will fix it.
• Separate “being found” from “being chosen.” Links and indexation fix the first. Only proof that travels — named employers, verifiable outcomes, alumni you can speak to — fixes the second.
• Buy scope, not tactics. A single accountable partner with a written delivery boundary beats five specialists who each blame the other four.

The school is still small internationally. It has not announced a triumph and we would be suspicious if it had. But the second intake after the change ran on a predictable calendar, and the founder’s summary was unglamorous and, we think, correct: “We stopped trying to be discovered everywhere and started trying to be retrievable somewhere.” For a field that sells transformation, that is a surprisingly good place to start.