Altra Tech. Incorporated (益瑞電子股份有限公司) AEO Assessment Report
by Novastacks AI

Good news first: Altra Tech's pages are server-rendered and crawler-accessible — the homepage, blog article, and product pages all return their content in the initial HTML, so AI crawlers and search engines can read them without executing JavaScript. This is the single most important infrastructure prerequisite, and Altra clears it. There is no client-side-rendering wall here.

Two caveats temper this. First, the top navigation dropdowns (Products, Latest News, Knowledge Base) are wired with inert javascript:; links, which risks hiding the paths to deeper category and blog pages from crawlers that rely on real hyperlinks. Second, the automated speed audit could not capture a result: the homepage returned no painted content (a 'NO_FCP' outcome) on two consecutive attempts — the same outcome as GTA, while Sansho's homepage rendered cleanly. We therefore report homepage rendering speed as unavailable rather than estimating it, and recommend a manual page-speed verification. The 171 KB homepage weight suggests speed is worth checking.

The pages give machines weak structural cues. Altra's homepage, product page, and application page each have no H1 heading at all — the single strongest 'what is this page about' signal — relying instead on a lone H2. Only the blog article carries a proper H1. None of the four pages use semantic HTML landmarks (no <main>, <nav>, <article>, or <section> tags), so a machine cannot easily tell navigation from content from footer. Two of the four pages also ship with empty meta descriptions.

The contrast with GTA is instructive: GTA's pages use full semantic landmarks, breadcrumb trails, and descriptive titles and meta descriptions throughout. Sansho is weaker than both. The fix is squarely code-level — add one clear H1 per page, wrap content in semantic landmarks, and populate every meta description.

Altra Tech ships a single Organization profile on its homepage, product, and application pages — and nothing else. Critically, the blog article — the company's strongest content asset — carries zero structured data: no Article markup, no author, no publish date that a machine can parse. There is no WebSite markup (which unlocks the site-search box in Google), no Breadcrumb markup, no Product markup, and no FAQ markup anywhere on the site.

GTA, by comparison, implements Breadcrumb and Store markup on its product and application pages. The gap is most costly on the blog: structured Article data is exactly what lets AI attribute a technical claim back to Altra as the author. Without it, Altra's best content is published in a format machines treat as anonymous.

We tested Altra against its core category terms in Taiwan and the result is uniform: it holds zero ranked keywords in Taiwan organic search and appears in the top 20 for none of the three head commercial terms (electronic-materials supplier, lead-free solder-paste supplier, probe-pin supplier). Both competitors, though modest, at least rank — GTA for 26 terms (9 in the top 10), Sansho for 12. On the six AI prompt tests, Altra surfaced in organic results for exactly one query: its own brand name — and even there a recruitment directory outranks its homepage. This is a small-sample signal, but it points one direction consistently.

For the head category term in Taiwan, the sources AI cites most are financial and stock-data sites, Chinese Wikipedia, and social video — an investor lens, not a procurement lens. When a buyer asks AI to compare Altra, Sansho, and GTA, the answer is built from securities-research and stock-analysis sources treating them as tickers, because no vendor-authored procurement content exists to override that framing. Altra sells as a one-stop distributor solving sourcing problems; AI currently has nothing from Altra that frames it that way, so it defaults to the financial framing the web already supplies.

On every consumer platform AI draws from — YouTube, PTT, Dcard — the 'Altra' namespace is owned by Altra Running shoes, with videos drawing 12,000 to 109,000 views and active forum threads, while Altra Tech the electronics distributor returns zero relevant results. Altra's own YouTube channel does not surface even for its own brand-plus-industry query. This is namespace capture, not sentiment: the company cannot build organic discovery under its name on these surfaces without consistently pairing its brand with disambiguating context (益瑞電子 + electronic-materials). Its only clean-identity surfaces are LinkedIn and its Google Business Profile.

Altra is not starting from nothing — it holds advantages neither rival has. It is the only one of the three with a real editorial blog (named authors, two technical verticals), it has the deepest informational page content, it carries a Google knowledge panel with a 4.1/5 rating, and ChatGPT already describes it accurately when named. The category itself is wide open: the three distributors share zero ranked keywords with each other, meaning no competitor owns the head terms — they are held loosely by directories and marketplaces. A fragmented category plus an existing content engine is the most favorable starting position of the three.