San Jose SEO for buyers who search in part numbers.

The three searches from the table above, counted. Each bar is page one; each color is who owns the page.

Counted by hand on October 24, 2026, US results. A part-number search is owned by resellers because they publish a page per part and the maker publishes a PDF. A certified-service search is owned by the local shops because they wrote the certification page. An “alternative” search is owned by forums because nobody else wrote the table. The pattern is the same in all three: the page that exists holds the slot.

Half your buyers are 8,000 miles away.

Valley hardware queries come from Taipei, Penang, and Munich as much as from down the 101. We split the architecture: spec pages rank globally, service pages hold the San Jose local pack.

“0.5mm pitch ffc connector 30 pin,” searched from the United States, Germany and Taiwan. Same distributors on page one. No maker anywhere.

We ran the same part query through Google’s US, German and Taiwanese results on October 24, 2026 and listed every domain on page one. A filled dot means the domain holds a slot in that country. The last row is the point: the companies that actually make the part — their names are printed on every listing — do not hold an organic slot in any of the three. Their part lives on a distributor’s page in Mansfield, Texas, or a marketplace’s in Shenzhen.

That is the global-intent half of this page in one table. A spec page written once, with a Mouser-grade field set and an hreflang pair per market, enters a page one that is the same in every country — because the engineers in Hsinchu and Munich type the same part number as the ones in San Jose.

METHOD: GOOGLE.COM WITH GL=US, GL=DE AND GL=TW, ENGLISH INTERFACE, PERSONALIZATION OFF; ORGANIC RESULTS AND DATASHEET PDFs COUNTED, SHOPPING UNITS AND IMAGES NOT. DOMAINS GROUPED BY COMPANY ACROSS COUNTRY TLDs.

Priced per indexed page, like a part. Google counts them; the count is the invoice.

A catalog is bought by the thousand, so the work is priced by the thousand. The unit is a part page that Google has listed as indexed in your Search Console — not a page we generated, not a page we submitted. If it is not in the index, it is not on the invoice.

Our catalog is 30,000 SKUs in PDFs. Realistic?

Yes — that's an extraction project, not a rewrite. We generate structured product pages from your existing data and let the long tail do the ranking.

Distributors outrank us for our own parts.

Because their pages are structured and yours are PDFs. Manufacturer pages with stock, specs, and schema reliably reclaim those SERPs.

Does the San Jose local pack even matter in B2B?

For services with a truck involved — assembly, rework, calibration, same-day stock — it's decisive. For components, global spec pages matter more. We do both, separately.

How much should I expect to pay for SEO in San Jose?

Read the bill of materials above: $1.90 per part page, billed only once Google has indexed it, $0.95 after the first five thousand, and a $790 monthly hold for up to 10,000 pages. For scale, the three organic results above this page on the search: a template site whose phone number is still the template’s placeholder, a national agency’s list ranking itself first, and a page built in Microsoft FrontPage that guarantees “first page top 3 on Google” and quotes a $99 service. One agency further down estimates $1,000 to $5,000 a month with boutique work from $3,000. None of them is priced per page, because none of them builds pages by the thousand.

Is SEO still relevant in 2026 for a component maker?

More than for anyone else on this site. An engineer typing a part number is the highest-intent search that exists, and the chart above shows who answers it today: distributors and marketplaces in every country, the manufacturer in none. AI answers draw from the same pages. The maker with a real page per part — specs, stock, cross-references — is the one both the engineer and the model can cite.

What does an SEO company actually do with a catalog?

Reads your data — PDFs, CSV exports, the ERP or PLM — and builds one page per part from the real fields: pitch, pin count, current rating, temperature range, stock, lead time, the drawing. Adds the schema that tells Google it is a product, the cross-reference table that answers “alternative,” and the hreflang pair for every market you sell into. Then reads Search Console by part family every month and fixes what it shows.

Who actually buys from a San Jose supplier?

Engineers who are not in San Jose. The three-country matrix above is the same part number searched from the United States, Germany and Taiwan: the same distributors hold page one in all three, and no manufacturer’s own page appears in any. Your buyer is in Munich or Hsinchu as often as on North First Street; the page has to be built for the search, not the ZIP code.

Can you work with our ERP or PLM data and our engineers’ review?

Yes. The pipeline reads what you already maintain — a CSV export, a database view, a PLM API — and the page template is reviewed by your engineers before the first batch ships. Where your site is code, the pages arrive as pull requests. Nothing is retyped by hand and nothing is invented by a model; a field with no source value stays blank rather than guessed.

Do you guarantee rankings?

No, and the datasheet’s absolute maximum ratings say so in a table. What is guaranteed is the billing rule: you pay per page Google has indexed, which is a number you can read yourself in Search Console. The page one above this one contains a $99 top-three guarantee from 2015; it also, two paragraphs later, warns you never to trust guarantees.

What is the first step, and what does it cost?

The crawl, below, and nothing. Within 48 hours you get a count: how many of your SKUs are indexable today, how many are locked in PDFs or behind a login, which fields we can extract from what you have, and the bill of materials for the first batch of a thousand.

Point us at the catalog. In 48 hours you know how many of your parts Google can see, and how many are locked in a PDF.

We crawl what you have — the site, the PDFs, the export you send — and return a count: SKUs indexable today, SKUs invisible, the fields we can extract from your data as it is, and the bill of materials for the first batch of a thousand. A number your engineers can check, not a proposal. Three fields are required.

OUTPUT: indexable.csv · locked.csv · fields.json · bom.pdf  ·  RUNTIME: ≤ 48 BUSINESS HOURS  ·  COST: $0