Service 02 · Product Visibility
Engineers and purchasing teams search for: "pressure rating PN16 flange DN50", "alternative to Siemens 6ES7531-7KF00", "3 piece stainless 316 ball valve". Your catalog has all of it — but it is locked inside images and PDFs. Google does not read it. ChatGPT answers: "I do not have specific data." So the buyer buys from a competitor whose specs are available as text. I turn your catalog into structured content: 1 product = 1 page of text + structured data + the phrases buyers actually use.
Request a Free AuditThe Problem
5,000 SKUs. Every SKU has a product photo, a datasheet PDF, and a price in a spreadsheet. The marketing team is proud: "our catalog is complete." But when an engineer searches Google for "flange PN16 DN50 stainless 316", your catalog does not show up. A competitor with 500 SKUs and tidy text pages is on page 1 instead. The buyer buys from them. You never know which opportunities passed you by.
Google does not read text inside images. A PDF can be indexed, but it is not structured. The specs that matter — rating, material, dimensions, certification — cannot be read by a machine.
Products live only in a downloadable PDF catalog, not on web pages. Google does not crawl download files well. No URL per product means no entry point from search.
The description says: "a high-quality flange for a wide range of applications." The engineer needs: "PN16, DN50, ASTM A182 F316, raised face, 4 bolt holes." No answer means no ranking.
Two Places Buyers Look
Buyers type a technical spec, a part number, or an OEM code. They compare 3–5 results and click the one with the most complete specs. Positions 1–3 get 60% of clicks. Positions 4–10 get 30%. Page 2 is barely seen.
Buyers ask: "What is an alternative to Siemens 6ES7531-7KF00 available in Indonesia?" or "3 piece stainless 316 ball valve for a chemical application — which brand fits?" The AI answers from structured data and credible text content. If your product has no structured data and no content that answers the question, it does not get cited.
What Gets Built
Structure first, content second. Without the right structure, even the best content struggles to get indexed.
Every page is written to answer a question from an engineer or a purchasing team, not to be a marketing brochure.
So the product is not only found, but cited correctly by AI.
We do not measure "keyword volume". We measure whether your product answers the questions buyers actually ask.
Measurement
Every number is recorded on day one, before any change. The monthly report uses the same method and the same data. No picking whichever metric looks good.
For the 50 questions: the share of search results your products occupy (organic + AI Overview + PAA + Shopping). Target: above 40% SoS by month 6.
The percentage of questions where your product is cited in ChatGPT, Gemini, or AI Overviews. Target: above 30% by month 6.
The number of RFQ forms or WhatsApp chats that come from organic search or an AI answer. Target: 3x baseline by month 6.
The percentage of the 5,000 SKUs that already have a canonical page + structured spec table + schema.org + FAQ. Target: 100% of priority SKUs by month 3, 100% of all SKUs by month 6.
The Outcome
A measuring instruments distributor: 2,400 → 10,500 sessions per month. 40% of new inquiries came from the AI search channel.
Not after the RFQ has gone out to 5 vendors. The engineer finds you during early research, puts you on the shortlist, then sends the RFQ.
Every product page is a search entry point working 24 hours a day. The results compound: the longer it stays live, the more questions it catches.
The visitor has read the specs, compared, and knows your product fits. The RFQ that arrives reads: "I need 50 pcs of flange PN16 DN50 316L" — not "please send me information about your products".
FAQ
An SEO agency: keyword research → blog content → backlinks. Me: technical catalog structure → structured product pages → structured data → measurement from buyer questions. I do not write blogs. I make your catalog readable by machines. That is the focus.
Images are for people. Text is for machines. You need both. The solution: the product page has both — an HTML spec table (machines), product images, and a PDF download (people). Not one or the other.
I use 50 real buyer questions from sales calls, email, and WhatsApp. Every month those questions are run against ChatGPT, Gemini, and AI Overviews through the API and checked by hand. What gets recorded: whether your product is cited, which URL, and at what position. Screenshots serve as evidence, using the same method every month.
Baseline in the first week. Indexing and crawling in weeks 2–4. Early ranking movement in the second month. Significant results in months 3–4. AI citations need authority to build, so they start showing in months 4–6. All of it measured against the same baseline. No "it looks like it went up".
Not necessarily. If the CMS supports custom page templates, JSON-LD injection, and clean URLs, we use what you have. WordPress, Webflow, a custom CMS, a static site — all work. What has to be rebuilt: a CMS that locks the URL structure, does not allow schema injection, or cannot be made fast enough.
A large catalog is exactly where the biggest results come from. 10,000 SKUs means 10,000 search entry points. But it is not done by hand. What I build: (1) a structured product page template, (2) a data pipeline from the catalog master that generates the pages automatically, (3) automated QA that checks the schema is valid, the specs are complete, and the canonical is unique. Priority: the 20% of SKUs that get searched most, first.
Send a list: "what do customers usually ask?" (for example: "flange PN16 DN50", "alternative to Siemens 6ES7..."). I check the Google position, the AI appearance, and whether your catalog answers. A 1-page report. Free. No pitch.