The US facilities manager evaluating your reverse osmosis unit at 10am is checking your NSF number against the NSF database in a separate tab. Make sure she can find it.

It is 10:14 on a Wednesday. A facilities manager at a US senior living operator is researching reverse osmosis units for a 240-unit refurbishment. Her shortlist has six manufacturers: two Korean, two Chinese, one Italian, one US-based. She opens the first Korean manufacturer site. The homepage loads in Korean with a small English flag in the top corner. She clicks it. The English page surfaces three product categories but no certification numbers, no flow rate specs visible without downloading a PDF, no distributor list for the United States, and a contact form routed to Seoul.

She has nine minutes before her next meeting. She closes the tab. By the time the manufacturer's sales team replies the next day, she has already requested quotes from the two manufacturers whose websites surfaced their NSF/ANSI 58 certification number, their flow rate per fixture, and their nearest authorised distributor in Pennsylvania within the first ninety seconds.

That manufacturer just lost a $180,000 procurement deal. Not on the device, which is competitive and properly certified. On the website. The artefact representing the product was built for the buyer in Seoul, not the buyer in Philadelphia.

Three structural reasons.

First, Western water treatment procurement is certification-led. NSF International, WaterSense, WRAS, KIWA, DWI, ACS, DVGW. Each of these is a body that issues certifications buyers cross-reference against the public database within their first three minutes of evaluation. The certification number is the trust signal. A site that does not surface NSF/ANSI 42 alongside NSF/ANSI 53 and NSF/ANSI 61 (different certifications cover different aspects of the same product) is a site the buyer cannot verify, and the buyer moves on.

Second, Western water treatment is distributor-mediated. Plumbers, mechanical contractors, facilities teams, and HVAC service companies in the US, UK, and EU rarely buy direct from overseas manufacturers. They source through regional distributors who handle install training, warranty, and spare parts inventory. A manufacturer site without a distributor finder forces every buyer through a "contact us" form that routes to Asia and takes 24 to 72 hours to reply. The deal is already gone.

Third, Western buyers evaluate on performance data, not marketing claims. Flow rate at typical inlet pressure, contaminant removal efficiency per third-party test, NSF 53 lead reduction percentage, replacement cartridge service life, recovery rate for RO systems. These are real numbers from real test reports. The manufacturer that surfaces them as a structured data table on the product page wins the technical-buyer evaluation; the one that buries them in a brochure PDF loses.

1. Certification clarity in the first 60 seconds

NSF/ANSI 42 (aesthetic effects), NSF/ANSI 53 (health effects), NSF/ANSI 58 (reverse osmosis systems), NSF/ANSI 61 (system components), NSF/ANSI 372 (lead-free compliance). WRAS approval number for UK. KIWA Watermark for the Netherlands and broader EU recognition. ACS attestation for France. DVGW for Germany. ISO 9001 certificate. Surface these on the product page itself with the certificate number linked to the relevant public database where the buyer can cross-reference in real time. A small badge with the NSF number that links to the NSF certification database is the single highest-impact element on the page.

2. Performance data as a structured table per product

Flow rate at 60 psi inlet, contaminant removal percentages per NSF protocol, capacity in gallons or litres, recovery rate for RO, service intervals, replacement cartridge specifications. Rendered as a structured data table on the product page itself, not as a brochure PDF download. Marked up with structured data so AI engines and search engines can extract the specifications. Buyers compare manufacturers by lining up these tables side by side; the manufacturer that makes the data accessible wins more comparisons.

3. English-primary regional sub-paths with proper hreflang

The /us/, /eu/, and /uk/ sub-paths each present native English content with region-specific certification positioning (NSF for US, WRAS for UK, KIWA and CE marking for EU), distributor information, and regulatory context appropriate to that market. Bidirectional hreflang signals to Google and AI engines which page belongs to which market. The home-market language remains primary on the corporate root, with regional flows existing alongside.

4. Distributor finder and dealer application portal

Interactive map showing your authorised distributors per region with current contact details, product line coverage, and certification status (some distributors are certified to install certain product lines but not others). A separate flow for prospective distributors to apply for territorial rights with qualification questions mapped to your existing channel partner programme. Most Asian-manufacturer sites have a static list with broken phone numbers or no distributor information at all. The deal closes in the time it takes a US plumber to find a phone number that connects to a human.

5. Technical documentation as findable structured documents

Installation manuals, maintenance manuals, spare parts catalogues, MSDS sheets for consumable cartridges, performance test reports from third-party labs, IFU documents per product per language. Hosted as findable linkable documents (not gated PDFs behind a sales contact form). EU MDR-equivalent requirements for technical documentation apply to certain water treatment categories; the documentation must be available in the official language of every member state where the product is sold.

6. AI search and category-query visibility

When a US plumber asks ChatGPT, Perplexity, or Google AI Overviews "best Korean reverse osmosis manufacturer with NSF 58 certification for commercial installations", you need to be in the citation list. The work that drives this is structured: Product and Service schema with proper certification properties, FAQPage schema on category pages, peer-reviewed citation density where applicable, clean entity graph with Wikidata Q-item and sameAs unification across LinkedIn, Crunchbase, and water industry directories.

A 12-to-20-week build. Three to four weeks of discovery and certification documentation gathering. Six to ten weeks of build covering the foundation, regional flows, and the certification clarity layer. Two weeks of distributor architecture (interactive map, application portal, partner-side dashboard). Three to four weeks of content production: performance data table population, IFU and DFU document migration with native technical writer review per language, and the AI search content layer.

Stack is Astro for the public marketing layer, Supabase for the distributor portal and authenticated content, Cloudinary or imgix for the technical and product image pipeline, and a translation workflow with native technical writers per target language. Schema rollout on every page: Product or Service with certification properties, Organization with sameAs to Crunchbase, LinkedIn, NSF database, WRAS database. Performance budget: WCAG 2.1 AA accessibility (US procurement requirement), Core Web Vitals passing on mobile, sub-2-second LCP under 4G simulation.

The work is sequenced so the certification clarity layer ships first, the performance data tables and distributor architecture second, the AI search programme third. Each phase produces a measurable outcome: the certification layer reduces buyer friction in the first 60 seconds; the performance data tables position the brand for technical-buyer evaluation; the AI search programme captures the discovery flow that increasingly precedes the contact form.

Three shapes of project I turn down on the kickoff call.

The "translate the existing home-market site into English" brief. The structural mismatch is not a translation problem. A machine-translated or even human-translated home-market page is the wrong artefact for a Western buyer. The right brief is "build a Western-buyer-flow site that lives alongside the home-market site". I will turn down the translate-only version and explain why on the call.

The "we have not received our first NSF or WRAS certification yet" brief. If the manufacturer does not yet hold at least one Western-market certification for one product line, the certification clarity layer (the highest-impact element of the build) has nothing to surface. The right move is to wait until the first certification is granted, then ship the site sized to that single certified product, then expand as further certifications arrive.

The "we want to sell direct to US consumers via Amazon and Shopify" brief. That is a different shape of business. Direct-to-consumer water treatment requires Amazon listing optimisation, retail packaging design, and consumer-facing marketing, none of which I specialise in. The procurement-grade B2B site I build serves the distributor-mediated channel; if direct-to-consumer is the primary growth motion, you need a different agency.

Three forces converged this year and the next eighteen months are the highest-impact window for this work.

NSF certification fee increases hit in early 2026. Manufacturers carrying the cost of multi-product NSF certification are looking harder at converting those certifications into Western-market deals to justify the recurring renewal cost. Sites that surface certifications properly improve the conversion-per-certification number meaningfully.

US municipal infrastructure spending accelerated through 2024 and 2025. The Bipartisan Infrastructure Law allocated significant water-system upgrade funding. Group Purchasing Organisations (Premier, Vizient, HealthTrust) are running RFP intake through digital portals that pull manufacturer data from the public website. A site that surfaces NSF clearances, performance data, and technical specifications in a procurement-portal-friendly way captures GPO inclusion at a rate manufacturers without these capabilities cannot match.

AI-driven vendor discovery is reshaping the early stage of the buyer journey. ChatGPT, Perplexity, Google AI Overviews are increasingly the first place a Western facilities manager or specifying engineer asks "who manufactures X with Y certification". The brands that ship AI-citation-ready content this year define the consideration set for the next eighteen months.

Why does a Western-buyer flow need to be separate from the home market site?

Western buyers evaluate on certification numbers, third-party performance data, and regional distributor accessibility. Home-market buyers evaluate on different signals. A single home-language site cannot serve both audiences without one feeling like a translation artefact. Regional sub-paths (/us/, /eu/, /uk/) with native English copy and proper hreflang is the pattern that works.

What certifications should the public site surface?

For US: NSF/ANSI 42, 53, 58, 61, 372 as applicable, plus WaterSense for water-efficient products. For UK: WRAS approval number, DWI listing. For EU: KIWA Watermark, ACS for France, DVGW for Germany. Across all: ISO 9001 certificate. Surface these on the product page itself with database links for cross-reference.

How important is the distributor finder?

Highest single-element importance. Western buyers source through regional distributors. A manufacturer site without a working distributor finder closes deals at fraction of the rate of a site with one. Build it; it is the single highest-conversion element on the site.

How long does the build take?

Twelve to twenty weeks for a typical mid-market manufacturer with three to eight product lines. Three to four weeks discovery, six to ten weeks build, two weeks distributor architecture, three to four weeks content production.

Why does AI search visibility matter for water treatment manufacturers in 2026?

Western buyers increasingly start vendor discovery on ChatGPT, Perplexity, or Google AI Overviews. A US plumber asking "best Korean reverse osmosis manufacturer with NSF 58" receives an answer with three to five cited sources. If your brand is not in those citations, you are not in the consideration set.

What does this cost?

Build $30,000 to $100,000 USD (£24,000-£80,000 GBP) depending on product line count and regional scope. Retainer $5,000 to $10,000 per month (£4,000-£8,000) for content, certification updates, AI search tracking, distributor portal maintenance.

The build, $30,000 to $100,000 USD (£24,000 to £80,000 GBP), 12 to 20 weeks

Lower end covers single-product-line manufacturers with US-only scope and existing English content. Upper end covers multi-product-line manufacturers with US plus EU plus UK scope, full performance data table population, distributor portal architecture, and native technical writer translation pipeline. Most engagements land between $50,000 and $75,000 USD. Project pricing is fixed-scope; surprises are absorbed inside the engagement, not billed as variations.

The retainer, $5,000 to $10,000 per month (£4,000 to £8,000), ongoing

Content production (technical articles, comparison guides, application case studies one to three per month), certification updates as new NSF or WRAS approvals arrive, AI citation tracking across ChatGPT, Perplexity, Gemini, Google AI Overviews, distributor portal maintenance, quarterly executive review with Western-market analytics. Three-quarter minimum commitment because AI search surfaces re-crawl on a four-to-eight-week cycle.

The free AI Citation Checker derives five buyer queries for your category and reports whether your brand is cited in the top 10 organic results that ChatGPT and Perplexity pull from, plus whether Google AI Overviews are running on those queries today. No login, no email, takes about 20 seconds.

Book a 30-minute call. Tell me your product lines, your existing Western-market certifications, and the three target markets that matter most for the next twelve months. By the end of the call you have a stack pick, a price range, a delivery window, and a clear sense of whether I am the right person for this engagement.