ARKS Infinity Logistics: a 51-page platform, built from a blank page.

The starting point

There was no previous website to improve. ARKS Infinity was a newly incorporated company with deep operational expertise and no digital footprint at all — no site, no content architecture, no search presence, no enquiry channel.

That is a harder brief than a redesign. A redesign inherits decisions. This one required every decision to be made from nothing: what the business actually sells, how buyers search for it, how to structure it, and how to make a first-year company credible to buyers who verify their suppliers.

Project objectives

• Explain a multi-layered service offering without overwhelming a first-time visitor.
• Establish credibility for a new company in a trust-critical industry.
• Create an architecture that could rank for commercial searches across services, destinations and commodities.
• Build clear paths from arrival to enquiry.
• Hand over a platform the client could extend themselves, without a developer.

The core challenge: three dimensions, one navigation bar

A freight forwarder does not sell one thing. It sells a service, applied to a commodity, moving along a corridor. A customer might need customs clearance for pharmaceuticals going to Canada, or door-to-door delivery of chemical dyes to the UAE.

Most logistics websites collapse that into a single flat services list, which fails two audiences at once. The exporter who knows exactly what they need cannot find their specific case. And search engines see one thin page trying to rank for hundreds of distinct queries.

The architecture had to hold all three dimensions without turning the navigation into a maze.

The solution: three parallel content systems

I built three independent but interlinked page families, each answering a different question a buyer arrives with.

Each family has its own hub page (/our-services/, /trade-lanes/, /commodities/) and its own dropdown. A visitor never has to understand the whole business — only the dimension they arrived thinking about.

Why this matters commercially

Real logistics searches are compound. Somebody types “customs clearance for pharmaceuticals to Canada”, not “logistics company”. A single services page cannot answer that. Three intersecting systems can, because any two dimensions combine into a specific, low-competition, high-intent query.

It also creates room to grow. A new corridor or commodity is a new page inside an existing pattern, not a redesign. The architecture was built to be extended.

WordPress development

This is a custom build on WordPress, not a theme install. The work splits across two pieces of code I wrote for this client.

Custom child theme

Holds the design system — the full stylesheet, the JavaScript, and the PHP that controls layout, the footer, blog templates and the site’s structural behaviour. Built on GeneratePress as a lightweight framework rather than a page-builder, which is why the front end ships without the DOM bloat a builder introduces.

Custom companion plugin

Holds the functionality that must survive a theme change: custom content types, the enquiry system, security hardening, block patterns and structured-data output. Separating theme from function is deliberate — the client can restyle in future without losing their data or their forms.

Built to be handed over

Pages are Gutenberg blocks with a shared component vocabulary — hero, section, card grid, feature, checklist, steps, call to action. The client edits content in the normal WordPress editor. No page builder licence, no developer dependency for routine changes, no rebuild required to add a page.

One thing I had to solve during handover: the client’s editor account was silently stripping inline SVG icons on every save, because WordPress filters unfiltered HTML for non-administrator roles. I added a filter to the plugin whitelisting the specific SVG elements and attributes the design uses. The client can now edit any page without breaking the iconography.

Lead generation and the quote journey

Two distinct paths, because two different visitors need two different things.

The quote form

A nine-field form built for shipment scoping rather than generic contact: name, company, email, phone, origin, destination, commodity, service required (ocean freight, air freight, customs advisory, end-to-end orchestration, project and heavy cargo) and shipment details. Origin and destination are the fields that let the operations team price a lane instead of replying with “what are you shipping?”.

The enquiry form

A separate contact form capturing commodity and destination alongside the usual details — again, structured so the first reply can be substantive.

A submission system that cannot lose a lead

I wrote a custom handler rather than installing a form plugin. Every submission is stored as a record in the WordPress admin and emailed to the team. If mail delivery fails — a genuine risk on shared hosting — the enquiry is still saved and visible. Spam is filtered with a honeypot field rather than a CAPTCHA, so there is no friction for real users.

Shipment status requests

The homepage carries a short status-request form — reference number and email — that routes to the operations desk. The dedicated tracking page explains the milestone communication process and states plainly that a self-service portal is still in development. The site does not claim live tracking software it does not have, which was a deliberate decision.

Performance and technical delivery

Methodology: three anonymous, cache-busted requests per page, median reported. Cache-busting forces a full origin render, so these figures represent the slowest realistic case — a cached response to a normal visitor is faster. These are server response and payload measurements, not Lighthouse scores, and I have not published a PageSpeed number because I did not run one.

Supporting measurements: 5 stylesheets, 6 scripts, 603 DOM nodes on a representative page, served over HTTP/2 with Brotli compression and LiteSpeed page caching. Those are lean numbers for a site of this size, and they are a direct consequence of not using a page builder.

Asset and rendering work

• Every content image converted to WebP.
• The route map built as inline SVG over a 49 KB raster basemap rather than a heavy JavaScript mapping library.
• Explicit width and height on inline SVG icons — without them an icon renders full-container-width if the stylesheet is slow or stale, which is a genuine layout-shift risk I hit and fixed during QA.
• Server-side page caching enabled at launch.

Security configuration

Verified live on the production site:

• Content-Security-Policy, Strict-Transport-Security, X-Content-Type-Options: nosniff, X-Frame-Options: SAMEORIGIN, Referrer-Policy and Permissions-Policy all present.
• X-Powered-By removed — no PHP version disclosure.
• XML-RPC disabled at both application and server level.
• REST user enumeration blocked. This was the most important find of the pre-launch audit: the default WordPress REST endpoint was exposing the administrator account name, which is half of a brute-force attempt. Closed before the site was public.

Building credibility for a first-year company

Early drafts of the content carried the numbers logistics templates ship with: millions of shipments delivered, a 99.6% on-time rate, a row of client logos, a fleet of owned warehouses.

None of it was true. The logos were invented company names. The warehouses did not exist. The company was in its first year.

I stripped all of it out before the site went live, and it was a commercial decision rather than a moral one. ARKS sells to procurement and compliance teams who run supplier due diligence. A claim that fails a background check does not merely fail to persuade — it ends the relationship silently, and you never learn why.

What replaced it

Credentials that survive verification:

• Corporate Identification Number, published in the footer — independently checkable on the MCA register, and legally required on an Indian company website.
• CBLR-qualified leadership — the founder’s customs broker examination under Regulation 6 of CBLR 2018, presented as a personal qualification, never as a company licence.
• Founder’s record in customs since 1993, stated consistently everywhere rather than inflated on one page and contradicted on another.
• ICEGATE and E-Sanchit familiarity — the customs documentation systems Indian shipments run through, with filing carried out by licensed customs broker partners.
• A real corporate address plus the Mundra operational hub, both published.

The badges I refused to add

The client could have displayed FIATA and ISO marks. I left them off, because their own accreditations page lists both as planned rather than held. A certification mark you have not earned is not decoration — it is a claim, and in this sector the wrong reader checks it.

I also repositioned the company as asset-light — corridor specialisation and vetted bonded partners — because that is what it is. The earlier “owned fleet” framing would have collapsed under the first serious question.

Development process

The project ran in four stages across roughly three weeks of build, plus a pre-launch audit and post-launch fixes.

• 1 · Discovery and architecture. Business and competitor research, requirements analysis against the client brief, sitemap and URL planning, navigation design.
• 2 · Design system and homepage. Visual language, typography, colour tokens, component library, then the homepage as the reference implementation every later page inherits.
• 3 · Content systems, in three phases. Phase one delivered the core services and first corridors; phase two the remaining lanes and commodities; phase three the network, credentials and company pages. Building in phases meant each pattern was proven before it was repeated twenty times.
• 4 · Conversion, SEO, hardening and launch. Forms and lead storage, metadata and structured data across all pages, image conversion, security headers, responsive and cross-browser QA, full audit, then deployment.

Post-launch I ran a further round of fixes on client feedback — corrected statistics, the credibility purge described above, accessibility labelling on third-party widgets, and the scroll and SVG bugs found in QA.

Results and strategic impact

Verified outcomes

Directly observable on the live site today:

• 51 indexable URLs, every one returning 200.
• Zero broken internal links across 47 checked.
• One H1 per page and unique, in-range metadata sitewide.
• Median server response 141–207 ms on cache-busted requests, with 79–115 KB HTML payloads.
• Six security headers live, version disclosure removed, user enumeration blocked.
• Pre-launch audit closed at zero outstanding issues.

Strategic outcomes

Built into the architecture, not yet measurable:

• A structure that can target compound commercial queries head-on rather than competing for generic head terms.
• A credibility position that survives procurement due diligence — increasingly the deciding factor for corporate freight buyers.
• Content infrastructure the client extends without a developer.
• A lead system where no enquiry is lost to email failure.

What is not claimed

No traffic, ranking, lead-volume or revenue figures appear in this case study. Analytics were still being configured at launch, so no meaningful measurement window exists yet. Publishing an estimate would contradict the entire premise of the project. When real data exists and the client has approved it, it will be added here.