Pyramid Talent
Website Ownership
Approach

Tech Stack

Rendering Strategy

Each page type gets the optimal rendering mode. Zero unnecessary compute — only what changes gets rebuilt.

The Pages Rebuild Problem — Solved

Caching Strategy — 3 Layers

Infrastructure & Scalability

• Cloudflare at the front — CDN, WAF, DDoS protection, rate limiting, bot filtering
• AWS Load Balancer — distributes traffic across EC2 instances, handles spikes without linear cost
• Auto Scaling Groups — scale up on demand, scale down off-peak. Cost proportional to actual traffic.
• AWS S3 + CloudFront — all media (images, videos, docs) from CDN, not origin
• OCI for origin compute — significantly cheaper than AWS EC2 for base compute layer

Security

Headless CMS for Marketing Team

• WordPress as headless CMS — marketing team already knows it, no retraining
• ACF — structured content fields for every content type
• CPT — for Jobs, Case Studies, Resources, Industries, Testimonials
• REST API / WP GraphQL — Next.js fetches at build time (SSG/ISR) or request time (SSR)
• Live preview — marketing can preview unpublished changes before going live
• Media via S3 — WordPress uploads offloaded to S3, served via CloudFront CDN

SEO Strategy

• Technical SEO: Schema markup (Organization, JobPosting, Article, BreadcrumbList), XML sitemaps with lastmod, canonical tags, hreflang for US market
• Core Web Vitals: LCP < 2.5s, CLS < 0.1, INP < 200ms — monitored via GSC and PageSpeed Insights
• pSEO: Dynamic page generation for industry × service × location at scale using CPT + ACF
• Metadata: Next.js Metadata API for dynamic title, description, OG tags per page
• Internal linking: Automated cross-links between related industries, services, resources
• Lighthouse target: 90+ across all core pages

GEO — Generative Engine Optimisation

Optimising for AI answer engines (ChatGPT, Perplexity, Gemini, Claude) alongside traditional search:

• Semantic HTML — clean, well-structured markup that LLMs can parse accurately
• Schema markup — structured data makes content machine-readable for Google and LLMs
• llms.txt — emerging standard (like robots.txt for LLMs) — tells AI crawlers what content is available
• Fast TTFB (<200ms) — AI crawlers time out on slow servers
• Authoritative content structure — clear headings, factual statements, cited proof points
• sitemap.xml with lastmod — helps AI crawlers prioritise fresh content

AIO — AI Overview Optimisation

For Google's AI Overviews (SGE):

• Answer-first content structure — lead with direct answer, then expand
• FAQ Schema — structured Q&A increases AI Overview inclusion likelihood
• E-E-A-T signals — Experience, Expertise, Authoritativeness, Trustworthiness on every page
• Clear entity definitions — each page defines who Pyramid Talent is, what it does, for whom

Performance

• next/image — automatic WebP conversion, lazy loading, correct sizing
• Font optimisation — next/font for zero layout shift, preloaded critical fonts
• Bundle splitting — dynamic imports for heavy components (Three.js, carousels)
• Prefetching — Next.js <Link> prefetches on hover, instant navigation
• No render-blocking resources — all non-critical JS deferred
• Image CDN via CloudFront — global delivery, compressed at edge

Accessibility

• Semantic HTML throughout — <nav>, <main>, <section>, <article>, <button>
• Single <h1> per page, logical heading hierarchy
• All images have descriptive alt text
• Keyboard-accessible interactions — all CTAs, carousels, modals operable without mouse
• Colour contrast — WCAG 2.1 AA minimum (4.5:1 for body text)
• Focus indicators visible on all interactive elements
• Respects prefers-reduced-motion

Cost Optimisation

• Static pages on CDN = near-zero compute cost per request
• ISR means only changed pages consume compute — not the full 100k+ site
• OCI for origin server — cheaper compute than AWS for base layer
• S3 + CloudFront for all media — removes load from application servers
• Redis caching — reduces DB queries, lowers compute requirements
• Auto-scaling groups — no idle cost during off-peak hours

Monitoring & Incident Response

• NetData — continuous uptime monitoring across all properties
• AWS CloudWatch — server metrics, error rate alerts, traffic anomaly detection
• Cloudflare Analytics — edge-level traffic insights, bot detection, WAF trigger logs
• GA4 + GSC — user behaviour, organic performance, indexing health
• SEMrush — keyword rankings, backlink monitoring, competitor tracking
• On-call response — clear escalation path, rollback via Vercel one-click