Build commerce infrastructure that scales with your revenue.

A 100ms improvement in page load time increases conversion by 1%. At $10M annual revenue, that is $100K from a single performance optimisation.

Ecommerce is one of the few domains where frontend performance directly converts to revenue. Every 100ms of additional page load time reduces conversion rate by approximately 1% — measured across thousands of ecommerce deployments by Google. Checkout abandonment correlates directly with checkout step count and payment method availability. Search result quality determines whether users find what they want or leave. These are engineering problems with measurable business outcomes. PROPELOO builds ecommerce platforms where performance is a first-class engineering requirement, checkout is optimised for conversion, and every component is measured against its revenue impact.

The full ecommerce engineering stack.

Production ecommerce has six engineering domains — each with direct revenue impact.

System Layers

• Storefront Layer: Product catalogue, search, filtering, product detail pages, cart — performance-optimised
• Checkout Layer: Multi-step or one-page checkout, address validation, shipping calculation, tax
• Payment Layer: Multi-PSP integration, 3DS2, saved cards, wallets (Apple Pay, Google Pay), BNPL
• Order Management Layer: Order processing, fulfilment routing, inventory allocation, returns, refunds
• Operations Layer: Merchant admin, product management, pricing rules, promotions, analytics

Core Technical Capabilities

• High-performance Storefront

  Next.js with ISR/SSG for product pages, edge caching via Cloudflare, Core Web Vitals optimisation (LCP < 2.5s, INP < 200ms), image optimisation pipeline and search with Algolia or Typesense for sub-100ms results.

• Conversion-optimised Checkout

  Single-page checkout with address autocomplete (Google Places), real-time shipping rate calculation, tax computation (TaxJar/Avalara), order summary and guest checkout — minimising form fields and page transitions that cause abandonment.

• Multi-PSP Payment Integration

  Stripe as primary PSP with Razorpay for India, PayPal, Apple Pay/Google Pay, BNPL (Klarna/Afterpay), 3DS2 for EU compliance, saved payment methods and smart retry logic for failed payments.

• Product Search & Discovery

  Algolia or Typesense for faceted search — instant results, typo tolerance, synonym handling, personalised ranking, dynamic filtering and search analytics to identify zero-result queries.

• Inventory & Order Management

  Real-time inventory allocation at checkout, multi-warehouse routing, backorder management, returns processing with restocking, refund automation and fulfilment partner integrations (ShipBob, EasyShip).

• Commerce Analytics

  Conversion funnel analysis, cart abandonment tracking, revenue attribution, product performance dashboards, cohort analysis for repeat purchase rate and A/B testing infrastructure for checkout experiments.

How we think about ecommerce engineering.

Every engineering decision in an ecommerce platform has a conversion impact. Performance, checkout UX and payment success rate are not metrics — they are revenue.

• Product page performance is conversion infrastructure

  A product page that loads in 1.2 seconds converts at a meaningfully higher rate than one that loads in 3 seconds. Core Web Vitals (LCP, INP, CLS) are not technical vanity metrics — Google uses them as ranking signals and users feel them as page quality. ISR (Incremental Static Regeneration) in Next.js serves product pages from CDN cache while updating in the background — the fastest possible delivery of frequently-changing content.

  Axiom:

• Every checkout field is a conversion risk

  The average ecommerce checkout has 23 form fields. Studies show that reducing to 12 increases conversion by up to 20%. Address autocomplete eliminates most address entry. Guest checkout removes the registration barrier. Saved payment methods reduce checkout to two taps for returning customers. Checkout engineering is UX engineering with a direct revenue coefficient.

  Axiom:

• Payment success rate is not 100% by default

  Card declines, 3DS authentication failures, network timeouts and issuer blocks cause payment failures that are not the customer's fault. Smart retry logic (retry declined transactions with different acquirer routing), fallback PSPs (if Stripe fails, route to Braintree), and clear error messaging with recovery paths recover a meaningful percentage of otherwise-lost revenue.

  Axiom:

• Search is the highest-intent customer touchpoint

  A customer who uses site search converts at 3-5x the rate of a customer who browses. Zero-result searches are abandoned sessions. Synonym handling ("sneakers" returning "trainers"), typo tolerance, and personalised ranking based on purchase history are not advanced features — they are the baseline for a search experience that does not lose high-intent customers.

  Axiom:

The ecommerce architecture decisions that matter.

Each choice has measurable conversion and operational implications.

• Custom build vs headless Shopify vs full custom?

  Impact: Headless Shopify (Shopify + custom Next.js storefront) for most D2C brands — Shopify handles payments, inventory and admin; custom storefront handles performance and UX. Full custom for marketplaces, complex B2B requirements or products that do not fit Shopify's data model.

  • Shopify + custom storefront (headless) — fastest to market, Shopify payments, customisation limits
  • Shopify with theme customisation — simplest, most constrained
  • WooCommerce — WordPress-based, large ecosystem, performance challenges at scale
  • Full custom — maximum control, 4-8 months to build, right for complex requirements

• SSR vs SSG vs ISR for product pages?

  Impact: ISR is the correct default for ecommerce product pages — pages are served from CDN cache (near-static speed) but regenerated in the background when product data changes. SSR only for pages with personalised content that must be accurate per-user.

  • SSG (static generation) — fastest, must rebuild on product change
  • ISR (incremental static regeneration) — near-static speed, updates in background, best for most ecommerce
  • SSR (server-side rendering) — always fresh, higher TTFB, necessary for personalised content
  • CSR (client-side rendering) — worst for SEO and LCP, not recommended

• Primary payment processor?

  Impact: Stripe for global D2C. Razorpay for India-primary. Adyen for high-volume merchants where custom acquiring pricing is justified (>$1M/month). Always integrate Apple Pay/Google Pay — wallet payments have significantly lower abandonment than card entry.

  • Stripe — best developer experience, global coverage, 2.9% + 30¢
  • Razorpay — India-focused, UPI/netbanking, 2% fee
  • Braintree (PayPal) — good for US/EU, lower volume pricing available
  • Adyen — institutional-grade, volume pricing, complex integration

• Search infrastructure?

  Impact: Algolia for most ecommerce with <5M monthly searches. Typesense self-hosted for cost-sensitive applications or catalogues over 100K products where Algolia cost becomes significant. Database full-text search only for catalogues under 1,000 products.

  • Algolia — fastest, best features, expensive at volume ($1/1000 searches)
  • Typesense — open source, self-hosted, 90% of Algolia features, lower cost
  • Elasticsearch — maximum flexibility, high operational overhead
  • Database full-text search — simplest, poor relevance for large catalogues

• Inventory management approach?

  Impact: Reservation + TTL (hold inventory for 15 minutes during checkout, release if payment not completed) is the standard approach — prevents overselling while not permanently allocating inventory during abandoned checkouts.

  • Real-time allocation at checkout — prevent overselling, requires atomic transactions
  • Reservation + TTL — hold inventory during checkout flow, release if not purchased
  • Optimistic (allow oversell, handle manually) — simplest, creates customer experience problems
  • Warehouse management system integration — enterprise approach, complex integration

• Multi-currency and international?

  Impact: Multi-currency with local acquiring (charge EUR for EU customers via EU acquirer) maximises conversion — eliminating foreign transaction fees for customers and reducing card decline rates. Start with display-only currency conversion and migrate to local acquiring as market revenue justifies the complexity.

  • Single currency (home market only) — simplest
  • Display currency conversion (charge in base currency) — easy to implement, customers see local price but are charged in foreign currency
  • Multi-currency pricing with local acquiring — best conversion, highest complexity
  • Geolocation + redirect to market-specific store — simple, fragmented data

What PROPELOO builds.

• D2C Ecommerce Platform

  High-performance Next.js storefront with Shopify backend or custom — product catalogue, search, cart, checkout, payments and order management.

• B2B Commerce Platform

  B2B ecommerce with account-based pricing, credit terms, bulk ordering, approval workflows, ERP integration and trade account management.

• Marketplace Platform

  Multi-vendor marketplace — seller onboarding, product listing management, split payments, commission accounting and dispute resolution.

• Headless Commerce Migration

  Migrate from monolithic ecommerce to headless architecture — Next.js storefront, Shopify or custom backend, performance improvement and CMS integration.

• Subscription Commerce

  Subscription billing with Stripe Billing — plan management, free trials, proration, dunning for failed payments, customer portal and revenue recognition.

• Mobile Commerce App

  React Native shopping app with deep-linked product pages, native Apple Pay/Google Pay, push notifications for order updates and personalised recommendations.

The ecommerce stack.

Storefront performance, payment reliability and operational tooling each require specific choices.

• Storefront

  Stack: Next.js (ISR), TypeScript, Tailwind CSS, TanStack Query, Cloudflare CDN

• Commerce Backend

  Stack: Shopify (headless), Custom Node.js/Go, PostgreSQL, Redis (cart/session), Elasticsearch / Typesense

• Payments

  Stack: Stripe, Razorpay, Braintree, Apple Pay / Google Pay, Klarna / Afterpay

• Search & Discovery

  Stack: Algolia, Typesense, Elasticsearch, Personalization API

• Order & Inventory

  Stack: Custom OMS, ShipBob integration, EasyShip, TaxJar / Avalara, Returns management

• Analytics

  Stack: GA4, Mixpanel, Hotjar, Datadog RUM, Custom conversion funnel

Ecommerce security protects revenue and customer trust.

Payment fraud, data breaches and account takeover are the primary ecommerce security threats.

• PCI DSS Compliance

  Card data must never touch your servers if you use a PCI-compliant PSP (Stripe, Braintree) with their hosted fields or Stripe Elements. Using tokenisation means you never handle raw card numbers — PCI scope is dramatically reduced to SAQ A.

• Account Takeover Prevention

  Credential stuffing attacks target ecommerce accounts with stored payment methods. Rate limiting on login, CAPTCHA after failed attempts, email alerts for new login from new device and automatic logout after password change.

• Payment Fraud

  Card testing attacks use stolen card numbers to make small test purchases. Velocity limits (max purchases per IP/email per hour), CAPTCHA on checkout, 3DS2 for risk-scored transactions and real-time fraud scoring via Stripe Radar or dedicated fraud tools.

• Inventory Manipulation

  Automated bots can add items to cart faster than humans, creating artificial scarcity or facilitating resale arbitrage. Rate limiting on add-to-cart, CAPTCHA for high-demand products and queue systems for product drops.

• Data Protection

  Customer email, shipping address and order history are PII under GDPR/CCPA. Data minimisation, right to erasure workflow, clear retention policies and encrypted PII storage.

• Dependency Security

  Ecommerce platforms have large npm dependency trees. Malicious packages targeting ecommerce sites (Magecart-style attacks) inject JavaScript to skim payment data. Content Security Policy, Subresource Integrity for third-party scripts and regular dependency auditing.

From concept to converting ecommerce platform.

1. 01. Commerce Architecture

   Platform approach (headless Shopify vs custom), payment stack, search infrastructure, inventory model and performance targets.

2. 02. Storefront Development

   Next.js storefront — product catalogue, search, PDP, cart and checkout UI. Core Web Vitals targets defined and measured.

3. 03. Payment Integration

   PSP integration, Apple Pay/Google Pay, saved cards, 3DS2 and payment failure recovery flows.

4. 04. Order & Inventory

   Order management system, inventory allocation, fulfilment integration, returns and refund processing.

5. 05. Search & Discovery

   Search implementation, faceted filtering, synonym configuration, personalisation and zero-result monitoring.

6. 06. Analytics & Optimisation

   Conversion funnel tracking, A/B testing infrastructure, performance monitoring and revenue attribution.

7. 07. Launch & Scaling

   Load testing, CDN configuration, payment monitoring and post-launch conversion optimisation.

Frequently Asked Questions