Anyone can publish a service list. The useful half is knowing when not to reach for something, so every page below carries the case against itself alongside the case for it.
The stack
What the work in each category actually consists of, and where it sits in the record. A skill rated out of a hundred tells you nothing about what the missing part is.
Magento & Adobe Commerce
Building inside the platform rather than around it: modules written through the plugin and preference system so an upgrade does not become a project, multi-store structures where shared logic is genuinely shared, and B2B account hierarchies with the approval chains a real procurement process needs.
- Magento 2
- Adobe Commerce
- Custom Modules
- Multi-Store & Multi-Currency
- B2B Commerce
- Magento GraphQL
- MSI
- Upgrades & Migrations
12+ multi-store platforms carrying 500K+ SKUs and 1M+ monthly users, led from 2023. Site speed improved 60% through Varnish, Redis, full-page cache, CDN and database tuning.
Shopify & Shopify Plus
Theme architecture structured so a merchandiser can work without a developer, private and custom apps on the Admin and Storefront APIs, and checkout extensibility through the supported extension points rather than through script injection that breaks on the next platform update.
- Shopify
- Shopify Plus
- Liquid
- Custom Apps
- Storefront API
- Admin API
- Checkout Extensibility
- Shopify Functions
Storefronts for consumer brands since 2018, across D2C fashion, horticulture and sports nutrition.
Frontend
Rendering strategy chosen per route against what the data actually does, rather than one strategy applied everywhere. Static for a catalogue, incremental for what changes, server-rendered for what is personal, with Core Web Vitals treated as an acceptance criterion carrying a number.
- Next.js
- React.js
- JavaScript
- TypeScript
- Tailwind CSS
- Core Web Vitals
- PWA
Backend
The services behind a storefront and the contracts between them. API design that is versioned, documented and authenticated properly, built for the consumers that exist rather than every consumer that might.
- PHP
- Node.js
- GraphQL
- REST APIs
- MySQL
- PostgreSQL
- MongoDB
90% of B2B order and quote processing automated, cutting approval cycle time by 40%.
Infrastructure & DevOps
Where a commerce platform stops being code and starts being a running system. Caching layers, search, queues and the release process, with the part that matters written down so a deploy is not something only one person can do.
- AWS
- EC2
- RDS
- S3
- Redis
- Varnish
- Elasticsearch
- RabbitMQ
- Docker
- Jenkins
- GitLab CI/CD
- Cloudflare
Magento 2 on AWS with EC2, RDS, S3, load balancing and auto-scaling, from 2020.
Generative AI
Applied to the problems that are genuinely language problems: catalogue content at a scale no copywriter can match, assistants grounded in your own product data rather than in a model’s memory, and the operational steps that are really text moving between systems. Certified with IBM and AcceleratorX in November 2026.
- Generative AI
- LLMs
- RAG
- Embeddings
- Prompt Engineering
- n8n
- Workflow Automation
- Evaluation
Measurement
Instrumenting what the business needs to decide with, rather than what is easy to collect. Event schemas defined before the dashboard, so two departments stop disagreeing about what a conversion is.
- GA4
- Google Tag Manager
- dataLayer
- MoEngage
- WebEngage
- Technical SEO
By technology
Every page below answers the same four questions about one technology: what it is, who it suits, where it stops, and how AI extends it. They overlap on purpose, because a replatforming decision is an architecture question before it is a Magento question, and an AI project is usually a process question before it is a model question.
Shopify
Shopify and Shopify Plus work where it moves the number: theme architecture, checkout, app consolidation, integrations. Including when the platform is wrong.
Think twice when
Quote-driven pricing, negotiated terms or multi-level approvals.
Magento
Magento 2 and Adobe Commerce: multi-store catalogues at 500K+ SKUs, B2B approval workflows, performance engineering and migrations run without going offline.
Think twice when
A single store with a few hundred SKUs and no custom workflow.
WooCommerce
WooCommerce builds and performance work: bespoke themes, plugin development, checkout and catalogue customisation, and the hardening these stacks lack.
Think twice when
Variant complexity, multi-currency or B2B pricing rules.
WordPress
WordPress builds, performance and infrastructure: bespoke themes, custom plugins, a real content model, hardening, and production hosting on AWS.
Think twice when
Commerce is the business rather than a section of the site.
Headless Commerce
Headless and composable commerce on Next.js, GraphQL and the Storefront API. When decoupling pays, when it only adds a deployment surface, and how.
Think twice when
The real request is a faster site, which is usually a theme project.
AI Commerce
AI for eCommerce from a consultant who starts with the process: catalogue enrichment, semantic search and operations automation, not demos.
Think twice when
A process nobody has fixed, where automation would make it permanent.
AI Search
AI search consulting for eCommerce: semantic search, Generative Engine Optimization, and being cited by AI answer engines. What works, what is speculation.
Think twice when
A catalogue of a few dozen products, where good navigation beats any search engine.
AI Automation
AI automation for commerce: agents, LLMs, RAG and MCP applied to quotes, approvals and reconciliation, with human checkpoints where output touches money.
Think twice when
A process that should be removed rather than automated.
Digital Transformation
Digital transformation for commerce businesses: modernising legacy stacks into API-first systems, with phased migration sequenced so trading continues.
Think twice when
The current system works and the constraint is commercial.
Next.js
Next.js commerce storefronts: rendering strategy, caching and revalidation, the GraphQL boundary, Core Web Vitals and SEO over Shopify and Magento back ends.
Think twice when
A slow theme, which is usually cheaper to fix than to replace.
Mixpanel
Mixpanel consulting and implementation: tracking plan and event schema, funnels and retention, identity resolution, audits of an existing setup, and governance.
Think twice when
Traffic and acquisition reporting, which GA4 does well and free.
eCommerce consulting
eCommerce Consulting
Most of the expensive mistakes in commerce are made in the first three weeks, before anyone writes code. Platform chosen before the sales process is understood, integrations scoped as an afterthought, a migration sequenced so the riskiest work lands in peak season.
Wrong choice when
If the decision is already made and funded, and the constraint is delivery capacity rather than direction, consulting is an expensive way to buy hands. Hire the build instead.
Ongoing ownership
Website Management
Most website problems are not technical failures. They are ownership failures: nobody decides when to update, nobody watches the thing that fails quietly, and the person who knows how to deploy is on holiday. Management is the arrangement that makes those decisions somebody's job before they become an incident.
Wrong choice when
This is the wrong engagement when everything runs on one platform and the work is genuinely that platform's upkeep. A single Magento store needs a patch cadence and a rehearsal environment, not an estate-level arrangement, and the platform maintenance pages describe that better and cost less. It is also wrong where there is an internal team that already has the discipline and simply needs capacity, which is a hiring question rather than a consulting one.
Store operations
eCommerce Management
Most stores are launched and then run by whoever has time. Catalogue quality drifts, promotions are set up by hand under deadline, stock accuracy is maintained by someone checking twice, and the reporting nobody trusts gets ignored. eCommerce management is treating the store as a trading operation with a rhythm, rather than as a project that finished.
Wrong choice when
This is the wrong engagement while the store is still being built, or while the platform itself is the constraint. A catalogue that cannot be merchandised because the attribute model is wrong needs the data model fixed first, and no amount of operational rhythm compensates for a platform the business has genuinely outgrown. It is also wrong for a business that wants campaigns run: that is marketing, and it is a different discipline with different people.
That is usually the right starting position. Most engagements begin as one of these and turn out to be another, which is what the first conversation is for.