Practical insights from building AI systems, running transformation programmes, and making businesses permanently more profitable. No theory — just what actually works.
Start Here
New to MarginOps? These four posts capture who we are and what we do.
Your marketplace isn't just a sales channel — it's a cost centre, a margin driver, and an operational liability. Here's what happens when you treat it as a margin lever.
Read →Vision AI that reads receipts and invoices, extracts every line item, and syncs directly to your accounting system. Practical AI that saves finance teams hours every week.
Read →Your churn rate is a vanity metric. What matters is the margin impact — and most subscription businesses can't calculate it because their data is spread across 4 disconnected platforms.
Read →The step-by-step process for auditing every line of your P&L — from COGS to cloud costs to pricing — and turning what you find into an EBITDA-tracked action plan in two weeks.
Read →The decision to go all-in on Claude wasn't about hype. It was about reasoning quality, enterprise trust, and agentic capabilities that let us build production systems — not demos.
Read →Overlapping tools, unaudited cloud bills, enterprise SaaS you barely use, unoptimised databases, and no idea what tech adds to your cost-per-order. Sound familiar?
Read →Agent teams, 1M token context, and finance benchmark leadership. How Opus 4.6 transformed our transformation work.
Read →Most AI initiatives in retail fail because they start with technology, not business problems. A no-nonsense guide for CEOs, COOs, and CFOs who want results, not slide decks.
Read →A journey through every Claude model generation and the specific capabilities that unlocked new consulting work at each step.
Read →A transparent, week-by-week breakdown of how we run a margin transformation — from the initial P&L audit to full handover with your team running the programme independently.
Read →The actual architecture behind deploying 7 production agents: Claude Code, Agent SDK, MCP for data connectivity, and the coordinator pattern for multi-department analysis.
Read →You don't need a machine learning team. You need your existing people to understand what AI can do and how to use it. Here's the six-week approach that actually works.
Read →Most transformation programmes fail because nobody tracks them. Here's how I build a programme where every initiative has an owner, a go-live date, and an EBITDA value.
Read →16.4 million customer profiles, RFM scoring, churn prediction, and discount suppression — all running on Snowflake for a fraction of the vendor cost.
Read →Cloud infrastructure, SaaS sprawl, fulfilment, marketing, pricing, headcount, and technical debt. Each one bleeds 1-3% margin. Together they shift EBITDA by 3-5 points.
Read →Keyword search treats "blue cocktail dress for wedding" as four separate terms. Neural search understands intent. Here's how we deployed it.
Read →465GB across five production databases. 257GB of recoverable space. And six index fixes that saved 12.2 CPU days of processing.
Read →Automation-first headcount optimization across 8 departments over 8 months. The principle: automate the work before removing the role.
Read →Welcome sequences, browse abandonment, cart recovery, VIP nurture — 190 automated flows replacing manual campaign builds.
Read →£38K/month to £15K/month in three months. Multi-cloud migration from AWS to Hetzner with CDC pipelines and zero downtime.
Read →No new warehouse. No new WMS. Just 14 targeted operational changes that delivered £1.02M in annualized savings.
Read →A £75K pricing vendor and a £45K CDP — both replaced with in-house builds that outperform the originals. Here's the economics.
Read →PE portfolio companies keep hiring AI teams without operational context. Here's what actually works for extracting value.
Read →90,000 customers identified who would buy at full price. Four-tier discount suppression. £8.3M in margin protection.
Read →Replacing a £75K/year vendor with an in-house engine. Elasticity modelling, real-time monitoring, and results by week five.
Read →AI chatbot handling first-line queries, proactive issue detection, and systematic CX improvements. No extra headcount needed.
Read →The gap between AI strategy and AI implementation is where most consulting spend goes to die. Here's what's different about operating.
Read →I'll review your P&L, tech stack, and operations to identify where AI can make a permanent difference.
Stay updated
Practical AI and margin improvement insights. No spam.