Notes from inside the enterprise IT stack.
Everything I have written down, in one place. The guides answer one question step by step - a checklist you can run. The essays argue a position from the operator seat. Real T-codes, real article numbers, real failure modes. All free to read.
Guides
13 liveAnswer-first. Each one takes a single question an operator actually gets asked, and works through it as a checklist you can run this week.
The ERP modernisation business case - the honest case, build-vs-buy, and the cost of not acting
How to build an ERP / S/4HANA modernisation business case that survives the steering committee: the honest obligation-driven case (not the inspirational one), the five modernisation options each with its trap, the brownfield-vs-greenfield-vs-selective decision, a five-year TCO with the data-remediation line, the weighted decision scorecard, and the cost-of-not-acting frame anchored to the 2027 ECC maintenance deadline.
Procurement & Vendor ManagementIT vendor & SaaS TCO - the true three-year cost, not the invoice price
What an IT vendor / SaaS total-cost-of-ownership model must contain: the licence line, the hidden bolt-ons and admin FTE, the three-year cost with the renewal uplift modelled, and the renewal mechanics (notice period, auto-renewal trigger, list-vs-floor price, the vendor's fiscal-quarter close) that decide what you actually pay. From the distributor side of the table.
Transformation & SSC/GBSThe SSC/GBS transition plan - FTE sizing, knowledge transfer, and the exit to BAU
What a shared-services, GBS or BPO transition plan must contain: the FTE sizing math (volume / productivity x ramp uplift), the four transition models, the four-stage knowledge transfer that advances on evidence, the governance and self-writing status, the hypercare-to-BAU exit criteria, and the retained organisation. From an operator who has run offshoring transitions, not pitched them.
Career & RolesThe Job Decoder - 6 questions that tell you what a role actually is
Most job descriptions are written for the applicant-tracking filter, for HR and legal, and for the internal pay grade - not to tell you what the work is. The Job Decoder is the six-question method to understand any role in about a minute: what they do all day, what they're really measured on, who they work with, the number-one failure mode, the skills that actually matter, and how to break in. Written by an operator who has hired into and worked across enterprise-operations roles on both sides of the table.
Transformation & SSC/GBSThe PMO status report that doesn't lie - defensible RAG, RAID, and gates
Most programme status reports are watermelons: green on the outside until they turn red overnight. A defensible PMO status has written RAG definitions everyone agrees on, a RAID log with owners not just entries, stage gates that are pre-agreed pass criteria instead of date theatre, and numbers that assemble from the trackers rather than being narrated. Written by an operator who has run transformation PMOs - here is how to make a status report nobody can argue with.
Transformation & SSC/GBSHow to size an offshoring or SSC transition - the FTE math nobody teaches
Most transitions are sized with optimism, not arithmetic - and the new team is understaffed for its first six months, exactly when its reputation is set. The real method: Target FTE = annual volume ÷ productivity, then a ramp-inefficiency uplift you actually fund; choose the transition model deliberately; run knowledge transfer in four evidence-gated steps; budget the dual-running window; and declare done at BAU exit criteria, not the go-live date. Written by an operator who has run SSC and offshoring transitions.
Transformation & SSC/GBSWhy transformations fail - the 7 failure modes and the operating spine that prevents them
Transformations almost never fail on the technology. They fail in seven predictable, organisational ways: no honest baseline, maturity scored to flatter, everything-is-Wave-1, strategic builds before quick wins, no operating model for the after, governance by status deck, and value asserted but never proven. The fix is an operating spine - assessment, capability maturity, roadmap, target operating model, governance, value. Written by an operator who has run transformation programmes.
AI Adoption & GovernanceAre you ready to deploy an AI agent? The 12-point operator readiness check
A free 12-point readiness check for deploying an AI agent on real enterprise data: the identity, completeness, consistency, governance and process questions to answer before go-live - because an agent amplifies whatever your data already is. From an operator who builds master-data and AI-readiness tooling.
Regulatory & ComplianceThe DORA Article 28 ICT register - what it must actually contain
What a DORA Register of Information has to capture under Article 28: every ICT third-party arrangement, which critical or important function each one supports, the criticality and sub-contracting chain, and the concentration-risk picture - not just a procurement list. An operator's view of the register supervisors actually open.
Regulatory & ComplianceThe EU AI Act deployer checklist - what an SME has to do before August 2026
A practical EU AI Act readiness checklist for deployers: build the AI system inventory, classify each system (prohibited, high-risk under Annex III, limited-risk, minimal), meet the Article 50 transparency obligations, cover the Article 26 deployer duties, and assemble the evidence - treated as a documentation exercise, not a legal one.
Master Data OperationsThe master data quality audit - the checks that actually matter
How to run a master data quality audit: the six dimensions worth scoring (completeness, validity, consistency, uniqueness, conformity, timeliness), how to score at field level, how to log and prioritise defects, and how to evidence the improvement over time - for the person accountable for the data, not the consultant reviewing it.
SAP / ERP OperationsThe S/4HANA cutover checklist - the run sheet that survives 3 a.m.
What an S/4HANA migration cutover checklist must contain: the readiness gates, the hour-by-hour run sheet, the master-data freeze, the data-validation checks (MASS/MASSD, MB52, FBL3N), the pre-agreed rollback criteria, and the hypercare plan. From an operator who has watched cutovers fail at 4 a.m.
Procurement & Vendor ManagementThe SaaS renewal negotiation checklist - terms that beat the discount
How to negotiate a SaaS or software renewal from the buyer's seat: find the notice period and start 90–120 days out, pull real usage, build the true 3-year TCO, read the vendor's leverage (quarter-end, deal registration, your renewal date in their CRM), establish a credible alternative, and counter the uplift with specifics. Written by someone who has run renewals from the vendor and distribution side.
Essays
9 liveWhere I argue a position rather than hand you a checklist - what I watched go wrong, and what I think it means.
Governance, not agents: the operating model most AI programs skip
Most AI agent projects don't fail on the model - they fail on the operating model: no decision rights, no use-case gate, broken data, runaway cost, no owner after launch. Why the first question isn't 'which agent should we buy' but 'who owns this, and how do we decide', and what an AI Center of Excellence actually is - a thin function with a charter, a GO / PILOT / PARK gate, eval standards, a risk register, and run-and-maintain ownership across the agent lifecycle.
Touchless ordering is a master-data achievement nobody sees
An operator's account of turning manual product creation into an AI-driven, rules-based, fully-automated master-data engine at distribution scale - ingestion to touchless quoting, the governance that holds it, and what actually breaks. From someone who came to master data from the reporting side, where the drift shows up first.
Your AI agent will amplify your data problems, not fix them
An AI agent is an amplifier - point it at fragmented, duplicated, or stale data and it scales the confusion, fluently. Why bad data no longer looks broken but looks persuasive, why enterprise systems were never built for autonomous decisions, and the three readiness questions - identity, completeness, source of truth - to answer before you deploy an AI agent on anything that matters.
Vendor-side observations on how enterprise IT procurement actually works
What the vendor and distribution side actually does when your IT spend comes up - whether it's a competitive RFP or a quiet renewal. The signals they read: is the competition genuine, who controls the channel and deal registration, the executive relationship map, your own internal fault lines, whether you can operationally switch - and the two renewal mistakes that cost more than any discount saves.
The cutover-window decision nobody makes early enough
Every S/4HANA migration has a cutover window - and the rollback decision inside it is the hardest call in the whole project, not technically but psychologically. The three questions the best teams answer before T-Day: the written go/no-go criteria, who holds the call, and where the point-of-no-return actually sits.
Three IT procurement mistakes I've watched repeat from the vendor side
From 15 years on the vendor-distribution side of enterprise IT deals: the three renewal mistakes that repeat across organisations - starting the conversation at T-30 instead of T-90, negotiating price instead of contract terms, and never asking what the distribution channel already knows - and what the best procurement teams do instead.
What supervisory authorities actually look for in DORA and EU AI Act compliance
Modern EU ICT and AI regulation is no longer a document-based compliance exercise - it's a governance-maturity assessment. What that means in practice for EU AI Act deployment governance (the four controls that make you defensible) and DORA ICT third-party risk (why your register is a procurement inventory, not a concentration-risk picture), and the shift underneath both.
The S/4HANA deadline math most organizations calculate wrong
SAP ECC mainstream maintenance ends December 31, 2027 - but for most mid-market organizations the real cut-off was mid-2025, because transformation capacity is finite. The lead-time math, the capacity crunch, why procurement misreads ERP sourcing, the readiness signals that beat RAG status, the rollback decision nobody pre-commits, and why hypercare is where the program is actually judged.