The knowledge portal
Everything we know
about making delivery work.
79 long-form articles across 19 domains. Free, ungated, and written for people who have to make these decisions rather than present them. When we are wrong, we revise the page and say so.
Domains
Where to start
8
AI-Native Delivery
What happens to a delivery system when writing code stops being the expensive part. Transaction cost, batch economics, and why verification becomes the constraint you have to engineer.
Move faster because the flow physics say to, not in spite of them.7
Agentic Systems
Systems that choose their own next step. Architecture patterns, tool design and MCP, context engineering, trajectory evaluation, and the operational reality of software that acts on its own.
Know when autonomy earns its unpredictability — and when it does not.5
AI & ML Engineering
Building and shipping systems whose behaviour is probabilistic: LLM application architecture, retrieval, evaluation, agents, inference economics and the operational discipline underneath.
Ship a non-deterministic system you can still be accountable for.5
Blockchain & Web3
Delivery where deployment is irreversible and bugs are withdrawals: smart contract lifecycle, audit economics, upgradeability, key management and the on-chain boundary decision.
Iterate on a system you cannot patch after the fact.5
Mobile & Cross-Platform
App store review as a hard release constraint, staged rollouts, the native versus cross-platform economics, and how to keep continuous delivery when a third party gates every release.
Run continuous delivery through a gate you do not control.8
Flow Beyond Engineering
The same queueing arithmetic, pointed at the rest of the business: hiring, procurement, legal review, the month-end close, marketing production and support. Little’s Law does not know what is in the queue.
Diagnose any function in the business the way you would a delivery team.6
Leadership & Decisions
Decision latency as a delivery metric, decision rights, operating cadence, and the uncomfortable fact that executives create most of the queues they later complain about.
Stop being the constraint in your own organisation.3
Agile Foundations
Where agile came from, what the Manifesto actually says, and the difference between the empirical ideas that hold up and the folklore that accreted around them.
Argue from first principles instead of quoting a framework.4
Scrum in Practice
The framework as written, the framework as practised, and the specific places the two diverge badly enough to cost you a quarter.
Run Scrum without the cargo cult.3
Kanban & Flow
Queueing theory for people who ship software: work in progress, cycle time, batch size, and why your team is slow despite being busy.
Diagnose a slow system by its queues rather than its people.3
Roles & Teams
What a Product Owner is accountable for, what a Scrum Master is actually worth, how to design team boundaries, and what happens to managers.
Staff a team so the accountabilities do not collide.3
Scaling & Portfolio
SAFe, LeSS, Scrum@Scale, Spotify-as-myth, and the dependency mathematics that decides whether any of them will help you.
Choose a scaling approach for a reason you can defend.3
Metrics & Measurement
DORA, flow metrics, SPACE, and how to build a measurement system that survives contact with an executive who wants a single number.
Report delivery performance without inviting gaming.3
Engineering Practice
Trunk-based development, continuous delivery, test strategy, technical debt as an economic decision. The practices without which agility is theatre.
Make the technical changes that agility actually depends on.2
Product & Discovery
Backlogs, outcomes over output, continuous discovery, opportunity solution trees, and how to say no to a roadmap you were handed.
Decide what to build, with evidence, at the pace you ship.3
Transformation & Change
Operating-model change at organisational scale: sequencing, funding models, resistance, coaching structures and exit criteria.
Run a change programme that leaves capability behind.3
Distributed & Offshore
Time-zone design, asynchronous defaults, vendor and captive-centre models, and the handover practices that decide whether distributed delivery works.
Run a multi-site team that is not slower than a single-site one.2
Governance & Compliance
Agile inside regulated environments: audit evidence, segregation of duties, change advisory boards, contracts and procurement.
Pass the audit without stopping the delivery.3
Antipatterns
Named failure modes with their symptoms, root causes and treatments. The diagnostic manual for organisations that already "do agile".
Recognise the failure you are living in and name it.Full index
Search everything
Filter by domain and depth. Nothing here is gated, and nothing links to a form you have to fill in to keep reading.
Level
79 of 79 articles
- Agentic9 minAgent Architecture PatternsSingle loop, prompt chain, router, orchestrator-worker, parallel fan-out and multi-agent. What each pattern is genuinely for, the coordination cost that multi-agent quietly reintroduces, and a selection table you can argue from.Practitioner
- Governance11 minAgile Contracts And ProcurementFixed-price fixed-scope contracts reliably produce the behaviours both sides claim to hate. A practical look at the structures that work, what procurement is genuinely protecting, and how to write a statement of work for iterative delivery.Advanced
- Governance11 minAgile In Regulated EnvironmentsRegulators ask for evidence of control, not for stage gates. How segregation of duties, traceability and change approval can be met by an automated pipeline, and how to make audit an output rather than a project.Advanced
- Foundations9 minAgile Vs Waterfall, The Honest ComparisonAn even-handed account of where sequential delivery genuinely outperforms iterative delivery, where it does not, why Royce is routinely misquoted, and how to choose per workstream rather than per organisation.Foundational
- Distributed12 minAsynchronous First DeliveryWriting first is a discipline with mechanics, not a personality trait. Decision records, RFCs, timeboxed asynchronous decisions, recorded demos, and how to stop async degrading into nobody replying for four days.Practitioner
- Blockchain10 minAudit Scheduling And The Code FreezeAn external audit is a scarce dependency with a lead time, not a stage gate. How to book it, scope it, prepare for it, survive the freeze, triage findings, and plan the re-review as cost rather than surprise.Practitioner
- Flow10 minBatch Size And The Cost Of Big ReleasesBatch size is the variable with the widest reach in delivery, and the one least often managed deliberately. The economics of why smaller is faster, safer and cheaper, and how to shrink batches when the release path fights you.Practitioner
- Scaling11 minChoosing A Scaling FrameworkAn even-handed comparison of SAFe, LeSS, Scrum@Scale, Nexus and Team Topologies, plus the option nobody costs properly — reducing coupling so that you need less coordination in the first place.Advanced
- AI Delivery9 minCode Review When Code Is CheapReview practice under high change volume. What a reviewer should look for when the author did not type the code, why pull request size discipline matters more than before, and who owns a defect nobody wrote by hand.Practitioner
- Agentic10 minContext Engineering And Agent MemoryContext is a scarce, expensive, actively managed resource rather than a place to put everything. Working versus persistent memory, what compaction loses, sub-agent isolation, and why long-running agents degrade.Advanced
- Mobile10 minContinuous Delivery Through App Store ReviewStore review is the one place a third party sits between your merge and your users. It is a fixed cost in the batch-size equation, not a wall, and the monthly cadence it seems to force is a choice.Practitioner
- Product10 minContinuous Discovery For Delivery TeamsDiscovery as a weekly habit rather than a phase. Customer contact, opportunity solution trees, assumption mapping, and how to run discovery alongside delivery without rebuilding a miniature waterfall.Practitioner
- Leadership10 minCulture Is A Lagging IndicatorCulture-change programmes fail because culture is an output, not an input. Westrum's typology, why psychological safety is produced by how incidents are handled, and how a leader changes culture by changing what is measured, funded and promoted.Practitioner
- Leadership10 minDecision Latency Is A Delivery MetricThe elapsed time between a question being raised and an answer being given is a queue, and usually the largest one nobody measures. How to make it visible, and why a leader's calendar constrains the whole organisation's throughput.Foundational
- Leadership10 minDecision Rights And Who Breaks A TieMost decision delay is ambiguity about who decides rather than difficulty in deciding. How to map decision rights honestly, why consensus and consultation get confused, and what to do when two executives both believe they own the same call.Practitioner
- Scrum10 minDefinition Of Done As A ContractTreated properly, the Definition of Done is your organisation's quality floor and the only enforceable meaning of "releasable". How to encode non-functional, security, accessibility and compliance requirements, and why per-team definitions fragment a product.Practitioner
- Blockchain10 minDelivering Software You Cannot PatchContinuous delivery assumes you can ship a fix. On-chain, you cannot, and a defect is a withdrawal of funds. The answer is not waterfall — it is moving iteration upstream and making the irreversible step as small as possible.Foundational
- Agentic10 minDesigning Agents With Bounded AuthorityThe design question for an agent is not which model it uses but what it is permitted to do. Authority boundaries, idempotency, blast-radius limits, prompt injection as a live threat, and audit trails that hold up.Advanced
- Distributed11 minDesigning Timezone OverlapOverlap is a quantity you design, not a weather condition you suffer. How much you need, what it is actually for, why follow-the-sun rarely works for feature work, and how to write it into a statement of work.Practitioner
- Metrics10 minDORA Metrics Without The GamingThe four key metrics are the best-evidenced delivery measures available, and they break the moment anyone is rewarded for them. Precise definitions, the specific gaming each invites, and how to report them so they survive contact with executives.Practitioner
- Foundations10 minEmpiricism Is The Engine Under Every FrameworkTransparency, inspection and adaptation are what actually make iterative delivery work. Here is why empirical control exists, the four conditions under which it fails, and where predictive planning is the correct choice.Practitioner
- Agentic9 minEvaluating AgentsThe unit of evaluation is a trajectory, not a response. Outcome versus trajectory assertions, step and cost budgets as first-class checks, replay harnesses, and why an agent that succeeds expensively has still failed.Practitioner
- AI Engineering9 minEvaluation Harnesses For LLM SystemsThe piece almost every team skips. How to build a golden set, design a rubric, use LLM-as-judge without being fooled by it, and turn "is this better?" from an argument into a measurement.Practitioner
- Enterprise Flow10 minEvery Value Stream Is A QueueThe queueing argument that explains slow software delivery is not about software. It is operations research, and it applies unchanged to hiring, finance, procurement, marketing and support.Foundational
- Enterprise Flow10 minFailure Demand And What Your Support Queue Is Telling YouJohn Seddon's distinction between value demand and failure demand turns a support queue into a defect signal for the whole business. How to classify it, who should own the fix, and why handle time is the wrong target.Advanced
- Mobile10 minFeature Flags And Staged Rollout On MobileOn mobile you cannot roll back, so flags are not a convenience — they are the safety mechanism that replaces rollback. Remote config architecture, kill switches, default-off discipline, flag debt, and rollout halts that fire without a human.Practitioner
- Metrics10 minFlow Metrics A Practical GuideFlow time, throughput, flow efficiency, flow load and flow distribution — what each one tells you, how to instrument them from an ordinary Jira or Linear board, and why percentiles and Monte Carlo beat averages and estimates.Practitioner
- AI Delivery10 minForecasting When Throughput ChangesHistorical throughput predicts the future only while the system is stable. What to do when capability is shifting under you — re-baselining, estimation practice, which DORA and flow metrics survive, and how to report it upward honestly.Advanced
- Transformation11 minFrom Project Funding To Product FundingThe funding model is the binding constraint on agility in most organisations. Why annual project budgets force temporary teams, how persistent product funding works, and how to hold the governance conversation with a CFO.Advanced
- Leadership10 minHow Executives Create QueuesThe specific, well-intentioned executive behaviours that generate delay downstream — mid-quarter reprioritisation, being the single approver, the pet project, utilisation targets and demands for certainty. The mechanism in each case, and the cheaper alternative.Advanced
- AI Engineering10 minInference Cost As An Architectural ConstraintModel-backed features carry a variable cost that rises with success. Context budgeting, caching, routing and cascades, batching, and the unit economics you need before the architecture hardens around you.Advanced
- Enterprise Flow10 minMarketing Production And Campaign FlowCreative production is a value stream whose dominant wait state is approval. How to design review rather than resent it, why simultaneous campaigns slow all of them, and how to measure output without gaming it.Practitioner
- Mobile10 minMobile Release Health And ObservabilityWhat to instrument on a client you cannot attach a debugger to, which signals deserve an automated rollout halt, and how to tell a real regression from noise when early updaters skew every number.Practitioner
- Mobile10 minNative Versus Cross Platform EconomicsThe shared-code decision is an economic one, not a technical preference. What genuinely shares, what never does, the residue that eats the savings, and how to make the decision once so it is not relitigated every quarter.Advanced
- Distributed12 minOffshore Engagement Models ComparedStaff augmentation, managed pods, project outsourcing, build–operate–transfer and captive capability centres. What each optimises for, how vendor margin shapes behaviour, and how to choose without discovering the trade-off two years in.Advanced
- Leadership11 minOrganisational Design Above Team TopologiesTeam-level design is well covered. The layer above it — divisions, shared services, centres of excellence, matrix reporting — is where most coordination cost is created. Where boundaries should fall, what each one costs, and why reorganisations destroy value.Advanced
- Product11 minOutcomes Over Output In PracticeHow to move from a feature roadmap to outcome commitments without losing the trust of stakeholders who want dates. Writing measurable outcomes, choosing leading indicators, and the ways OKRs fail.Practitioner
- Enterprise Flow10 minProcurement And Legal Review As ConstraintsThe two queues that silently gate everything else. Why adding buyers and lawyers rarely helps, how risk tiering and pre-approved templates reduce work in progress, and how to measure a function whose job is partly to say no.Advanced
- Enterprise Flow7 minPublished Commitments And The Debt You Cannot RefactorTechnical debt can be refactored because you own both sides of it. Anything you have published — an API, a price, a filing, a plot point, a licence — is coupling to people outside your control, and it only ever accumulates.Advanced
- AI Delivery9 minPutting AI In The Delivery PipelineWhere assistants belong in the lifecycle beyond code completion — test generation, large-scale migration, decision capture, triage, review assistance — which of these pay back immediately, which are traps, and what provenance a regulated estate has to evidence.Practitioner
- AI Engineering10 minRAG Architecture That Survives ProductionRetrieval quality sets the ceiling on answer quality, and chunking decides retrieval quality. Hybrid search, reranking, grounding and citation, re-indexing, and how to tell which half of the system broke.Practitioner
- Transformation10 minResistance Is InformationPeople who push back on a change usually have reasons. Treating resistance as data rather than obstruction surfaces risks you missed, and occasionally tells you the change should stop.Practitioner
- AI Delivery10 minRunning A Fleet Of Coding AgentsSeveral coding agents working a codebase in parallel. Decomposition so they do not collide, worktree isolation, why review capacity becomes the ceiling almost immediately, and the failure modes that are easy to miss.Advanced
- Agentic10 minRunning Agents In ProductionThe operational surface of a system that acts on its own. Sandboxing, durable execution, circuit breakers and runaway cost, rate limiting, trajectory observability, on-call, and how you stop one mid-run.Advanced
- Scrum9 minScrum As Written Vs Scrum As PractisedThe framework is a short document with very few rules. Most organisations keep the meetings and discard the mechanisms. Here are the five divergences that cost real money, and how to tell which ones matter.Practitioner
- Transformation11 minSequencing An Agile TransformationMost transformations fail on order of operations rather than on intent. Diagnose before intervening, fix structure before ceremony, engineering before scaling, and define the exit before you start.Advanced
- Blockchain10 minSmart Contract Security As A Delivery PracticeSecurity fails when it is an end-stage gate. How to make threat modelling, stated invariants, fuzzing, forking, static analysis and adversarial review part of the ordinary development loop, and where formal methods earn their cost.Advanced
- Scrum10 minSprint Planning That Is Not TheatreHow to run planning so it produces a goal the team can act on rather than a number somebody will be held to. Covers sprint goals, capacity versus forecast, right-sizing instead of points, and the usual failure modes.Practitioner
- Engineering11 minTechnical Debt As An Economic DecisionRecovering the original meaning of the metaphor, distinguishing the four kinds of debt, and making the interest payment visible in terms a finance function can act on. Including the cases where not paying it down is the correct answer.Advanced
- Engineering11 minTest Strategy For Teams That Ship DailyA test suite is a portfolio of instruments bought to answer different questions, not a coverage number to be maximised. How to design one for daily deployment, including flakiness, contract testing and the legitimate role of testing in production.Practitioner
- AI Engineering9 minTesting Non-Deterministic SystemsAssertion-based testing assumes one correct output. Model-backed systems do not have one. What replaces equality checks, where deterministic tests still belong, and what the test pyramid becomes.Practitioner
- Foundations9 minThe Agile Manifesto, Read ProperlyThe Manifesto is sixty-eight words plus twelve principles, and almost nobody who cites it has read it recently. What it actually says, what it was arguing against, and which parts have aged.Foundational
- Antipatterns10 minThe Agile Theatre DiagnosticA short set of sharp questions whose answers reveal whether delivery has actually changed or whether the organisation has only changed its vocabulary, plus the order in which to treat what you find.Foundational
- AI Engineering9 minThe Demo To Production GapA convincing prototype is a small fraction of the work. What sits between a demo that worked and a system you can operate, and how to reset executive expectations after they have seen the demo.Foundational
- Scaling11 minThe Dependency Mathematics Of ScalingCoordination cost grows faster than team count. This is why, how to measure the dependencies you actually have, and why removing them beats managing them almost every time.Advanced
- Antipatterns11 minThe Lighthouse Team TrapThe protected pilot team, exempted from the rules everyone else lives under, produces a success that cannot transfer and a political aftermath that makes the next attempt harder. What to do instead.Practitioner
- Enterprise Flow10 minThe Month End Close As A Batch ProblemThe close is the largest batch in most businesses. Treating it as a batch size problem rather than a resourcing problem explains why it takes as long as it does and what actually shortens it.Practitioner
- Blockchain10 minThe On-Chain BoundaryDeciding what genuinely belongs on-chain and what is there for narrative reasons. Cost, latency, privacy and immutability as the four axes, the trust that oracles reintroduce, and the ordinary software that makes up most of a working system.Practitioner
- Leadership10 minThe Operating CadenceThe meeting system is infrastructure, not overhead. What genuinely needs to be synchronous, why a weekly forum adds four days of average wait to every decision routed through it, and how to audit a calendar without losing coordination.Practitioner
- Scrum9 minThe Retrospective ProblemRetrospectives decay into complaint or silence for structural reasons, not cultural ones. How to retro on the system rather than the team, what to do about actions nobody owns, and when to stop running them.Practitioner
- Roles9 minThe Scrum Master Value ProblemThe role is under commercial pressure because too many people in it do work that is not worth the money. An honest account of what a good one does, what a bad one does, how to tell them apart, and where the career goes next.Practitioner
- Scaling9 minThe Spotify Model Was Never A ModelSquads, tribes, chapters and guilds were a description of one company at one moment, published by people who later said it was not a template. Here is what gets lost when the org chart travels without the engineering culture.Practitioner
- AI Delivery9 minThe Transaction Cost CollapseAI-assisted engineering collapses the cost of producing a change. It does not collapse the cost of verifying one. This is the economic argument for what that does to batch size, and where the constraint moves next.Foundational
- Agentic10 minTool Design And The Model Context ProtocolThe quality of an agent is mostly the quality of its tools. How to design interfaces for a model rather than a human, why error messages are instructions, and what MCP standardises without solving.Practitioner
- Engineering9 minTrunk-Based Development And Why It MattersBranching strategy is a batch-size decision with organisation-wide consequences, not a matter of developer taste. Why merge cost rises with branch age, what replaces the long-lived branch, and what must be true before trunk-based development works.Practitioner
- Blockchain10 minUpgradeability Is A Trust DecisionEvery upgrade mechanism is a trust assumption imposed on users. Proxy patterns and their trade-offs, timelocks and signer topology, emergency pause, and how to document honestly what you are and are not guaranteeing.Advanced
- Metrics10 minVelocity Is Not A Performance MetricStory points were invented to stop teams being compared on hours. Reporting velocity upward reintroduces exactly the problem they were meant to solve, and the damage is predictable. What to report instead.Foundational
- AI Delivery9 minVerification Is The New BottleneckWhen producing a change is cheap, the constraint becomes establishing that it is correct. What that means operationally for review capacity, test confidence, and the shift from writing code to specifying and checking it.Practitioner
- AI Delivery7 minVerification Without An OracleVerification is the constraint once AI makes production cheap. In software it at least has an oracle — tests, types, telemetry. In creative and editorial work there is none, and the arithmetic gets much worse.Advanced
- Antipatterns11 minWater Scrum FallThe most common enterprise delivery shape. Requirements and design settled upfront, iterative build in the middle, batched release and hardening at the end. How to recognise it, why it is rational, and how to dismantle it.Foundational
- Roles10 minWhat A Product Owner Is Accountable ForThe role carries one accountability and it is not backlog administration. What authority it requires to function, why proxy product owners fail, and how the job changes when several teams share one product.Practitioner
- AI Delivery9 minWhat AI Does Not ChangeThe corrective piece. Little's Law, dependency structure, Conway's Law, approval queues and funding models are all untouched. Coordination cost and conceptual integrity get harder at volume, not easier.Foundational
- Roles10 minWhat Happens To Middle ManagementThe hardest unanswered question in most transformations. Why self-organisation is heard as redundancy, which management jobs become more valuable rather than less, and how to sequence the change so managers become allies.Advanced
- Agentic9 minWhat Makes A System AgenticThe word covers everything from a scripted pipeline with a model in it to open-ended autonomy. The distinguishing property is who chooses the next step, and giving that away is a cost you should be able to justify.Foundational
- Enterprise Flow7 minWhen The Form Sets Your Batch SizeThe advice to reduce batch size assumes batch size is yours to choose. Sometimes the medium imposes a floor — you cannot ship a quarter of an episode, half a filing or a partial print run. What changes is the advice, not the economics.Practitioner
- Flow6 minWhy Busy Teams Are Slow TeamsUtilisation and speed are opposites past a certain point. This is the queueing argument that explains why adding people to a loaded system makes everything later, and what to do instead.Foundational
- Flow10 minWork In Progress Limits In PracticeEveryone agrees WIP limits work. Almost nobody sets them correctly. This is where to place the limit, what to do the first time it bites, and how to survive the political conversation that follows.Practitioner
- Enterprise Flow9 minYour Hiring Pipeline Is A Queueing SystemTime-to-hire is a distribution, not an average. Open requisitions are work in progress, the interview loop is usually the constraint, and scheduling is a batch delay you can design away.Practitioner
- Mobile10 minYour Old Versions Never DieEvery binary you have shipped is still running somewhere. Version fragmentation is a permanent condition, not a transitional one — and every API change becomes a compatibility negotiation with your own past.Practitioner
The library is the argument. The engagement is the work.
If any of this matches what you are living through, the fastest next step is a conversation rather than another article.