PRINCE2
PRINCE2 is a structured method for governing projects, built around one question a project must keep answering: is it still worth funding? It says almost nothing about how the work gets done and a great deal about who may decide that it continues. The page opens with one project set up the PRINCE2 way, finished, then covers what the method is, its seven principles, seven practices and seven processes, tolerances and exceptions, two projects it exists to stop, and where it degrades.
- Current version
- 7th edition, 2023
- Structure
- 7 principles · 7 practices · 7 processes
- Core question
- Still justified?
A PRINCE2 project on one screen — finished
Most people who look up PRINCE2 want to see what a project run that way actually looks like. Here is one: the retailer replacing its online shop from the WBSWork Breakdown Structure — the project’s scope as a treeA tree of everything a project will deliver, broken down until each piece is small enough to estimate and give to one owner — and, by definition, nothing the project will not deliver. Its children must add up to exactly their parent, which is what makes it a scope boundary rather than a task list.Full article →All terms page — nine months, a fixed priceagreed scope, agreed sumA contract where an agreed scope is delivered for an agreed sum, so the supplier carries the cost of anything that takes longer than planned. That is why fixed-price work comes with the strictest change control: every change to the scope is a commercial event, not just a planning one.All terms with an agency — as its project board approved it on Monday 3 February 2025. These are the decisions that matter from its project initiation documentation, the file PRINCE2 calls the PID. Later sections cite the lines by number.
1.Why — the business case
- Reason: the old platform’s contract ends on 31 October 2025, and renewing it costs €380,000 for another year
- Benefits: €240,000 a year less to run the shop, and conversion from 2.1% to 2.4% within six months of launch — about €550,000 a year in gross margin
- Costs: €1.2 million to build — €950,000 fixed price to the agency, €250,000 of the retailer’s own — and €140,000 a year to run
- Dis-benefit, said out loud: customer service handles orders about a fifth slower for the first four weeks, while it learns the new back office
- What would end the project: a forecast launch after 31 October, when the renewal falls due and the first year’s saving is gone; or usability tests in stage 2 showing checkout completion no better than today’s
2.Who decides
- Executive: the retailer’s commercial director — owns the business case and the money; one person, never the project manager
- Senior Users: the head of e-commerce, for conversion, and the head of customer service, for adoption and the dis-benefit in 1.4
- Senior Suppliers: the agency’s delivery director, and the retailer’s head of IT for hosting and the ERP feed
- Project manager: on the retailer’s payroll, not the agency’s, running the project day to day inside the stage’s tolerances
- Team managers: the agency’s two team leads, and the retailer’s payments team lead, who owns WBS line 4.3
- Assurance, none of it reporting to the project manager: finance for the business, the customer-service quality lead for users, an independent architect for the supplier side
- Change authority: the project manager up to €5,000 a change, the Executive with the Senior Users up to €20,000, the board above that — from a change budget of €40,000
- Project support: a PMOProject Management Office — the office that sets project standardsA unit that sets project-management standards across an organisation, collects reporting from its projects and sometimes employs the project managers. Its value depends on which of those it does: a PMO that only gathers status reports adds a layer without adding a decision.All terms analyst keeps the registers and logs — never the same person as assurance
3.Stages — the board authorises one at a time
- Starting up, December 2024: mandate from the leadership team on 2 December, project brief and outline business case, initiation authorised on 20 December
- Initiation, 6–31 January 2025: this file, the full business case and the project plan; project authorised on 3 February
- Stage 2, Foundations, February to April: environments, design, accounts and catalogue; the boundary asks whether usability tests support the conversion in 1.2
- Stage 3, Checkout and migration, May to July: checkout, card payments, the first full migration rehearsal; the boundary asks whether 15 September still holds
- Stage 4, Launch, August to October: tests, acceptance, cutover on Monday 15 September, hypercare to Friday 10 October, closure on Friday 24 October
- At every boundary the board makes one of three decisions: authorise the next stage, ask for an exception plan, or close the project early
4.Tolerances — the seven performance targets
- Time: launch on 15 September, no later than 29 September; each stage may end a week early or late
- Cost: the project up to 5% over its €1.2 million; each stage up to 10% over its own budget
- Scope: nothing in WBS branches 4, 5, 6 or 8 may move; up to two other work packages may finish within four weeks after launch
- Quality: checkout loads in under 2 seconds for 95% of requests at three times last Black Friday’s peak, and no severity-1 defect is open at go-live
- Benefits: conversion may land at 2.3% instead of 2.4% without the board; below that, the board re-decides 1.2
- Risk: no threat above €50,000 of exposure without an owner and a response; open threats together within the €80,000 risk budget
- Sustainability: the old shop’s data-centre racks switched off within 30 days of launch; product pages under 1.5 MB, from 4.2 MB today
5.How the board hears about it
- Checkpoint reports: each team manager to the project manager, every Friday
- Highlight report: the project manager to the board every second Friday — one page, a forecast against each line of group 4, the top three risks, the issues decided
- Exception report: within two working days of any forecast breaking a stage tolerance — the cause, the options, a recommendation
- The board meets at stage boundaries and when an exception arrives — not weekly; a forecast beyond project tolerance goes past it to the leadership team
6.Products — what is delivered, before how long it takes
- The project product: a new shop taking orders and card payments for the whole catalogue, with accounts and order history migrated — accepted by the Senior Users after user acceptance (WBS 7.4) and four weeks of hypercare (WBS 8.4)
- The product breakdown is the WBS’s thirty-one work packages; each product description is its dictionary entry, like 4.3 on the WBS page
- Work packages are authorised one at a time, each with its own tolerance — card payments, 4.3: 72 hours, give or take 10%
- A quality register records every check: which product, who checked it, against what, and the result
7.Learning, closing and after
- Lessons read before starting: the 2018 replatform migrated orders once, on the night, and lost two days of order history — so this plan has two full rehearsals
- The lessons log is read at every stage boundary; a lessons report goes to the PMO at closure
- Closing: acceptance confirmed, handover to operations, follow-on actions recommended — refunds through the checkout, asked for in month three, among them
- Benefits reviews at three and six months after launch, in December 2025 and March 2026 — owned by the Senior Users, reported to the Executive, after the project has stopped existing
8.Tailored for this project
- The agency’s teams run two-week sprints inside each work package; the sprint review is the checkpoint in 5.1
- No separate team plans: the sprint backlogs are the team plans
- Risks, issues and lessons live in one tool with three views, not in three documents
What PRINCE2 is — a method for deciding, not for building
PRINCE2 is often compared with Scrum, and the comparison confuses almost everyone the first time. The two are not alternatives. Scrum describes how a team turns a list of work into working software. PRINCE2 describes how an organisation decides to start a project, keeps checking whether it should carry on, and stops it when the answer changes. One is about building. The other is about deciding and funding.
The method’s own definition explains the rest. A project, in PRINCE2’s words, is “a temporary organization that is created for the purpose of delivering one or more business products according to an agreed business case”. Every part of that sentence is load-bearing. Temporary means it has a planned end and a closing process. Organization means roles and decision rights, not just a team. And an agreed business case means a written reason for spending the money — the reason the board keeps re-reading in line 1.5.
Five integrated elements
The 2023 edition describes the method as five elements that only work together. Four of them existed before; people is new as a named element, and the edition puts it at the centre of the diagram.
| Element | What it is | In the hook |
|---|---|---|
| Principles | Seven obligations; break one and it is not PRINCE2 | The shape of the whole file |
| People | Leading the team and the change, and how the project communicates | 1.4, 2.2, group 5 |
| Practices | Seven aspects managed all the way through; called themes until 2023 | Groups 1, 2, 4 and 6 |
| Processes | Seven sets of activities: who does what, and when | Group 3 |
| Project context | The organisation, the contract, the delivery method, the scale — what tailoring adapts to | Group 8 |
Four levels, and why the board is not a steering committee
PRINCE2 separates four levels of authority. The business layer, outside the project, commissions it: it issues the mandate, appoints the Executive and sets the project’s tolerances. The project board directs. The project manager manages, day to day. The teams deliver. Each level works inside limits set by the one above it, and reports to it in a fixed format — which is exactly what group 5 of the hook writes down.
The board looks like a steering committee and is a different thing. A steering committee is usually a forum: many people, an agenda, a status update, no single owner of the decision. A PRINCE2 board has one decision-maker, the Executive, advised by the people who speak for the users and the suppliers. It meets when there is something to decide, and its decisions have only three possible outcomes, listed in line 3.6.
Where it came from, and what it is called now
The name is an acronym: PRojects IN Controlled Environments. In 1989 the UK government’s computer agency, the CCTA, adopted a version of an earlier method, PROMPT II, as its standard for IT projects and called it PRINCE. PRINCE2 followed in 1996 as a method for any kind of project, was refreshed in 2009, updated in 2017, and rewritten as the 7th edition in 2023. Ownership moved from the Cabinet Office to AXELOS, a joint venture with Capita, in 2013, and PeopleCert bought AXELOS in 2021.
Two renamings are worth knowing before an interview or a tender. In 2023 the seven themes became the seven practices; most training material online still says themes, and it means the same seven things. And PeopleCert has since turned PRINCE2 into the name of a whole family. The method itself is sold as PRINCE2 Project Management, Version 7, with a Foundation exam of 60 questions and 36 needed to pass. MSP is now PRINCE2 Programme Management, MoP PRINCE2 Portfolio Management, and M_o_R PRINCE2 Risk Management.
The seven principles — and what each looks like when it is missing
The principles are the non-negotiable part. The method says so in one sentence: unless they are applied, “it is not a PRINCE2 project”. A project may drop documents, merge processes and combine roles and still be PRINCE2. If it breaks a principle, it is not, whatever its paperwork says. The last column is the useful one: it is what each principle looks like from the outside when it is gone.
| Principle | What it obliges | In the hook | What it looks like missing |
|---|---|---|---|
| 1. Ensure continued business justification | A reason worth the money at the start, and still worth it at every review; if not, stop | 1.5, 3.6 | A business case written for approval and never opened again |
| 2. Learn from experience | Look lessons up at the start, record them throughout, pass them on | 7.1, 7.2 | A lessons report written at the end that nobody reads at the start of the next project |
| 3. Define roles, responsibilities and relationships | Business, user and supplier interests all represented, and decision rights written down | Group 2 | Five people who each think someone else decides |
| 4. Manage by stages | Plan the next stage in detail, the rest roughly; authorise one stage at a time | Group 3 | One approval at the start for the whole budget |
| 5. Manage by exception | Agree limits for each performance target, then leave the level below alone inside them | Groups 4 and 5 | A board that wants weekly detail, or one that hears nothing until it is too late |
| 6. Focus on products | Agree what will be delivered and to what quality before how long it takes | Group 6 | A plan of activities with no way to tell whether anything is finished |
| 7. Tailor to suit the project | Fit the method to the size, risk, team and delivery method | Group 8 | Thirty documents on a three-month project — or “we tailored it” as the answer to every gap |
The third principle gained a word in 2023: roles, responsibilities and relationships. The addition is not decoration. A project borrows people from several organisations for a few months, and the edition’s new chapter on people says that how well it performs depends on “the strength of the relationships between them”. A role chart says who decides. It does not make the head of customer service, line 2.2, trust the agency enough to report a problem in week three rather than week twelve.
The seven practices — formerly themes
If the principles are what must hold, the practices are what must be managed from the first day to the last. Each is an aspect of the project that never gets “done” once. Each also comes with the documents it adds, which is where the method’s reputation for paperwork comes from — and why group 8 of the hook exists.
| Practice | The question it keeps asking | What it adds | In the hook |
|---|---|---|---|
| 1. Business case | Is this still desirable, viable and achievable? | Business case; benefits and sustainability management approaches | Group 1, 7.4 |
| 2. Organizing | Who decides this, and who must know? | Role descriptions; communication management approach | Group 2 |
| 3. Plans | What, by whom, how, when and for how much? | Project, stage and exception plans; team plans optional | Group 3, 8.2 |
| 4. Quality | How will we know it is good enough? | Product descriptions, quality register, quality management approach | 4.4, 6.2, 6.4 |
| 5. Risk | What might happen, and what would we do? | Risk register, risk management approach, risk budget | 4.6 |
| 6. Issues | This happened — who decides what now? | Issue register, change budget, change authority | 2.7 |
| 7. Progress | Are we still inside the agreed limits? | Tolerances, reports, the exception procedure | Groups 4 and 5 |
PRINCE2 recognises six kinds of issue, and naming them tells you who handles each. A problem has an immediate negative effect. A concern still needs its timing and impact assessed. An event outside the project may affect it. A business opportunity is an unexpected upside. A request for change proposes changing the baseline — the approved versions of the plans and products. An off-specification is a product that will not meet its quality specification.
The change budget and the change authority in line 2.7 are what keep requests for change from clogging the board. Take the request the WBS page describes in month three: refunds through the new checkout, “since that is obviously part of payments”. The dictionary entry for 4.3 excludes refunds, so it is a request for change, not a clarification. The agency priced it at €18,000 and a week of stage 3. That is above the project manager’s €5,000 and below the €20,000 the Executive and the Senior Users may approve together, so the two of them decided it in one conversation. They deferred it to after launch, and it reappears in line 7.3 as a follow-on action. The board never needed to meet.
The seven processes — one project’s life, start to finish
The processes are the timeline: what happens, in what order, and who does it. Read them in sequence and you have the whole life of the shop project, from the mandate on 2 December to the closure on 24 October.
| Process | When, and who | The question it answers | In the hook |
|---|---|---|---|
| 1. Starting up a project | Before the project exists; the Executive and the project manager | Do we have a viable and worthwhile project? | 3.1, 7.1 |
| 2. Directing a project | From the first authorisation to closure; the board | Do we start, continue, change course or stop? | 3.6, 5.4 |
| 3. Initiating a project | The initiation stage; the project manager | What exactly are we committing to, before any significant spending? | 3.2 |
| 4. Controlling a stage | Inside every stage; the project manager | Is the stage on track, and what needs a decision? | 5.2, 5.3 |
| 5. Managing product delivery | Inside every work package; the team managers | Is the agreed product being built to its description? | 6.3, 8.1 |
| 6. Managing a stage boundary | End of each stage except the last, and for any exception plan; the project manager | Is the business case still sound, and what is the next stage? | 3.3, 3.4 |
| 7. Closing a project | At the end — or early; the project manager, then the board | Is the product accepted, and who owns what happens next? | 7.3, 7.4 |
Three rules hide in that table. A project has at least two stages, and the first is always the initiation stage — which is why the shop’s board authorised initiation in December and the project only in February. Managing a stage boundary runs at the end of every stage except the last, because after the last one there is nothing left to plan. And closing a project is the same process whether the project succeeded or was stopped: PRINCE2 treats cancellation as a proper ending, with acceptance of what exists, handover, lessons and follow-on actions, rather than as an embarrassing exit.
Managing product delivery is where Scrum lives. The project manager authorises a work package — what to produce, its description, its tolerance, how to report — and the team manager accepts it, delivers it however the team works, and hands it back. That handshake is the only place PRINCE2 touches the work itself, and it deliberately says nothing about what happens inside.
Tolerances and exceptions — how a board stays out of the way
A tolerance is the permitted deviation from a plan before the level above has to be asked. PRINCE2 sets tolerances for seven performance targets, and the 2023 edition added the seventh, sustainability. Each level hands the one below a smaller allowance than its own. The business layer gives the board project tolerances. The board gives the project manager stage tolerances. The project manager gives each team manager work-package tolerances. As long as the forecast stays inside, nobody above needs to hear more than the regular report.
| Target | What its tolerance limits | In the hook |
|---|---|---|
| Time | How much later or earlier than the agreed date | 4.1 |
| Cost | How much over or under the agreed budget | 4.2 |
| Scope | How far the plan’s products may vary | 4.3 |
| Quality | How far a product may vary from its quality criteria | 4.4 |
| Benefits | How far benefits may fall short of, or exceed, the business case | 4.5 |
| Risk | Limits on the plan’s aggregated risks | 4.6 |
| Sustainability | Limits on the agreed sustainability measures — new in 2023 | 4.7 |
An exception is not something going wrong. It is a forecast that something will end outside its tolerance. That distinction does most of the work: the project manager raises it the moment the forecast crosses the line, not when the date has passed, which is what gives the board time to choose. Here is the one the shop project had, step by step, as PRINCE2’s exception procedure runs.
- The issue, 11 June 2025: the payment provider moves the shop’s 3-D Secure certification slot back by three weeks. The payments team manager reports it to the project manager the same day.
- The forecast: stage 3 now ends three weeks late, against a tolerance of one week in line 4.1. The launch would move to 6 October — outside the project’s own tolerance too.
- The exception report, 13 June — two working days, as line 5.3 requires. Option A: wait, and ask the leadership team to accept a 6 October launch. Option B: run the performance and security tests against the provider’s test double, certify in parallel, and absorb a week and a half in stage 3 for €22,000. Option C: launch on the provider’s older payment page without the new 3-D Secure flow, and upgrade in November. Recommendation: B.
- The board decides, 16 June: option B, and asks for an exception plan. Option C is refused against 4.4 — a payment flow the customer will see change twice in three months is a quality problem, not a saving.
- The exception plan replaces the rest of stage 3, with its own tolerances, and the board approves it on 20 June. The launch stays on 15 September; the €22,000 comes from the risk budget in 4.6, because a provider delay was already on the register.
Two projects the method exists to stop
Both cases are British public projects audited by the National Audit Office, in the country whose government wrote PRINCE2. Neither failed for lack of a method on paper. Each is mapped to the line of the hook that answers it.
| Case | What the auditors found | The line that answers it |
|---|---|---|
| FiReControl, 2004–2010 | Approved on a £120 million estimate; a board “not operating as an effective decision-making forum”; five Senior Responsible Owners | 1.5 and 3.6 — a stop condition, and a board with three answers |
| FiReControl | “No single, authoritative owner of the user requirements” | 2.2 — Senior Users who answer for adoption |
| BBC Digital Media Initiative, 2008–2013 | No senior responsible owner; a red risk rating that took six months to reach the executive board | 2.1 and 5.3 — one Executive, and an exception within two days |
| BBC Digital Media Initiative | The business case approved in April 2010 never revisited | 3.3 and 3.4 — every boundary re-reads group 1 |
FiReControl — seven years before the stop
FiReControl was meant to replace England’s 46 local fire control rooms with nine regional centres sharing one national computer system. The department and the Treasury committed to it in 2004, the National Audit Office found, “on the basis of very broad-brush and unrealistic estimates of costs of £120 million and an anticipated overall net saving of £86 million”. The first comprehensive assessment came three years later and said £340 million. The regional buildings came first: the first centres were finished in June 2007, three months after the IT contract was signed with EADS.
The method’s safeguards were present on paper and absent in practice. In 2008 the government’s own review body found that “the project board was not operating as an effective decision-making forum”; in 2009 it found “no single, authoritative owner of the user requirements” and a “cultural failing to share bad news early”. The project had five Senior Responsible Owners and four Project Directors. The department considered terminating in 2008 and 2009 and carried on. When it finally cancelled in December 2010, it had spent £245 million and finishing would have taken the total to £635 million. The auditors put the waste at a minimum of £469 million, and the nine empty centres had already cost £32 million to maintain.
The decision in December 2010 was, in the auditors’ words, justified: continuing would cost another £390 million and a year’s delay, while cancelling and upgrading the local rooms would cost £310–400 million. That is the comparison line 1.5 asks for at every boundary — the cost still to come against the alternatives, with the money already spent left out. A sunk cost cannot be recovered by finishing, so it belongs in the lessons report, not in the decision. FiReControl made that comparison once, six years late.
The BBC’s Digital Media Initiative — nobody to tell
The Digital Media Initiative was to let BBC staff create, share and archive video and audio from their desks. Siemens was contracted in February 2008; the contract ended in July 2009 with 18 months used and no working system, and the BBC took the work in-house. The National Audit Office’s memorandum of 27 January 2014 found that “the BBC did not appoint a senior responsible owner to act as a single point of accountability”. Development and the business case sat in one division, the benefits in the divisions that would use the system, and nobody owned the gap between them.
Reporting failed in exactly the place line 5.3 guards. The programme’s risk rating turned red for October to December 2011; the executive board heard in June 2012 and the BBC Trust in July 2012. By May 2012 the work was 15 months behind the timetable in the business case, and the business case approved in April 2010 had not been revisited. The BBC cancelled in May 2013, having spent £125.9 million — £98.4 million net of what it recovered from Siemens. The executive board itself named, at closure, “a failure to recognise the severity of the issues in the reports they had received”.
Where it degrades — and what PRINCE2 cannot do
For an engineering team, PRINCE2 mostly shows up as stage boundaries: predictable moments when funding is re-examined and evidence is expected. Teams that fare well treat them as real reviews and bring outcome numbers. Teams that fare badly bring a percentage complete, which answers a question nobody on the board asked. The ways the method decays are all versions of the same thing — the documents survive and the decisions they were for do not.
- A business case written once for approval and never reviewed — the first principle absent while the paperwork claims otherwise.
- A board that meets but has never stopped anything, which makes every stage boundary a ceremony.
- No tolerances agreed, so every deviation is either escalated or hidden.
- Assurance that reports to the project manager, so the board hears only one voice.
- Issues logged as risks, so a problem that has already happened waits for someone to reclassify it.
- The full document set on a small project, where governance costs more than the risk it manages.
- Benefits nobody measures after closure, because the project that promised them no longer exists.
When to use it
Use it when
- Funding is released in stages and someone must justify each release.
- Public sector or regulated work where recognised governance is expected or written into the contract.
- Several organisations are involved and decision rights must be written down.
- The organisation genuinely wants the ability to stop a project cleanly.
Avoid it when
- Continuous product development with no defined end — there are no stages to authorise.
- A small project where the documents would outweigh the work itself.
- Expecting it to tell the team how to build — it deliberately does not.
- The board will never use its authority to stop, which turns the method into ceremony.
Sources and method
What this page stands on
- 6 sources
- 2 documented failures
- Checked October 16, 2026
The method
the five elements, the principles, practices and processes, tolerances and the exception procedure- BookPeopleCert — Managing Successful Projects with PRINCE2, 7th editionPeopleCert · 2023
- GuidePeopleCert — PRINCE2 7 Foundation Quick Reference Guide (Official Training Materials) (opens in a new tab)PeopleCert International Ltd. · 2023
- DocumentationPRINCE2 Project Management (opens in a new tab)PeopleCert
- DocumentationPRINCE2 — The Language Of Success (opens in a new tab)PeopleCert
The failures this page learns from
FiReControl and the BBC’s Digital Media Initiative, in “Two projects the method exists to stop”- Regulator reportNational Audit Office — The failure of the FiReControl project (HC 1272) (opens in a new tab)National Audit Office · Jul 1, 2011
- Regulator reportNational Audit Office — British Broadcasting Corporation: Digital Media Initiative (memorandum to the BBC Trust) (opens in a new tab)National Audit Office · Jan 27, 2014
Researched with Claude. Finding and cross-checking the primary sources was done with Claude (Anthropic). Selection, writing and fact-checking: the Anatomy of IT Product editors.
Found this useful?
Share it with someone who is working on the same problem.