Design Responsibility Matrix: A Guide and Template for UK Architects

What is a design responsibility matrix?

A design responsibility matrix is a table that records, for every element of a building, who is responsible for designing it, at which stage, what they must hand over and which other designers they must coordinate with.

The RIBA Plan of Work 2020 calls it the Responsibility Matrix. Its purpose is stated plainly: it makes clear what tasks will underpin the production of information and who will undertake them. The Plan of Work asks for it to be produced in Stage 1, and says it needs to focus on the boundaries between Stage 2 and Stage 3 tasks, and between the design team and any specialist subcontractors at Stage 4. Those two boundaries are where most gaps open up in practice.

A good matrix answers four questions for every element. Who designs it? What do they receive from others before they can start? What must they return, and in what form? Where does their design stop and somebody else's begin? If a row cannot answer all four, the element has an owner in name only.

The matrix is a different document from the appointment schedule of services and from the contractor's design portion (CDP) schedule in the building contract, although it should agree with both. The schedule of services says what the architect will do for the client. The CDP schedule says what the contractor will design. The matrix sits underneath both, at element level, and shows how they fit together. The matrix is also distinct from the information management responsibility matrix that an ISO 19650 project may require, which allocates information management tasks rather than design tasks. The two are compared in ISO 19650 for Architects.

Why a technical director cares

Most disputes about who should have designed something begin with an element that nobody had written down. The curtain walling designer says the structural engineer was to size the support brackets. The structural engineer says the bracket loads were never provided. The architect says the specification required the contractor to coordinate. Every statement is arguable, because the documents did not say. A matrix does not remove the argument, but it moves the starting point to a written allocation rather than an assumption.

If a design element appears in the specification but not in the matrix, nobody has agreed to design it.

Who owns what across the design team, engineers and specialists?

Ownership falls into four groups: the lead designer and architect, the consultant engineers, the contractor's design portions and the specialist subcontractors. Each owns a different kind of decision, and the interfaces between them need to be named separately.

The first job in building a matrix is to sort the building into four groups of responsibility. Most practices find that two groups are easy and two are where the work lies.

Prescriptive and descriptive information

The RIBA Plan of Work 2020 describes the choice at the heart of the matrix. For each building system, the design team either delivers prescriptive information (the design is complete and the contractor builds it) or descriptive information (the design team states what the system must achieve, and a specialist develops the solution). The matrix records which, element by element. The specification then has to be written to match: a prescriptive clause for a prescriptive row, a performance clause for a descriptive row. A mismatch between the two is one of the quietest ways to create a gap. The guide to contractor-designed portions and performance specifications covers how to write the descriptive side.

Where principal designers fit

The matrix is also the practical tool a principal designer needs when coordinating design risk. The CDM principal designer and the Building Regulations principal designer on a higher-risk building each need to know which designer owns which element, otherwise a risk is closed by one party that has no authority over it. CDM vs Building Regulations principal designer sets out how the two roles differ, and the principal designer duties guide covers the duties themselves.

What should a design responsibility matrix template contain?

A usable template has one row per element and six columns: the element, the stage, the responsible party, the type of information, the deliverable and the interface. Everything else is optional.

The template below is deliberately short. Long matrices go unread. The six columns are enough to settle most arguments, and each one does a job.

How to fill in each column

Element. Use the same tag, type reference or system code that appears on the drawings and in the schedules. If the matrix uses its own names, it will drift from the other documents within a month. Where the practice classifies work with Uniclass or CAWS, add the code as a separate column rather than replacing the tag.

Stage. Show the stage at which the design is completed, not just started. A curtain wall is started by the architect at Stage 2 and finished by the contractor after award. Both facts belong in the row.

Responsible party. Name one party per row for each task. Where a row genuinely splits, split the row. "Architect and contractor" in a single cell is a sign that nobody has decided.

Information type. Prescriptive or descriptive. This is the column that links the matrix to the specification, and it is the one most often left out.

Deliverable. State what the responsible party hands over, in terms a reviewer can check: calculations, drawings at a stated scale, test evidence, a design report. Include what the design team will receive from the designer of a portion, and when, so that review rights have a date attached.

Interface. List the neighbouring elements and say who owns the junction. This is the column that prevents two specialists each assuming the other designs the connection.

Optional columns worth adding

Three further columns earn their space on larger projects: the planning or building regulations requirement the element supports, the insurer or warranty provider requirement if one applies, and the clause or section number in the specification. The clause column is the most useful because it turns the matrix into a cross-check. If a row has no clause, or a clause has no row, the pre-issue review has found a problem before the contractor does.

How does the matrix tie to the specification and to a contractor's design portion?

The matrix allocates responsibility, the specification states the requirement and the contract makes the allocation enforceable. When the three disagree, the contract and specification usually govern, so the matrix has to be written to match them, not the other way round.

A matrix kept in a spreadsheet on the project server has no contractual force on its own. Its value is that it forces the allocation decisions to be made early and then drives what is written elsewhere. Three documents must agree with it.

The specification

Every descriptive row needs a performance clause that states the result required, the standard it is measured against, the evidence that proves it and the design information the specialist will receive. Every prescriptive row needs a clause that is complete enough to build from. A clause that says "to be designed by specialist" without inputs or deliverables, for a row the matrix marks prescriptive, is a contradiction. The same applies in reverse: a detailed prescriptive clause for a row the matrix gives to the contractor creates doubt about who is responsible if the prescribed solution fails.

The contract and the CDP

Under the JCT suite, a contractor's design portion is normally created by using the Contractor's Designed Portion Supplement with a traditional contract, and identifying the portion in the contract documents together with the requirements the contractor must meet. The contractor then becomes responsible for the design of that portion, and the standard of that responsibility depends on the form and on any amendments made to it. The matrix is the practical record of which elements are CDP and which are not. If the matrix says an element is a CDP and the contract documents do not identify it, the architect has a difficult conversation ahead. Take legal or insurance advice on the exact wording for each appointment and check the standard of duty against the contract edition in use.

Because the contract terms govern, a technical director should read the CDP provisions of the form being used before the matrix is finalised, and take legal advice on amendments. The matrix cannot create responsibility the contract does not.

Review rights and design reviews

Reviewing a specialist's submission is not the same as taking over their design. The matrix should record, for each CDP row, what the design team reviews, for what purpose and in what time. A review comment that goes beyond the agreed purpose risks being read as a design instruction, which can blur the responsibility the matrix was written to fix. The specification should use the same words: what the contractor submits, what the architect reviews, and what review does and does not mean.

Why does the matrix so often stop at Stage 3, and how do you fix that?

The matrix stops at Stage 3 because it is written to support planning and design development, and nobody is assigned to keep it alive once technical design starts. It should be revised at least four times: at the start, at the end of Stage 3, before tender and after the contractor's design proposals are known.

This is the failure the audit guide lists as root cause 07. The matrix exists, it was sound at Stage 3, and it says nothing about Stage 4 packages. By then the contractor's design portions are being defined in the specification and the contract documents, and the matrix is no longer the document being used. Often it is not opened again until a defect appears. See Why Specifications Fail the Same Way in Every Practice for how that fits among the other causes.

There are three common reasons. The first is ownership: the person who wrote the matrix was the project architect at Stage 2, and technical design is run by someone else. The second is format: the matrix is a spreadsheet that nobody opens during drawing production. The third is timing: the real allocation decisions at Stage 4 depend on the procurement route, which is often not settled until late. The RIBA Plan of Work 2020 recognises this directly. It notes that the procurement strategy influences when building systems will be designed, and may require Stage 4 to be undertaken in two parts, with specialist subcontractors designing building systems after the building contract has been awarded.

A checklist for keeping the matrix alive

• ✓One named owner for the matrix, with the role written into the project quality plan.
• ✓Every element tag in the matrix appears on the drawings and in the schedules, and the reverse.
• ✓Every descriptive row has a performance clause, a design brief and a list of required returns.
• ✓Every CDP row appears in the contract documents as a contractor's designed portion.
• ✓Every interface row names the party who owns the junction.
• ✓The matrix is revised at each handover point and the revision is dated and recorded.

For a practice running many projects, the simplest control is to add one line to the pre-issue review: compare the matrix with the specification. The pre-issue technical review checklist includes this as the final check, and the RIBA Stage 4 guide shows where in technical design it belongs.

Questions technical directors ask about design responsibility matrices

What is a design responsibility matrix?

A table that records, for every element of a building, who is responsible for designing it, at which stage, what they hand over and where their design meets someone else's. The RIBA Plan of Work 2020 calls it the Responsibility Matrix and asks for it to be produced in Stage 1.

Who is responsible for the design responsibility matrix?

The lead designer normally prepares it and keeps it up to date, with input from each consultant. On a design-and-build project it passes to whoever holds design coordination after novation or award, and the handover should be recorded. The practice should name one owner in the quality plan.

What is the difference between a design responsibility matrix and a contractor's design portion?

The matrix is a coordination document that covers every element and every designer. A contractor's design portion is a contractual allocation of design responsibility for particular elements to the contractor. The matrix should show which rows are CDPs, and the contract documents must identify the same portions.

When should a design responsibility matrix be produced and updated?

The RIBA Plan of Work 2020 expects the first version in Stage 1. It should be reviewed at the end of Stage 3, aligned with the specification before tender, updated when specialists are appointed and controlled through construction.

What should the matrix say about prescriptive and descriptive information?

For each building system it should record whether the design team provides prescriptive information, where the design is complete, or descriptive information, where the requirement is stated and a specialist develops the solution. The specification clause must follow the same choice.

Is there a standard template?

The RIBA Plan of Work toolbox provides a pro-forma responsibility matrix. A practice can also use a shorter version with six columns: element, stage, responsible party, information type, deliverable and interface, as set out in this guide.

Where Avoice fits

The matrix is a coordination document, and coordination is where software helps once the practice has settled the rules.

In Avoice, the pre-issue Technical Review compares the specification against schedules and drawings, which is the check that catches a descriptive row with a prescriptive clause, or a tagged element with no clause at all. Practice Rules can require that every performance clause states a result, a standard and the evidence that proves it, and the clause library can hold the practice's standard wording for contractor-designed portions so that inputs, interfaces and deliverables are not left to the author's memory. The Compliance module keeps principal designer duties and the CDM risk register alongside the design, so that a risk is assigned to a designer who actually owns the element. Construction Admin then tracks submittals against the design portions after award.

The boundary stays where it was. Software can hold the structure, compare the documents and flag the mismatch. The technical director decides who designs what, and the contract determines what that allocation means.