Value Driven Design Logo

6.25 - The Pugh Matrix

The Pugh Matrix is a systematic technique, devised by Prof Stuart Pugh, to evaluate design concepts to find a solution which best meets the customer need before working on detail design and development.

Two inputs are needed: customer requirements (usually in the form of an early PDS) and a number of possible solutions to satisfy these requirements. The technique helps to extend both the requirements and solutions by showing where the requirements have not been fully defined and by helping to generate more possible solutions.

Why control Concept Convergence.

Simple everyday products have to be designed to satisfy many requirements. Many solutions are possible and some fit the requirement better than others. Competitors may also be looking to profit by providing a solution. So how do we make a logical choice? Using ‘gut feel’ in the concept stage of new product development creates product failure risk because they are, in Pugh’s terms, 'conceptually vulnerable'. For Example:

i) The product will fail because it does not address all the requirements adequately.

ii) The product meets the requirements but we do not fully understand why. The team will be unable to argue for or against it logically.

Some teams have dominant members who push a preferred solution. This tendency to override others opinions could take the team to the next stage without making a thorough decision analysis. This must be avoided.

What Do We Have to Achieve?

Solutions must be compared on their merits, referring between the requirements and solutions, cross checking between the two. The team needs a concise understanding the requirements and a consistent understanding of what is being compared. The requirements can be taken from a product design specification (PDS) or the 'hows' from the first stage of a QFD exercise. The solutions can be produced from any of the idea generating techniques (brainstorming, analogy, etc.) and by considering competitors products.

The team define requirements and propose possible solutions. Neutral solution comparisons lead to a common understanding of why a particular solution has been chosen. The team will gain confidence to be able to add any solution to the consideration and analyze it on the same unbiased basis. The technique enables convergence to a team preferred solution within a few iterations.

How to control Concept Convergence.

The Pugh Matrix highlights shortfalls in solutions and shows where alternatives meet the requirement more fully. Using ideas from other solutions helps build better alternatives. Therefore, a weighted rating matrix is NOT appropriate. At concept stage there is an open field of design choices. Not like a customer weighing product offerings with fixed features.

Individual team members produce ideas for solutions or part solutions on their own allowing focus on their area of expertise (e.g. mechanical, hydraulic or electrical) not influenced by others.

The Pugh Matrix can be used at various design levels for choosing and refining ideas. At the high level it can be used to choose between several roughly thought out concepts from which a number of favourites can be chosen for detailed analysis against finer criteria. Leading ideas will appear that should have more detailed design work carried out on them. There are a number of ways of detailing these, from which favoured solutions and new ideas will emerge (Figure 1).

Alternatives

1

2

3

4

5

6

7

8

9

10

11

12

Choose Alternatives for Analysis

4

6

10

11

Choose alternatives & derive options with more detail

6

11

Choose alternative for detail design

a

b

a

b

c

d

11

c

Continue through the detail stage refining the alternatives and the criteria for analysis

Figure 1: Refining alternatives to converge on the best concept

Process steps to control concept convergence using the Pugh Matrix.

1. List all requirements.

Use short simple statements that allow a comparison of solutions. For some statements the team may wish to show whether the value for the requirement is a: maximize, minimize, target or an attribute. Avoid using numbers if possible - at this stage it will be difficult to say whether a solution will produce a particular performance. It is easier to state which of two solutions might be better or worse.

It is important to define criteria in terms that can be measured as a group. Write the definitions down. If they are not written down false assumptions could be made – leading to time wasting argument and clarification later. For example – “ease of use.” A clear definition of what “ease of use” in this context will be required. For example: minimum back bending, least number of buttons to press or one handed operation. Later, “ease of use” may need to be broken down into these criteria and solutions compared directly against them as some ideas might be better in some areas than others. If the Pugh matrix is used after a QFD exercise there should be little trouble with this step, much of this will have been developed already in the 'hows' of stage 1.

2. Discuss the solutions.

Each idea should be sketched and described to everyone on the comparison team. Elaborate each idea to a similar standard and level of detail so that no one idea is given an advantage through its presentation. More time should be spent by the team trying to improve on the ideas presented or using colleague’s ideas to trigger new ideas. Add them to the list before starting the analysis.

Some solutions will be variations on a theme. In the early stages they should be grouped together and treated as one. If the overall idea shows promise it can be taken to the next stage of analysis and these alternatives brought out for more detailed study.

3. Prepare a matrix.

List the requirements down one side of a matrix in a rough order of importance and the possible solutions along the top. It is best to set up the matrix this way since our eyes naturally scan horizontally making it easier for us to study the matrix later.

Identify the solutions with numbers or letters rather than words (descriptions can be long and misleading). Alternatively, make the matrix large enough to put A4 or letter size sketches along the top representing the solutions. Leave blank columns to add new solutions as they are developed. A large matrix can appear to take over the office. The advantage is that it visible to everyone to see and discuss. Some companies have “war rooms” for this purpose.

4. Select a datum.

For the comparison each feature of each solution will be compared with one common solution (the DATUM). If the requirement is to replace an existing product it is usually best to take this as the datum as most involved in the study will be familiar with it. In the case of a new product use the idea which at first appears to be the best.

5. Compare solutions.

Compare each requirement of each solution against the datum, asking if the solution is better than, worse than, or the same as the datum for meeting this requirement. Put a '+', '', or '=' in the appropriate matrix square. Moving from left to right across the matrix makes it less likely for ‘favourite’ solution to be given bias.

6. Assess the scores.

Add the +'s and 's for each solution. The +'s and 's do not cancel each other out. The stronger concepts will have the most +'s. However, they will probably have some 's too.

7. Generate further concepts.

Using the strong solutions, look along the rows where they have minus signs to see if any of the other proposed solutions have a plus. Can the positive feature from this solution be incorporated into the leading solution - or any other solution?

Do not fill out the matrix in full for the sake of it (especially large ones with more than ten solutions). Look for any patterns which are emerging and ask why some ideas appear to meet some criteria better than others. Do they have anything in common? Can something be done about the solutions with lot of 's? If not, do not waste time analyzing them. Keep the picture as the idea might hold a useful feature.

8. Re-run the evaluation.

Add the new solutions generated in stage 7 (and any solutions, or modified solutions, thought of in the meantime) to the matrix and re-assess them using one of the modified solutions as the datum.

During stage 7 it might found that none of the solutions is better than another. This could be due to ambiguous criteria, or some of the concepts are really the same. Look for these. If this is not the case, re-run the analysis using another datum! If two solutions have close scores, use one as the datum and compare them.

Do not consider cost as a requirement. It cannot be assessed at this stage, it causes too many arguments in the team and in the long run it does not matter because it can NOT be treated as an independent criteria; it is value for money which matters - what are you getting for the cost - not the cost alone. Criteria which act as cost indicators will be as results from a Design for Assembly, Manufacturing Analysis or some kind of complexity factor[1]

If the ideas have been generated by a number of small teams (subsets of the overall project team) these teams should compare the solutions independently at first since it is easier for everyone to have their say in a small team. When this is done the teams should compare their results. Usually, they largely agree but one team may have given a plus where another team has given a minus. Upon discussion of these issues the teams will find one of the following possibilities:

  • Misunderstanding regarding a feature of a proposed solution.
  • Feature has an unknown element and until such time as one team can find hard evidence to support their opinion the teams will continue to disagree.
  • The criteria have not been defined adequately.
  • Team member(s) have some past experience of the feature and has some proof as to whether it will work or not.

Extension to concept selection.

The technique is designed to improve concepts and allow a preferred solution to grow. Some people like the idea of weighting the requirements and rating the extent to which the solutions meet these requirements but weighting/rating methods have two main weaknesses at this stage in the design cycle:

i) Numerical answers can be regarded as unquestionable. This is not true as small changes in weights or ratings can lead to disproportionally large changes in score. The differences between the highest scores may be so small to really say which is best.

ii) To give a factor a rating from 1 to 5 is slower and more tiring for the assessors than deciding if something is better than or worse. People’s mood changes from day to day - and so do their ratings.

The use of the Pugh Matrix has been criticized because there is no way to tell to what degree a solution fulfils an important requirement, or even which are the important features. The Pugh Matrix in conjunction with QFD can resolve this. Part of the first stage of QFD is to carry out a technical assessment of competitors products. The targets set for product requirements depend upon the answers from this study. Include these products in the Pugh Matrix. This highlights the features in competing products that have to equalled or beaten to reach the targets. Use the leading solution as the datum and compare it against the competition to find how close the proposed solutions are.

What do we do now?

If a single stage Pugh Matrix controlled convergence exercise is conducted it will not produce a final solution. At least two may be shown in an equally favorable light. The final choice is on the value for money each offers. A costing exercise, and some work on Failure Modes and Effects Analysis (FMEA) will help assess some of the risk.

Concept convergence is a continuous process. The Pugh Matrix can be used in latter design stages to refine solutions in detail against finer requirements. It can be used once to help make a choice at the concept stage. The greatest benefit comes when it is used in all stages. The final solution is most likely to have grown from more than one of the ideas. Team members feel they have contributed and usually are more committed to the favoured solution.

SUMMARY OF BENEFITS

  • Confidence that the developed solution is the best available and addresses all the requirements.
  • The technique highlights the strengths of the preferred solution.
  • Promotes team ownership of the final solution by forcing members to be open about their opinions on the solutions and by giving them the feeling they have all contributed to the solution and decision making process.
  • Difficult for one idea to succeed through the forcefulness of an individual.
  • The degree of risk on the chosen concept can be assessed and development work focused on this.
  • Decisions regarding why a solution was chosen can be traced via the matrix.

A Worked Example of Concept Convergence

A MOTOR VEHICLE AUDIBLE WARNING SYSTEM.

This example (Ref: Pugh, S., 'Concept Selection - a method that works.' ICED '81, Rome) shows The Pugh Matrix used to assess the suitability of a number of solutions to the broad requirement for: 'A means of providing an audible warning of the approach of a motor vehicle'.

Notice, it does not ask for a car horn. A vague description discourages preconceptions and promotes more possible concepts.

Sketches in Figure 2 show 12 possible solutions. Figure 3 shows the Pugh Matrix used to assess the solutions. The traditional car horn was used as the datum. Concept 4 is very strong. The cost of this design at the time of the study was very high and represented poor value for money to most users. Today a similar concept is used for sirens on emergency service vehicles, and almost without exception for car alarm systems. It shows that it is worthwhile monitoring the technology to see if the costs fall low enough to make the concept viable in future.

This example shows cost as one of the criteria. Pugh carried out this exercise before he found out it was a mistake to consider cost too early.

The Pugh Matrix - Example

Figure 2. Design options for audible warning system


Concept

1

2

3

4

5

6

7

8

9

10

11

12

Criteria

Values

Datum

1

Sound Level

105 – 125 DbA

=

-

+

-

+

+

-

-

-

=

+

2

Frequency

2000 – 5000 Htz

=

=

+

=

=

+

=

-

-

=

+

3

Resistance to Corrosion & Water Ingress

-

-

=

-

-

=

-

+

-

-

=

4

Resistance to vibration and Shock

=

-

=

-

=

-

-

=

-

-

-

5

Resistance to Temperature Changes

=

-

=

-

-

-

=

=

-

=

=

6

Response Time

=

-

+

-

-

-

-

=

-

-

-

7

Complexity: No of Stages

-

+

=

+

+

-

-

-

+

-

-

8

Power consumption

-

-

+

-

-

+

-

-

-

=

+

9

Ease of Maintenance

=

+

+

+

+

-

-

=

+

=

-

10

Weight

-

-

+

-

-

-

=

-

-

-

+

11

Size

-

-

=

-

-

-

-

-

-

-

-

12

Part Count

=

=

+

=

=

-

-

+

-

=

-

13

Service Life

=

-

+

-

=

-

-

-

-

-

-

14

Manufacturing Cost

-

=

-

+

+

-

-

=

-

-

-

15

Ease of Installation

=

=

=

=

+

-

=

-

-

=

-

16

Shelf Life

=

=

=

=

-

-

=

=

=

=

=

Totals

+

0+

2+

8+

3+

5+

3+

0+

2+

2+

0+

4+

-

6-

9-

1-

9-

7-

12-

11-

8-

13-

8-

9-

Figure 3. The Pugh Matrix for audible warning system concept evaluation


[1] e.g.:- Complexity = k(NpNtNi)1/3/f.

Where k - a constant, f - No of functions, Np - No of parts, Nt - No of part types, Ni - No of interfaces.

 
Further Reading