An Important Measure of Performance – Product Recalls

US Vehicle Recalls and Sales

The BBC correspondent Brian Hanrahan became famous during the Falklands war when he counted the number Harrier Jump jets taking off and returning to the British aircraft carriers.  But data published by the National Highway Transport Authority (NHTSA) shows that the vehicle manufacturers are now recalling more vehicles in the USA than they are selling.

The data downloaded from the NHTSA website has been analysed to compare the number of vehicles recalled due to ‘vehicle defects’ with vehicle sales since, to calculate the ‘Recall Ratio’  as an measure of industry performance.

The data shows the ‘average car’ sold in the US market will be recalled at least once to correct a safety defect. But perhaps the most worrying conclusion for car manufacturers. is the long term trend of increasingly frequent product recalls. It is not clear if the rising trend in the number of recalls reflects a cultural change as consumers have become unwilling to accept ‘dodgy products’, or is a bye-product of increasing complexity in vehicle design. Although the data shows that recalls are not limited to the more complex high-tech systems

So,if you need training and support to understand product liability and develop a cost effective quality management system please visit the  Stunell Technology  website  where you will be able to download a brief report based on the NHTSA data.

(Although similar recall data for the UK and European markets is not published, UK motorists can use the VOSA website to find details of any recalls for a particular vehicle model.)

3 Dimensions of Risk

As Australian student Erin Langworthy queued with other tourists waiting jump off the Victoria Falls Bridge over the Zambesi River in South Africa, I doubt she expected this:-

Although adrenalin addicts and thrill seekers may enjoy the excitement of bungee jumping they rely on the operators using ‘risk assessment techniques’ to identify, control and eliminate the risk of death or injury. In fact we all have a legal and moral responsibility to consider how our actions may affect other people, and to take reasonable precautions to avoid causing them harm.

Risk Management models  recognize the ‘Three Factors’  that you can identify and actively manage to minimize the risk of unintended loss or injury in any situation. These are:-

The severity of the foreseeable consequences of ‘failure’, which may be reduced if you have a planned response to the initial failure, that contains the problem and prevents injury.

How frequently failure may occur, called occurance which can be improved by eliminating the opportunities for error in design, production and operation.

Your ability to detect the causes of failure and take corrective action to avoid the consequences of failure

We can then assess each of the ‘3 Dimensions of Risk‘ independently using a 1-10 rating scale, where ’1 is good’ and ’10 is bad’, to decide on appropriate actions to improve the current situation. (Of course the rating scales used must  be ‘calibrated’ against some internal or external standard if we are going to be consistent in our judgement, but that will be the subject of another blog!)

Risk Priority Numbers combine Severity, Occurance and Detection scores

These independent scores can be multiplied together to calculate a Risk Priority Number or RPN. However different combinations of severity, occurance and detection scores can produce the same RPN value. For example, the severity, occurance and detection scores that combine to produce an RPN of 175 are shown here. So we must avoid ‘simplistic approaches’ and understand what the RPN score actually means.

Risk Management strategies must identify unacceptable risks and and take appropriate action to prevent, detect or mitigate known failures so that the ‘potential consequences’ don’t occur. However, we can only succeed if we are willing to contemplate what may go wrong, and learn from past mistakes.

Risk management is not optional!

The recent disclosure that a ‘rogue trader’ at UBS lost $2 billion before anybody noticed, has focused attention once again on the apparent failure of the bank’s risk management procedures. While the sums of money involved in this case are extreme, the problems caused by rogue employees, managers and directors are not unique to the banks and financial services organisations.

Although the UK the National Fraud Authority has estimated financial fraud cost the UK Economy £38 Billion last year, it may be a mistake to assume that deliberate fraud for personal gain is the  greatest risk  to your business. In fact Mr Adoboli, the trader at UBS, was not attempting to move the money out of the bank for his own use, but was bending the rules in an attempt out-perform his colleagues. In that sense at least he could be trusted to act in the ‘best interests of the bank’ even if he was grossly incompetent and out of control.

So I have been wondering:-

•  How much damage is caused by other rogue employees managers and directors, motivated by the desire to produce truly extraordinary results, making irrational decisions?
• How can we identify and control the risks created by those who aspire to be recognised as ‘star performers’?

If the management team don’t understand their products and processes, it seems unlikely that they will understand the risks to the business if they go wrong, or spot the warning signs when the process is out of control. Managers therefore need to understand the detail of what happening, and move away from the culture of ‘management by results’ that ignores disasters during their gestation period, waiting until it is too late before responding to them.

An effective risk management strategy must recognise the potential hazards to the business, identify the possible causes and put in place effective controls to prevent or contain the mistakes and mitigate its potential effects. That demands a level of transparency and accountability that is sadly lacking in many organisations, and we therefore need to re-establish some basic business ethics.

Most of the tools and techniques required to regain control are already well established, but are under utilised because many senior managers enjoy the power to make irrational decisions which they explain, are based on their intuitive knowledge of the business. For these individuals ‘evidence based decision making’ is an unnecessary burden, and you can trust them because they have not broken the bank yet!

Process Capability

 

What CPk means

‘Process capability’ can be used to describe the ability of any process to deliver the required outcomes, but in a manufacturing process it has a particular meaning as discussed below:-

The numerical value of  the Process Capability Index (Cpk) depends on the repeatability of the process and the level of variation considered acceptable, so you can get ‘improved process capability’ without changing the process by relaxing the permitted tolerances.

During the design process you may need to establish the effect of ‘variation’ on the performance of the product, (using design of experiments or ‘parametric analysis’) in order to decide on the production tolerances. In the case of geometric tolerances you may also need to do a tolerance stack-up to find the cumulative effect of all the permitted tolerances.

However, if the ‘design requires very close tolerances’ and the process ‘is not capable’ then the variations in the production process will cause unacceptable variations in product performance or durability, and a low ‘process yield’ with high levels of scrap and rework.

In such situations two approaches (which are complimentary) can be used:-

• Find and eliminate the causes of variation within the process to improve repeatability, using Process FMEA’s, fault trees and other similar techniques, and introducing additional ‘controls’ on the inputs to the process to reduce or eliminate variations in the output, and / or
• To review the design to ‘desensitise it’ to the levels of process variations that are found, so that the design tolerances can be legitimately be relaxed, without compromising product quality.

In some organisations this ‘feedback loop’ between design and production comes under the general heading of Design for Manufacture and Assembly (DFMA) while other companies apply QFD techniques to understand what they ‘must control’ in order to maintain product quality.

However, the very nature of ‘new technologies’ means that they are vulnerable to ‘uncontrolled variations’ in the production process, which may not have been stabilised or optimised, and the consequences of apparently minor changes can very significant. (People only discover the need for better controls when the process or product ‘fails’)

If you can systemise the knowledge you can then (and only then) adopt a scientific approach to product and process development, using predictive analysis to simulate performance and durability. Without that body of knowledge you are forced to ‘suck it and see’ as you move forward by ‘trial and error’ – with high levels of scrap and rework – and frequent project delays.

Process capability can become a big issue as you increase production volumes and need to achieve cost reductions, because poor process capability always introduces high levels of re-work and drives the need for ‘inspections’ – but you can never ‘inspect-in quality’. As soon as anybody starts talking about ‘selective assembly’ you should ask why the process capability is so poor!

Can you prevent failure?

The recent engine failure on the Quantus A380 has shown once again that an apparently minor problem can have serious consequences. Thankfully the crew were able to make an emergency landing without any significant injuries, but the safety of the aircraft was  ‘at risk’ due to the ’uncontained engine failure’ that allowed flying debris to penetrate the wing and damage the aircraft’s control system.

The ongoing investigation will undoubtedly find the root cause of the failure enabling Rolls Royce, who made the engine, to prevent any repeats. However the incident reminds those involved in product development and manufacturing that we must recognise ‘potential problems’ before disaster strikes – and put in place controls that eliminate the causes of failure or give adequate warning so that action can be taken to avoid or reduce the consequences.

Failure Modes Effects Analysis (FMEA) is used by engineers to identify possible failure modes (what could go wrong), and the effects of failure when it occurs. However the real benefit of using FMEA techniques is that it forces you to think about ‘how you could generate the failure condition’ and to identify the potential causes of failure, so that you can prevent or at least detect them.

The challenge in the FMEA process is to find appropriate controls that will reduce or eliminate the mistakes that could be made in the specification, design or manufacturing processes.  However when that is not possible, or the consequences of failure are very severe, we must find ways of detecting the problem and warning people so that they can avoid the potential consequences.

The tools and techniques already exist that will enable you to significantly reduces the risk of failure in your products or processes, but they will only work when you believe failure is possible and put the controls in place!

Managing Customer Expectation

This year we have spent a lot of time with our clients trying to understand and prevent failures, while investigating problems found in the product development process and ‘in service’.  In some cases the mechanisms of failure have been complex and the solutions technically challenging – while others are frustratingly obvious and simple to solve when you ask the right questions!

However, all too often the ‘root cause’ has been a mismatch between the real and imaginary market requirements. While customers live in the real world, and define the real requirements for your products, producers frequently inhabit a virtual world filled with abstract descriptions and computer models of  the ideal customer and what he should require. Experience shows that the difference between these definitions of reality may be the difference between success and failure.

Ready Frozen Screen wash

Customers have a right to complain when we don’t listen to what they are telling us about the ‘market requirements’ or we create expectations that we don’t fulfill. Which is why I was unimpressed (but amused) by the container of ‘Ready to Use Screen Wash’ with ‘Anti-Freezing Properties’ – frozen solid in the car boot this morning!