Managing a Product Recall!
By David Hutchins, Chairman, David Hutchins International Limited.
First published in 'Tell Me' the journal of the Quality Methods Association.
Risk Management is unfortunately an imperfect science. No matter how careful we are, even a probability of one in a million will, by definition occur on average one in a million times. It also seems that Murphy's Law prevails because it is often our impression that these occurrences seem to 'pick their moment'.
Like heart attacks, cancer, road accidents and muggings, product recalls happen to other organisations but could never happen to us! That is what everyone thinks until it does happen.
If it did happen to your organisation, could you cope? Would you know what to do? Could the company survive the cost let alone the bad publicity? Does your Risk Management policy include a Product Recall procedure?
Twenty years ago, when Products Liability legislation was more a discussion point than reality, there was considerable speculation that Product Recall would be one of the most crippling side effects of the European Draft Directive on Product Liability if it ever became Law.
Since then, of course, it has become law. The manufacturer of a product being held liable if a defect in the product or in its design could be determined to have resulted in an injury. The product would be regarded as defective if it caused the injury.
Also ISO 9001:2000 has two specific clauses related to Risk Management. For example:
Clause 8.5.2 Corrective Action states: The organisation shall continually improve the effectiveness of the quality management system through the use of the quality policy, quality objectives, and audit results, analysis of data, corrective and preventive actions and management review.
Clause 8.5.3 Preventive Action states: The organisation shall determine action to eliminate the causes of potential nonconformities in order to prevent their occurrence. Preventive actions shall be appropriate to the effects of the potential problems.
In addition to this, there is also the ISO 14000 environmental standard and all the Regulations related to Health and Safety Act reference Occupational Health and Safety Systems (OHSAS 18001), which today form part of practical Risk management. Please see the courses on all of these topics in our Risk Management section of courses.
So what has been the impact on the number of Product Recalls since the introduction of all this legislation?
It is difficult to say whether or not there are more than there were twenty years ago because it is doubtful if anyone collated the data. However, judging from the number that appear in the national press and are announced on the radio, it would certainly appear that there were more today.
In February 1998, Vauxhall Motors announced the recall of a large number of one of its brands due to the discovery of a defect in the design of a handbrake. Just prior to that Ford announced something similar. Mercedes narrowly avoided a complete disaster when its new baby car was found to be so unstable that it was capable of rolling over in relatively foreseeable conditions. The problem is not just restricted to the American and European auto manufacturers. Toyota and other Far Eastern manufacturers have all had their share despite their renowned Quality Control capabilities.
But away from the Automotive industry other product providers are equally as vulnerable, even those whose very name is associated with Quality. For example, a headline in a copy of the Daily Telegraph ran - “Salmonella shuts M & S sandwich factory. Marks and Spencer were forced to withdraw about 80 000 sandwiches due to traces of salmonella being discovered. In another example, B&Q took advertising space of 17 x 25 cm advertisements in National media to say “ Product Recall - stop use immediately - Landia Halogen Organiser Table Lamp. DO NOT TOUCH THE METAL TELESCOPIC LAMP ARMS - Immediately switch off the mains supply at the plug. Return to ….etc. These were examples of defective products. In this next example, again involving a 17 x 25 cm announcement. In this case, there was nothing specifically wrong with the product itself but….KP regret to announce that the above promotion offering free “milk caps” in packets of KP Skips has caused distress amongst some customers because of a small number of instances of people accidentally putting these “milk caps” into their mouths. The advertisement went on to say - “Whilst the product itself is unaffected and completely safe to eat and the “milk caps” conform strictly to European safety legislation, the Company regrets any distress amongst its customers”.
There have been numerous others across a wide range of products including the Moulinex Juice Extractor, Dimplex Heaters - a dangerous risk of fumes and fire, Habitat “jelly” candleholders - a risk of catching fire if the candle burns to its base, Mr Sheen Aerosols - a risk of leakage, Mother & Baby Magazine - a recall of free rattles distributed with their November 1995 issue, an unspecified defect in the rims of Mavic cycle racing wheels, a risk of fracture in Shimano Mountain bike cranks, ‘J’ bolt defects in Flying Trapeze Gym sets and other related products, Micromark fans, Berchet Rocking Horses, Ericsson mobile phone chargers, Kwicksave light bulbs, Servis Rainwave Washer Dryers, Great Mills Bench Grinders, Mascot Electronic mobile phone chargers, and many many others.
It is clear that a Product Liability situation can arise at any time and almost always without warning. It can happen to the most quality conscious organisation as well as the careless. No one could ever say with absolute certainty that it could not happen to them if they manufacture or distribute products into the market place.
What then can be done to protect ourselves? There are four possibilities:-
1. Ignore it and hope it doesn’t happen. (if you are lucky, this is the least expensive approach. If you are unlucky, it could wipe out the company).
2. Attempt to insure against it. In other words pass the risk on to others. But this assumes that insurance is available. In many cases it will not be.
3. Attempt to reduce the risk through the application of the Quality disciplines. Not only will this reduce the risk, the right approach will cut cost significantly as well.
4. Minimise the impact through the establishment of a Product Recall procedure.
With the exception of item 1. The remainder are each affected by the others. The higher the risk, the higher the premiums if indeed insurance can be obtained at all. The risk can be reduced but not completely eliminated through the application of Quality related disciplines. A Product recall procedure can greatly reduce the costs and minimise the effects if the worst happens.
The establishment of a recall procedure is analogous to the purchase of a fire extinguisher: one hopes that it will never be required, but if it is, it could prove to be a life saver.
A Product Recall will always be the result of the discovery of a real or potential hazard which will require the fastest possible location of every item responsible for the risk and its removal from the market.
One would imagine that if we knew where every such item was, then withdrawal should be easy, but experience indicates otherwise. For example, in order to be prepared for such an event, ITT in the 1960s established a scheme through which every purchaser of its steam irons was logged onto a data base. Further more, there was an incentive built in to encourage subsequent second hand users also to record the change of ownership.
Even with this level of information, and only months after the initial distribution of the product and intensive television and other media advertising the company was only able to recover about 70% of the defective products. The remainder simply disappeared without a trace. Attempting but failing to recover such items does not absolve the distributor of his responsibility so any one of the missing 30% would remain a hazard both to the user and to the company.
Clearly, failure to be able to act quickly, or with a disorganised approach, could result in chaos, and the resulting costs will be appreciably higher than if a proper plan is ready. A product recall plan, to be effective, must be capable of instant implementation. In the worst case decisions will need to be made on the recall of products from overseas, the tracing of individual purchasers, television and newspaper advertising, the withdrawal or freezing of stock in warehouse and retail outlets, stock replenishment, the issuing of instructions for the safe use of the product or its disposal, the subsequent monitoring of the health of those who may have inhaled or consumed doubtful chemicals, foodstuffs, gasses, radiation etc, not to mention the internal costs of product modification, replacement, and the administration of the plan.
The State of the Art.
Apart from the pharmaceutical industry, where the Medicines Act imposes certain obligations on licence-holders with regard to withdrawal and recall from sale, British industry at present has little in the way of a formal approach to the problem of product recall.
Owing to the relationship between the manufacturer and the eventual consumer which existed before the passing of s. 6 HASAWA and product liability legislation, there was little or no pressure on Industry to develop recall strategies. It was rarely possible for the ultimate consumer to bring a successful action against the manufacturer, because it was not usual to buy the product from him directly, and no contract of sale existed between them.
Because the impact of Product Liability legislation to date has been far less dramatic than was pessimistically predicted before its introduction, there has been little pressure on industry to consider product recall as a serious risk. Unfortunately, however, as the above examples demonstrate, product recalls are a significant reality.
Further more, the insurance companies do not look favourably on the idea of product recall insurance. This is due to the fact that the most costly recalls are likely to be initiated as the result of a new design fault. The more limited recalls will most likely be caused by a manufacturing fault, such as a mix-up with labels. The insurance companies generally regard innovation as a development risk which should be borne by the company. Consequently, for the unfortunate company confronted with the reality, exposure may be unavoidable.
Should a recall be required, speed of action is of the essence. This is not the time to have long discussions as to who should be in charge and to direct the plan. This should have been decided long before such an event. Ideally, and in general, this responsibility should fall upon the shoulders of the Quality Manager because he will be the most familiar with the systems in operation. However, there will be many specific cases where others may be more appropriate - for example, the Risk Management Department.
Whoever is selected, they will be responsible for calling together a team of qualified people as soon as a situation demands and should report directly to the Chief Executive for this purpose. They will also be responsible for leading the resources assigned to deal with the media should the need arise. In the case of urgent or serious product recalls, which may be costly, likely to damage the companies reputation, or receive any form of national or local publicity, it should usually be necessary to obtain Board approval for any announcements or interviews that may take place.
A recall should always be initiated when it has been ascertained that there is a definite or potential danger to life or health, and when the continued use or circulation of the product is likely to result in legal action. A recall may also be initiated when it is found that a fault in the product, or lack of performance, is likely to affect the reputation of the company and when the lack of performance fails to justify the claims made in advertisements. This would also include drugs, where the dosage advised was found to be at variance with requirements.
The first stage in a recall programme will be some form of notification of a hazard. This could come from a number of sources and in many forms, depending upon the product. The most likely sources include:
(a) The user.
(b) An independent test house report.
(c) government sources.
(d) own reliability test programme or research.
(e) Research Institutes and laboratories.
(f) Overseas sources.
Hazard Evaluation Report.
Immediately a hazard is identified from any source, it will be necessary to evaluate the level of danger, and the category of recall, if required.
Levels may be set as:
1. User level recall. Likely to be extremely expensive. Particularly in the case of consumer products. It would be essential in all urgent cases and probably in most serious cases. It entails recall from users, retail outlets and warehouses. Subsequently it would be necessary to reimburse the owners for losses sustained as a result of the withdrawal. The order for such a recall would always be made at Board level.
2. Retail level recall. Less serious cases to do with probable law breaking, particularly in cases where an existing product has been in widespread use for some considerable time, and now information has revealed a minor hazard.
3. Wholesale level recall. Will entail product recovery from the first stage in the distribution chain and may be applied in the case of non-safety critical situations.
4. Limited recall. Will mean the recall of individual batches and consignments and will probably occur most often when manufacturing faults have been discovered or suspected in an established product.
This evaluation should be carried out by the manager responsible for the recall programme. In any case, it is recommended that a hazard report be prepared and circulated to all concerned within two days of instigation, regardless of the action taken. The report should include a description of the type of hazard, the category, and the recall level. Apart from good housekeeping, hazard reporting provides a good source of data for future risk probability evaluation. As with accidents, near misses are far more numerous than actual events. By recording near miss recalls, the company will soon be able to evaluate the likelihood of an actual event and its consequences.
The report should report a nil return, where applicable, on the following aspects:
1. No health and safety hazard, no action required for product improvement or modification. The hazard evaluation report should state, ‘No hazard expected’, followed by the date and name of the responsible person(s) making the assessment and brief reasons for the assessment.
2. If a hazard is considered possible, all relevant information must be obtained for assessment by the recall team. The evaluation information should state:
(a) Whether any disease or injuries have already occurred.
(b) Whether the use of the product in any particular circumstances could lead to a hazard (ref. Vacwell Engineering Ltd, 1971).This should be supported wherever possible by the appropriate supporting evidence or research data. It should also give the following:
(c) An assessment of the particular segments of the population at risk, e.g. work people in general, work people in specific occupations, the general public, medical patients, people engaged in leisure pursuits, men, women or children, etc.
(d) An assessment of the seriousness of the hazard
(e) An assessment of the probability of occurrence.
(f) An assessment of the consequences of occurrence.
(g) An assessment of the appropriate remedial action required.
(h) Estimated total cost of the recall.
(i) Estimated consequential cost of the hazard.
(j) A statement signed by the Chief Executive authorising the initiation of the recall.
Product recall initiation.
Upon receipt of the agreement by the Chief Executive or his nominee to initiate the recall programme, a hazard warning notice may be required. This should contain all the information necessary to bring the risk to the attention of all those likely to be exposed to the hazardous product and the circulation may include:
1. Field specialists including service staff.
2. Appropriate Government or local authority departments.
3. Police and other emergency services.
4. Press, local and national where applicable.
5. Radio and/or TV, local and national where applicable.
6. Retail outlets.
7. Customers, where possible.
8. Trade journals and Trade Associations.
9. Professional Institutions and research Establishments.
10. Export agencies/ licensees, foreign Governments, etc.
11. Relevant test houses.
Depending upon circumstances, the warning notice may include:
(a) A statement of the hazard and likely consequences.
(b) Clear instructions and illustrations to enable the repair, correction, removal, immobilisation or shielding of the product.
(c) Warning or hazard signs, which may be detached from the notice and attached to the product, and its container.
(d) Instructions appertaining to the return of the product, and a recall return card, preferably reply paid. This could take the form of a tear-off portion of the letter.
Recall effectiveness assessment.
Once the recall procedure is under way, it will be necessary to assess the effectiveness of the procedures adopted and to consider the possibility of stepping up the campaign. Similarly it will also be necessary to make a decision to terminate the recall when all possible items have been returned.
Post - recall investigation.
The initiation of a product recall must be regarded as a catastrophic failure of the quality plan. Even in the case of a limited recall, entailing the return of a small number of items from a retailer, there must be a post mortem. The post mortem should be conducted by the most senior personnel from the departments implicated, headed by a senior executive and supported by the quality manager. The objective will be to determine where the organisational weaknesses are that allowed the event to occur. Failure to do so means that potentially the same thing could happen again. This should not entail a witch hunt. Except for specific cases where negligence may have been involved, the problem is more likely to reflect on the management system in place than on the actions of any individual. The most effective approach is to examine the management processes to identify where the weaknesses occur.
The cross functional project by project improvement process when correctly applied is the most effective method for resolving such problems. However, typical places to look for causes are:
(a) Lack of relevant training.
(b) Insufficient reliability testing of prototypes, components and pre-production models.
(c) Inadequate research, in the case of products using chemicals; insufficient knowledge of existing data (Vacwell Engineering Ltd, 1971).
(d) Lack of co-ordination of production feasibility before release for production.
(e) Inadequate specifications for the testing of incoming materials or components.
(f) Lack of co-ordination between design and other departments, particularly market research, reliability engineering and production, resulting in the production of inadequate specifications.
(g) Too much pressure to rush designs through, with the consequence that short cuts are taken.
It is hoped that the prospect of being faced with the consequences of having to implement the procedures described in this article will be sufficient to convince everyone of the importance of attempting to “get things right the first time”. During recent years the number of major product recalls seems to have been on the increase, and by now must be a major source of worry to industry. Quite apart from the actual costs of managing such a project, the damage to the reputation of the company can be lasting. Also, there is the question of negative advertising. Most organisations spend enormous sums attempting to convince the world that they are the best. Now they have to spend the same sums of money explaining that their product is actually quite harmful. Even with the most intense application of quality related concepts the risk is never zero. The Ford Motor Company for example is probably one of the biggest advocates of Failure Mode and Effects Analysis (FMEA) and even they get it wrong sometimes! Maybe Six Sigma or Lean Sigma programmes are the best form of defence?
David Hutchins is a Board member of the Quality Methods Association and obtained his Masters Degree at the Department of Engineering Production at Birmingham University. He specialises in the implementation of Business performance improvement concepts in a wide range of countries and cultures.
David Hutchins can be contacted at David Hutchins Innovation Limited by using the CONTACT US button at the top right hand corner of this web site.
Ref:My Docs\Articles\Product Recall.doc. For Quality World. Written Feb. 1998.
This topic is also included in the DHI courses leading to the CQI Diploma in Quality Management