Are safety KPIs counter-productive?

A colleague who works with a major international oil and gas company said “Safety KPIs achieve the exact opposite of their original intent”. Whilst he was undoubtedly being provocative to stimulate internal debate, the basis for such a bold statement can be seen in organisations that have comprehensive reporting mechanisms but still suffer from major safety incidents.

Safety performance metrics face an inherent difficulty in that managers have a strong temptation to make their numbers as good as possible when they report their performance.  This will be true for safety just as it is for production, sales, financials and any other business performance metric.  Unfortunately, this tendency is the exact opposite of what is required for good safety performance. 

A common theme in many of the major catastrophic industrial incidents is the fact that warning signs were there but were not reported.   If there is a bias that emphasises the good and glosses over the bad, there is likelihood that the first indication senior management get of problems will be too late.  The report of the Baker Panel into the explosion at BP’s Texas City refinery in 2005 highlighted the fact that clear warning signs were not acted upon and that filtering and softening of bad news from the plant to BP’s executive management was a contributor to the lack of priority given to solving the refinery’s many safety issues.

Trust plays a big part in the willingness to report bad news upwards.  Two factors will be barriers to open and honest reporting of safety related concerns in an organisation:

·         a history of senior levels reacting badly to negative news

·         a history of reported problems being ignored

Both will add further bias toward reporting good news and hiding bad news.

Therefore, there is a danger that the bias toward reporting good news will undermine the objective of safety performance metrics:  to provide information to senior managers to make decisions to improve safety, such as increasing capability and capacity of staff, increasing capital investment and increasing maintenance budgets.

This bias towards good news in safety KPIs is aided by the lack of rigour often found in the measurement of safety metrics.    Whether incidents or near misses (or near hits as they perhaps should be called) are reported will remain a function of how much importance the whole organisation places on improving its safety.  It is the organisation’s culture, its values and the behaviours they drive, that will determine the diligence with which safety metrics are reported.

·         Does the foreman view an injury to one of his team as an opportunity to learn, for him and the whole organisation, or something negative to try and avoid reporting if possible?  Are injured employees encouraged to report to work to ensure that there is no lost time incident to report?

·         Will the Maintenance manager accept an unplanned valve overhaul onto his already overloaded critical maintenance schedule?   Or will he try a quick patch and tell Operations it’s good to go to avoid having to ask for more money to complete critical work?  How will the plant general manager react to the double impact of missing the critical maintenance KPI and having to spend more than budgeted to fix the valve?

The use of targets for certain lagging safety performance indicators, such as lost time incidents, injuries and losses of containment, requires careful handling for similar reasons.  That they need to be measured is not in doubt, and is often a legal or regulatory requirement, but the only morally and logically correct target is zero.  Targets for incidents like injuries or deaths can be used to focus an organisation in a situation of chronic underperformance, but once performance is within acceptable, industry average ranges they should replaced with an ambition for zero.

For safety KPIs to achieve their intents the following is required:

Management systems: as a foundation the requirements for safety reporting and KPIs must be clear and unambiguous to all and there must be some provision to verify the accuracy of reported information

Trust:  all levels of the organisation must trust that bad news will be received in a constructive way and that it will not lead to a witch-hunt

Learning and continuous improvement:  all incidents and near misses have to be viewed by the whole organisation as opportunities to learn

Leadership: leaders need to visibly demonstrate the behaviours that encourage open and honest reporting of incidents.

Tackling these areas will set the cultural basis for safety KPIs to achieve their original intent: visibility of real frontline safety performance that then provides an accurate picture of the organisation’s risks.