How should you show incident data on your bowtie risk assessment? If done the right way, it will give you great insights into the quality of your safety management system as well as the quality of your risk assessment. However, if done the wrong way you miss out on the power of bowtie which will make the transfer of information to your target audience a lot more difficult.
DO: improve the quality of your bowtie with incident data
Incident data provides you with great feedback on the performance of your barriers, but it also provides you with feedback on the quality of your bowtie diagram. There are two things you can expect to encounter when displaying incident data on your bowtie:
- Container elements
- 1% elements
Container elements are elements that are too generic. For example, operator error or bad weather conditions. In an incident, these event descriptions are more specific. Therefore, it might be the case that events like ‘slippery road conditions’ and ‘poor visibility due to fog’ are actually more detailed threats of the higher-level threat ‘bad weather conditions’.
TIP: If this happens, try to review the incidents and check if you can split ‘bad weather conditions’ up into more specific threats. Doing this, will help you define your barriers and will also result in your barriers becoming more detailed.
1% elements are threats or consequences that almost never occur. For example, ‘a hurricane’ causing a loss of containment. For some organizations, there is only a 1% chance this will happen. Thus, you will probably never see any incident data of the barriers linked to the hurricane threat in your bowtie.
TIP: Try to cover this while you are setting the scope for the bowtie. Even though an event or threat is highly unlikely you might want to cover this in the bowtie. However, if you have genuine reasons not to include it, for instance, if the frequency is simply not high enough to fit the scope, you can leave it out of the bowtie.
DON’T: limit your incident analysis to just bowties
We are not saying it is not possible to analyze incidents with the bowtie. However, it is not the purpose of the bowtie. Bowties are used to create an overall picture of the risk an organization has to deal with. A ‘finished’ bowtie will show various incident scenarios, their causes and what the organization has to do to prevent this from happening.
Incidents, however, are on a much more detailed level because you want to know all the very specific details about these incidents. You can get the basics of the scenario from the bowtie, for example, one threat leading to one consequence and the barriers in between. An incident investigator wants to know why these barriers failed and digs deeper until the underlying cause of the barrier failure is found. This detailed level of data is not necessarily shown on the more generic higher-level bowtie. In fact, doing this could make your bowtie overly detailed which, in the end, can lead to having up to 100+ very specific (perhaps too specific) bowties.
TIP: The advice we would like to give is to use bowties for risk assessments to understand what can go wrong and use incident analysis, like Tripod Beta, BSCAT or Barrier Failure Analysis, to dig deeper into why things went wrong. Use the combination of both to aggregate incident data on your bowtie to gain understanding of the reliability of your barriers, by displaying their performance.
DO: assess your barriers with a combination of data sources
Incident data on your bowtie provides new insights for risk-based decision making. It is important to look at your bowtie with more than one data source. Incident data only shows barrier failure counts (and some effective barriers in case of near misses) and this might give you the wrong feeling about your barrier performances.
TIP: To get a more realistic idea of your safety system, you can combine incident data with for instance audit data. With AuditXP you can proactively ask people that are part of your safety management system, questions about your barriers and see the results (in combination with incident data) on your bowties. This allows you to see failure modes, but also strong barriers. Combining incident and audit data will enable you to proactively find the success factors of your barriers.
DON’T: take away the attention from your incident data by displaying too much information at once
If you try to communicate a bowtie you might find it difficult where to start. Often there is so much information included in the bowtie that you simply need to divide the information into batches, so it becomes comprehensive for the target audience. Therefore, try to think of a bowtie like of a story you would tell. Every leg of the diagram (from threat to consequence) is an own story. By taking this approach you will find it easier to take the audience on a journey they are able to understand.
Example of a bowtie with perhaps too much information
TIP: Use the display profiles (show different pieces of information), filters (hide irrelevant information) and expansion buttons (begin high over and then proceed to a detailed view) to collapse the diagram to only the part of information relevant to your story. This will make it possible to really set emphasis on the core of what you are trying to communicate.
Curious to learn more? Contact us .