Conformance is how well a process performs within its CTQs and specification limits. Variation is the opposite of it- the difference in the process output over time. Variation is often caused by elements which are part of the process itself (common cause), or elements which are external to the process (special cause).

Common Cause variation is effected by elements within the process. They are mostly random and independent of each other; conforming to the normal distribution curve ie stable. However, a stable process does not mean good or satisfactory output- it merely means the process is consistent within specification limits and the variations are contained within these limits. In the formula y=f(x), there are many x’s but low impact x’s (described in iSix Sigma’s dictionary). Tweak the process to get improvements in output.

Special Causes of variation are arising from unusual circumstances. They are not random, do not reflect historical trends and is not normally distributed. If Special Cause variation is detected in a process this process is considered not stable. In the formula y=f(x), there are few x’s but high impact x’s (described in iSix Sigma’s dictionary). It’s probably not a good idea to change the process first, but dig deeper into the root cause of this kind of variation. Tweaking the process won’t make the special cause variation disappear because these variations may have nothing to do with the process. Introducing a process change may result in worse variation than before -the normal ‘ups and downs’ of the process may turn into irregular spikes.

Hotels sell ‘moods’ and ‘senses’- it’s inevitable these elements influence comment forms or process output but what’s the connection between variation cause and customer experience? This brings us back to that particular set of dots outside specification limits in that project run chart that had the food & beverage team bewildered. If the Five Whys and Cause-And-Effect can’t provide a direction, it’s time to look at the ‘history’ of the guest who contributed to the data point. Were there any complaints logged on this particular guest prior to this project’s process? In the hotel industry, the guest goes through processes back-to-back. Any ‘unhappiness’ or ‘joy’ from the previous process is retained and brought forward to the next process, the degree which depends on how the guest interacts with hotel employees, hotel products and the customer recovery process. For instance, the guest ‘suffers’ a ‘bad’ check-in (room not ready & guest waits unreasonably; assigned a smoking instead of a non-smoking room, failed elevator, etc.) and is still unhappy at dinner time gives a poor rating to the food & beverage service. Properly identifying data points like this tells a clearer picture of project performance- what to change and what not to change.

About the Author