State-of-the-art problem-solving methodologies include trend charts (o
r run charts) and Pareto analyses. A trend or run chart represents dat
a in its time sequence. The user of a trend chart typically looks for
trends the data from time to time, such as gradual improvement or dete
rioration. The user also looks for spikes and dips in the data to iden
tify any sudden performance changes. An inherent problem with using tr
end charts that the data is treated as if it is deterministic. Failing
to understand that each data point represents one realization of a ra
ndom number leads incorrect conclusions. Actions based on these conclu
sions increase variation in the process and cause more product failure
s.