KPIs

KPIs, or "key performance indicators," are rarely actually data sets that are good indicators of performance, because such data is most often not easy to identify and/or measure. Instead, KPIs in reality are usually facile data sets chosen solely for their ease of measurement and ideation (copying whatever others use in an industry) - a more accurate name for these would be "facile non-indicators of anything" (FNAs).

They are also a symptom of data disease wherein leaders feel empowered by pretending to be data scientists, pointing to numbers and graphs, but without really grasping what data is and how it should be used. A basic concept in data science is "garbage in, garbage out" - if your original data is worthless or questionable or unrelated, then even the best analyses of said data can only be worthless or questionable or unrelated as well.

Today, KPIs pollute the minds of executives everywhere, and they remain the bane of many employees, but the word "KPI" is not old. According to Google's Ngram Viewer, the words "KPI" and "key performance indicator" only appeared in the late 1990s, starting to gain popularity from the early 2000s. Their older cousin "scientific management," however, was extremely popular from much earlier, peaking in 1913; certainly Frederick Winslow Taylor's popular book, The Principles of Scientific Management published in 1911 influenced this. Taylor mostly considered manufacturing and manual labor (e.g. shoveling), where production was easily quantified, and in these cases, providing bonuses to those who produced more predictably resulted in more production.

But scientific management had roots in another industry - slavery, which many workers compared Taylor's system with in 1911. The very popular Cotton Plantation Record and Account Book published in 1847 or 1848 by Thomas Affleck had detailed instructions for plantation owners to:

• Track cotton production of each slave (per acre, per bale)
• Evaluate the asset value of each slave ("negroes accounting")
• Depreciate values of slaves due to health or age (e.g. at middle age)
• Providing bonuses to productive slaves and dealing punishments to less productive slaves

Today, modern corporations sadly sound quite similar, but repetitive labor like cotton-picking is much easier to measure than the complex and unstructured work that modern corporations today focus on. Irrespective of what is being measured, employees' negative perception of the slave driver's whip and stopwatch has not changed.

More often than not, executives push and shove their KPIs down employees' throats, feeling as they have accomplished something - but what is the result? Certainly not increased performance.

• Moral damage: distrust of leadership, (accurate) fear that the new KPIs will be used to "catch" and punish staff.
• Motivational damage: disconnect between arbitrary KPIs and perceived actual value will foster unhappiness; simplifying people's contributions into mere numbers will also generate the perception of treating people as robots, drones, or slaves.
• Perverse incentives: KPIs very often create perverse ways of attaining "success" in the arbitrary data points chosen to define success - for example, if a writer is paid per word, then of course the writer will naturally write wordy, long-winded documents; however, wordiness is not real value.
• Material damage: increased turnover as staff decide to move elsewhere; few people enjoy being treated like slaves. Effectiveness falls as well, as executives fervently focus on the easily measurable aspects, while discarding valuable aspects that are not easily measurable, despite their importance.

It is important to remember that these methods were originally developed for managing slaves, who had no ability to quit or resist. It is also important to remember that data naturally predisposes people to focus upon it, thus overlooking information not included in the data. The ease of measuring something in no way is related to how important it is.