2. INSIGHT 1
Many businesses are reaping rewards from big
data analytics.
But there are also some areas of
disappointment. Experts caution that big data,
like any other, is only as good as the questions
being asked – and that some algorithms can
make unhelpful assumptions.
3.
4. There’s a saying in the sciences, “Statistics
means never having to say you are certain.”
In any massive data analysis, for instance,
there will be random correlations that look
“significant” but actually are noise, not signal.
5.
6. Big data needs a hard outcome metric for
performance, but the most readily available
metrics may not actually be the most important
variables in organizational flourishing.
7.
8. INSIGHT 2
An outcome metric like an executive’s earnings
performance, while ignoring his role as a boss
and his impact on the morale, loyalty, focus,
and stress levels of his direct reports, may
result in a false indication of who’s really the
best boss.
9.
10. EXAMPLE OF GOOGLE
At Google, the bastion of algorithms emerging
from giant data sets, are refused by engineers
to use as a method to decide on promotions.
As Laszlo Bock, head of hiring at Google
explained, the very fact the company knows so
much about algorithms lets it see their limits.
11.
12. The assumptions built into a test can
themselves be biased against certain traits and
so discriminate unfairly.
13. MANAGERIAL RELEVANCE
A manager – like the demotivating petty
tyrant–can force his people to work hard to
meet quarterly targets
Using an outcome metric like an executive’s
earnings performance, may result in a false
indication of who’s really the best boss.
14.
15. MANAGERIAL RELEVANCE
But the biggest objection comes from the fact
that the strongest predictor of a person’s future
behavior is their past performance itself.
And that performance gets evaluated best by
people who know that person well.