And what it fails
The performance analysis
Big data has entered into every facet of the industry
and today it is being used to gauge people, so that a
company can hire and retain the best ones. Many
companies are mainly using big data and analytics:
• To predict who can be the best person for a given job
based on his skills, experience and the job
• Predict how well a given person can do a job, based
on his prior experience.
Big data is only as good as the questions being posed to it. Even the
right kind of data can yield wrong results when we don’t ask the right
Say we were analyzing the performance of an employee over the years.
Instead of focusing only on his skills and deliverables, we also need to
• What kind of boss did he have: supportive or demanding?
• How conducive was the workplace atmosphere?
• Did they impact his failures, or did the employee perform well despite
• Managers who focus too much on performance force
people to work hard to meet quarterly targets, while
destroying the emotional climate that sustains the
life-blood of any organization.
• On the other hand, leaders themselves work the
hardest and inspire others to do more too. In such an
environment employees are more interested and
involved in their work and productivity increases.
Each of these atmospheres have a huge impact on
employee performance. So performance analyses
should include such factors too.
How do we know what is the metric that drives the success or failure
of a person?
What is the impact of his surroundings on his failures?
What are the differences between his successful and futile projects?
• The questions we could ask, the directions in which
we could think and the possibilities are infinite.
• We cannot design a customized algorithm that takes
care of all the possibilities and can analyze the
behavior of every person, to predict his performance.
• Even the best of algorithms have loopholes, due to
which we risk losing good employees or hiring bad
Would you fly in a plane that you designed,
even if you were the best?
• Employees at Google refused to using such predictive modeling based
on data analytics in order to hire and promote employees.
• This wasn’t because the algorithms used were black boxes; but
because Googlers knew so much about them that they knew it could
fail horribly sometimes.
• They cannot be trusted to predict success or failure of a person in a
The best practices
• The best analyzers of people are people themselves.
So when we want to know how well a person can do a
job, its best to ask people who have worked with him.
• It’s important to add that humane touch to our data,
because mathematics can take us only so far. In order
to cover the last mile and truly understand a person,
we need to take that leap of faith.