Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

The

"People Lie, Numbers Don't" Approach to HR Analytics



- Sumeet Varghese

A few workshops and a couple of speaking engagements later, I find it immoral
not to update what I have written earlier on the marriage between HR Analytics
and Big Data.

So here goes:

1. What is the future of Datafication in HR? Does HR already have a lot of
Data?

The idea that Data (of the wide and "wild" variety) is required to run any form of
Analytics (Big or Small) has not really caught on with "some" HR professionals,
atleast the ones I have spoken to. When I say "some" you may take it to mean
"many", since I always keep statistical sampling requirements in mind whenever
I strike a conversation with anyone in HR (a "few' therefore might just represent
the "many" out there, If I am right about my sampling). This means that when I
speak to you, I always consider you first as part of an important data-set - either
by virtue of your title, your experience in the HR function or your HR
caste/"gotra" (XLRI, TISS, IIM and so on). I know you may not like this statistical
approach to meeting and conversing with people selectively and treating them as
nothing more than "samples", but trust me, this approach is not as demeaning as
the kind of classificatory or analytical exercises you perform by asking me which
state I am from, which religion I practice or whether I love Modi as much as you
do or not. You see we are always collecting data from each other, whether others
like it or not, but what we do with that data in HR is something I will reserve for
another post some other day. For the moment though, let me clarify: HR
Departments are caught in a position where they may have too much of data and
very poor analytics capabilities to leverage the data or strong analytics
capabilities but hardly any worthwhile data to mine for insights. Frankly, I have
no clue which one is better!

Anyways, the sad/good news is that we need to sort out many serious data
related issues before we can discuss the extent to which our neighbors in
Marketing, Customer Service, Finance and Operations use Big Data Analytics:

Is the Data we capture of any value at all? If it is not valuable today, will it
be valuable tomorrow?

Unfortunately, no one has an answer. In one case, I asked a group of HR
professionals whether we should track the number of loo breaks that senior
executives took during a workshop and whether it would serve any purpose.
Obviously, there were blank stares. My question of course wasn't pointless. The
loo like the office water cooler and the cofee/tea vending machine points in an
organization is as much a space for exchanging workshop feedback as it is for
updating each other on some juicy company gossip. While this may be a small
unwanted detail to be avoided by the HR professional it might certainly be of

interest to a Data Scientist, assuming it can offer some interesting clues -


Elementary, my dear HR Professional, Elementary! Strangely, with a little bit of
luck I was able to work out a correlation (not causation) between the pathetic
condition of the loos used by the top management at one firm and the MD's
constant refrain that the organization lacked "ownership". In fact, nobody (and
this included the organization's top brass) bothered to complain about the stink
because they thought it wasn't "their job".

Should the Data we capture be of the same type or of different types? Aren't
we folks more qualitative rather than quantitative?

Frankly, HR professionals must closely study the kind of work being done by
their Marketing and Customer Service Analytics teams to figure out that Big Data
Tools have evolved to a point where they rarely ever care a byte if your data is
structured and/or unstructured. So, if you have an employee's leave records in
XL, the poor unsuspecting chap's FB Posts and Tweets for a full year in Word, the
person's performance appraisal history in PDF, and his/her Compensation Data
in any format that your ERP spits out, some meaningful analytics can still be
derived even if this employee record is a gibberish amalgamation of data. For
instance some recruiters have long studied behavior patterns of candidates
before, during and after the various stages of a screening process that they have
been subjected to. Thanks to these studies we now know that if you handle a
particular stage of the recruitment process poorly, candidates are two times or
three times more likely to badmouth the company's products and services on
social media. Obviously, in this case process feedback (quantitative and
qualitative data) at each stage of the screening cycle has been correlated with
social media behavior (qualitative data) of the candidates.

Why should we capture Data?

Asking this question (considering the order in which it appears) is a bit like
placing the cart before the horse (line managers will vouch this is precisely the
image of HR they have). Unfortunately, many of us in the HR fraternity have
fallen in love with the practice of hunting for data only when businesses want it.
Also, quite understandably, businesses don't explain why they want the data?
They know they can do the analytics themselves because we haven't been able to
figure out why businesses ask for specific types of data in the first place. At some
workshops, consultants have asked me why clients look for specific data to be
pulled out from either an ERP and/or freely floating electronic and physical
folders containing all types of files (some including data captured on papyrus - I
am mildly exaggerating, here). My answer: that's just their way of checking
whether we are busy or not.

2. What kind of new/nice Analytics are we going to see in the near or far
future?

People Models

We all know Google has done some heavy number crunching over quite so many

years to figure out 8 attributes they would love to see in Googlers who manage
other Googlers (please do read the HBS Case Study: Google's Project Oxygen) and
that this model feeds their recruitment and succession planning processes. Its
quite possible this people model might be revised in the next 10 years as new
challenges emerge with the business and the model (not necessarily in that
order). I remember studying Van der Waal's equation in school - the final
derivation of the equation, which obviously had more variables than the one
initially proposed, was developed to fit the "reality" out there because
tests/experiments revealed the equation had not quite nailed it. If People Models
are "work in progress", People Analytics Departments can rub shoulders with
their scientific peers - if not, such models run the risk of being exposed by a
Copernican revolution (which obviously would happen on the business side
first!). We do know for a fact that the famed/notorious 25 layered (rounds)
screening process (possibly, state-of-the-art at that point in time) at Google gave
way to a 4 layered (rounds) screening process partly because business managers
wanted "good" people in "quickly". I am assuming, Van der Waal was under no
such pressure.

Operational Experiments

Google did a great job of experimenting with plate size to figure out an optimal
shape that could meet its target of kicking employees back into shape (guilt and
shame worked powerfully to reduce the number of trips employees made to fill a
small plate) and help them reduce their calorie intake. I have seen such
experiments to control wastage of food during lunch breaks. At one
manufacturing firm, the HR Department set up a Scoreboard to show how many
kilos of food was wasted the day before and so on. Obviously, such loud displays
helped control the menace to an extent. At another place, a young engineer
decided to stick graphic photos of poor children dying of hunger right next to the
serving area. Consequently, people got the message and while some folks
attributed their loss of appetite to the pictures some said it made them more
sensitive about the quantity of food they loaded on their plates. Unlike the
operational experiments Google undertook, the examples I cite may not have
been the result of any meticulous planning, rigorous measurement or even
continuous experimentation. At the same time, I cannot help but point out that
HR is expected to change employee behavior in numerous ways for a variety of
reasons. That, to say the least, is exactly what HR is expected to do (if we hear
our Line Managers correctly). Therefore, If earlier, HR did not have the tools to
study, analyze and mold employee behavior, thanks to Big Data Analytics it now
has a wide and bewildering array of tools that have the potential to predict and
regulate employee behavior, on a mass scale.

While most of the examples here pertain to food, I am hopeful that Operational
Experiments in HR will extend to other more promising areas of employee
experience as well. I remember the case of a "desi" (no HR Degree / no Strategic
HR Experience) HR Head who was asked to hire a Costing Manager. The
company he works for has a lone manufacturing unit outside Delhi. Once the
Costing Manager was on board, the company realized he had no job since he
needed data on work in progress - timely data on finished and unfinished goods

and inventory, almost on a daily basis. As the company did not have an MIS or
any practice of tracking anything remotely called " operational and production
data", the HR Head secured permission from the MD to circulate chits of paper to
collect such data from the company's 300 odd employees (all semi-skilled) at the
end of each working day. In exchange for 10 rupees every day, each employee
was asked to accurately mention on the chit the quantum of stock they were
sitting on. The scheme went down well with the workers and the Costing
Manager discovered he had enough and more data to occupy himself for a full
year. The MD was so pleased, he decided to increase the amount to Rs. 20 per
day. If a Desi MIS can be generated on the fly through an operational experiment,
I am sure HR can conduct many experiments to help businesses unlock value
from Data.

Dashboards and Visualization

For HR Departments that continue to labor with PowerPoint and XL, software
like Tableau and Sisense (not that I am in love with these) can appear to be the
proverbial oasis in a desert formed by data. They can make data analytics
visually stunning and beautiful and for a change, even make business leaders fall
in love with HR. However, these are low hanging fruits on a long journey. The
primary objective of an HR Analytics Department cannot be the creation and
transmission of Dashboards and Data Visualization - although these can greatly
help Line Managers to arrive at their own inferences and conclusions, especially
where they doubt HR to offer some stellar insights.

HR Metrics

Dashboards are made up of various kinds of metrics. Thanks to the "proliferate
or perish" treaty that HR Professionals became signatories to sometime in the
past decade, various types of HR Metrics (in the order of 1000s) are available
today with leading ERP vendors. Someone recently claimed they have developed
3000+ HR metrics to track - now that's taking this proliferation business a bit too
far. Unfortunately businesses don't share HR's love of metrics. Moreover, what
irks them the most are the totally different ways in which teams within the same
organization measure the same metric. Recruitment alone throws up various
ways to measure an important metric like "time to hire" depending on how
exactly you identify the base line. Worried probably by the confusing signals the
HR fraternity was sending out to the business community, SHRM instituted
standard ways of measuring some common metrics like Cost of Hire and so on.
However, I really wonder how these standards can be applied across
geographies or even industries.

3. What kind of skill-sets will HR professionals of the future need thanks to
Big Data?

If Big Data Analytics is taken to its logical conclusion by "illogical" (I'll explain
this in a while) Departments (Whether, they be IT or Operations or even HR), HR
professionals won't be around and the best part, HR skills won't be required. I
and a senior friend facilitated a workshop recently for a group of finance

professionals. Everything from our travel and stay onwards to getting the
participants to the venue from various regions was seamlessly managed by the
Finance team. We were personally shocked (truth be told, we had mixed feelings
and didn't know whether to laugh or cry) to not find a single HR professional
play a role anywhere from need identification to vendor shortlisting and
screening to trainee coordination to venue booking to feedback collection. When
we left, the chaps said they have more work lined up for us - just that we would
have to re-title the entire intervention to avoid detection by the company's HR
Department and to prevent generating the impression that Finance is stepping
into HR territory. Already, many traditional HR processes (requiring what was
traditionally termed "HR skill-sets") are either being outsourced or perhaps (in
the case I recounted) taken over by line functions. Now if the rest (which is not
much - though some serious-minded folks might give it some meat and call it
"strategy" or "business partnering" or "talent management") is to be carefully
considered, automation will slowly catch up. Google's recruitment algorithms
have done away with the need to have a hiring manager for some positions. At
the same time its retention algorithms help it to predict who is likely to leave.
While I do not immediately foresee job-destroying algorithms to entirely replace
a generation of HR professionals, I am hoping a new breed of HR professionals
with algorithm-dismissing/refining skills will be able to find their feet in the Big
Data landscape. It is quite possible, that the HR Professional of the future will be
more analytical, a wee bit statistical, certainly programming friendly as well as a
domain expert having a more integrated view of HR.

Whither Psychometric Testing?

Some time back a well-known Psychometric Testing company that offers various
types of IT platforms to conduct millions of tests online came up with an
exceptional (and I must add "sensational") claim about the poor quality of sales
"talent" in leading B-Schools in India. They claimed they had used tests that were
rigorously developed. As an old hand at psychometric testing, I knew more than
to believe such claims. Interestingly, companies the world over are looking to IT
to bring them platforms that "shorten" the process of screening thousands. I
don't care, a gentleman told me once, what you do with my people, as long as
there are numbers and reports. I am hopeful Big Data Analytics will take on these
charlatans and their acolytes in interesting ways. There is already encouraging
research to show that your FB behavior, if analyzed well, can successfully predict
your Big 5 personality traits - so the good news is we may not need psychometric
tests in the future. But the sad news is our data footprints will be stored
somewhere to be analyzed someday. I won't be surprised if my Google Calendar
and Map are combined and analyzed someday (they already have, by the way) to
tell you how I handle projects (how many meetings), travel, stay, and money
(based on bookings data probably integrated from some other source). However,
I must concede that those data-sets along with others should provide more data
points to take a judicious people decision than simple answers to a set of
questions hosted on an IT platform designed by an IT team that hasn't the
faintest understanding of scoring and interpretation.

4. What are some common "heuristics" Line Managers are known to use for

hiring/firing and everything in between?



Across organizations of all types, you will bump across various types of people
heuristics (unexamined People Models) at work - rules of thumb that may have
served line managers well and which play a very large role in several people
decisions at a firm (contrary to what we HR professionals think). This is the
human version of Big Data Analytics at work, perhaps! From among the few that
I have been able to identify, I find the one involving a senior finance manager at a
large Indian firm, quite interesting. This gentleman hires juniors who meet one
criteria: they should have cleared their CA (Chartered Accountancy qualifying
exams) in the 3rd or 4th attempt. The way he sees it, such candidates are ready
to stretch more than those who have cleared it in the first attempt. Now, anyone
who has taken the exam thrice will confess that the pain of preparing for the
exam three times over and clearing it can be excruciating indeed. Whether that
preparation makes them more industrious and persevering (atleast, in the eyes
of this finance professional) is a matter of debate for statisticians and
behaviorists alike. While there is no independent study out there that can
establish whether these 3-timers are more persevering and industrious than the
first-timers, our finance manager continues to operate on the basis of this
heuristic and what is more, over time, has been able to build and retain a team of
productive professionals using the same logic.

5. How can HR Analytics explore such "heuristics" that drive people
decisions in a firm?

Every people heuristic is a fit subject of research for a budding HR Analytics
professional. Armed with statistical tools, behavioral analysis models and an
understanding of how people form perceptions about groups and individuals, an
HR Analytics Department should statistically examine those "notions" or
"assumptions" about people that might be actually preventing organizations
from attracting, hiring, promoting and retaining talented people.

One gentleman at a leading telecom company confessed using a particular
heuristic to screen out candidates: he would ask the candidate to share his/her
contact numbers during the interview. If the candidate used the services of a
rival telecom operator (as would be evident from the number he/she provided),
he/she would be dismissed from the interview. My friend's logic (based
obviously on years of experience - Big Data Analytics) for the summary rejection
is based on the idea that such individuals are never loyal to the brands they work
for. If they were, they would avail their company's services and not that of a rival.
In his scheme of things, people lied but the numbers didn't. If our Big Data
Analytics program operates on a similar premise: people lie but numbers don't,
we risk repeating the same mistake that my friend from the telecom sector
makes. You see - my friend never asks where the candidate lives and more
importantly, whether this place has adequate network coverage or not or
whether the area in which the candidate usually operates has a good number of
telecom towers for his customers or not.


About Me:

I head a Human Capital Management consulting firm based out of Ahmedabad in
India. We provide a wide range of HR Consulting and Training solutions to
clients across India and the Middle East.

It might interest you to know that my public and in-house Workshops on Big
Data and HR Analytics offered in several Indian metros since 2014 has been a hit
with Senior and Middle Management HR Professionals and Consultants.

I have offered more than 10 editions of the Workshop in almost all metros across
India and have had participants from companies like Aon Hewitt, White Spaces,
Pricewaterhouse Coopers, Greaves Cotton, Titan, Capgemini, HCL Technologies,
Maharashtra Natural Gas, Tech Mahindra, Toyota Financial Services, J K Lakshmi
Cement, Kotak Securities, Welspun India, Nestle, Subros, Toshiba, Sun Life, L&T
Shipbuilding and several others.

I have also had the honor of facilitating a Workshop on Big Data and HR Analytics
for the entire HR Department at Atul Ltd. and Claris Lifesciences Ltd.

Moreover the feedback has been terrific...

Jyotirmoy Bose
CEO, White Spaces
Sumeet brings simplicity to a complex emerging field like HR Analytics!

Divya Singhal
Manager HR, Aon Hewitt
Good and effective training .

Shalini Kapoor
Manager, PwC
Well-conducted Workshop. Extremely helpful for HR Professionals in a Business
Partner role.

Sanjay Shelke
Assistant Manager, Greaves Cotton
Trainer presentation skill is very good. Overall, session was very excellent.

Bhaskar Joshi
Chairman, Membership Services Committee, Bombay Management
Association
Thank you very much for your support. The Workshop on "HR Analytics - Turning
HR into A Decision Science" would have not been possible without your help. The
feedback from the participants was very positive. My deep appreciation for all that
you have done


Surabhi Sahay
HR Specialist, Aon Hewitt
Good introduction to Big Data.

Should you wish to know more about the Big Data & HR Analytics Workshop,
here's a link to our current advert: http://tinyurl.com/lu99bhd

In case you'd like to meet me to discuss this further or any of our other solutions,
feel free to call/mail.

Kind Regards,
Sumeet

We work on Behavior - the Space b/w Strategy and Outcomes

Sumeet Varghese, SPHR, SHRM-SCP

Founding Partner, Your HR Buddy


11- Krupa Shree, Ghosha Society Road,
Jai Ambenagar, Thaltej
Ahmedabad - 380 054
Phone: 91-9998822978
W:www.yourhrbuddy.org
L: http://in.linkedin.com/in/sumeetvarghese

Public Workshops from July to September:

The HR Auditor Masterclass @ BMA:
https://app.box.com/s/bv4lk7rzflu0q99j2l6y0mf9in9e822w

Big Data & HR Analytics: http://tinyurl.com/lu99bhd

HRCI Certification Exams Prep Course (PHR, SPHR, GPHR, HRBP,
HRMP): http://tinyurl.com/ow5sjvm

You might also like