Marcus Baker: People Analytics at Scale
People Analytics Conference 2022 Winter
6. Lay a strong foundation
Bring all the players to the table. Truly
understand what they need and get
them to understand what it will take to
Have end-users, data stewards, and
subject matter experts elaborate on what
data they have and what data they trust,.
Make complex thing
A complicated, hard to interpret, and
convoluted to measure, metric is often of
less value to the organization, even if its
Simple metrics sell. Simple metrics
drive action. Simple metrics connect
data to reality.
Say no to scope creep
Agree to a path, develop a plan &
timeline, and stick with it
Your resources are limited so its better to
build a high quality, narrow scoped
product out the door than a low-quality
wide scoped one.
The business context is always changing.
As People Analytics practitioners we need to
be ready to adapt and meet our customers
where they are at.
We need to help the business attain its
goals, if we’ve lost sight of their goals they
lose interest in our Products.
Leaders want to get the most out of their
workforce. Workers want to get the most out
of their role and organization. Data can help
them both - a win win.
rst means we put the customer at
the center of the products we build -
frequently involving them in the design,
development, and maintenance process.
The Business Context
9. Ask yourself:
Does my organization have a lot of custom metrics
they are unwilling to change?
Does my organization already have suf
insights to hold them over or do they need more
information right now?
Does my organization have its data ready and
cleaned for plug and play analytics?
Does my organization have the right technologies
to take advantage of?
11. Common Tools and
Tech to Build
When you build from the ground up you need to be able to
aggregate, clean, analyze, and visualize the data on your own.
- AWS (S3; Redshift)
- Starburst Trino
- Starburst Trino
- Google Data Studio
Data Platform Tools Analytics Platform Tools Analytics Product Tools
13. Common Tools and
Tech to Buy
When you buy you need to be prepared to implement the
solution, which can take weeks to months. You need to have
your data prepped to
ow into the newly acquired tool.
Analytics Products on the Market
15. Structure Governance Discoverability
The data schemas you build need to
be well-structured and designed for
The data you own, data about your
organization and its workers, needs
to be secure and kept relevant.
elds are used for their
intended purpose, & for the ease of
consumers, data must be documented.
Data & Metrics Libraries
When building your platform:
Well-written code On Rails End-to-End Process
16. Raw Data
Optimizing the data
17. HR Data Sources
Data powers insight. HR organizations, their peer functions, and the business, have an array of tools and datasets
that are relevant to telling peoples’ stories that we need to access and mine.
Core HCM ATS CRM
18. Ensure that the structures of today are going to
the needs of tomorrow
Limit rework & focus on work that matters by
ditching bespoke tables or calculations that can
only be used in one place for one purpose.
Build with the
future in mind
20. Governing the right Access
Intersection of who and what a user can see
We have colleagues of many different
backgrounds and experiences - our
products need to be usable by them.
Our gold star is to have data
insights powered, without manual touch.
Human intervention is key to many areas in
people analytics but cyclical work can stand
on its own.
We may never match Finance’s or Sale’s or
Operation’s numbers perfectly, but we need
to tell the story - not the stats to the
hundredth of a percent.
We need to energize our data consumers by
ensuring their interactions and experiences
with out products are top-notch.
22. Will people use this? What are users return? Can action be taken?
Leaders have limited time. Ensure
that your product puts what they
need in front of them, fast.
Quick to Load
Do your users perceive a positive-
return on their investment for using
As commonly said trop in People
Analytics is “insight without action is
overhead.” How do we connect insights
Does a metric bene
Does a metric have business impact?
Does a metric tell us if we are on track?
When building your products, ask yourself:
Quick to Insights
Tailored Experience Assist in their Work
ow Does a metric help us strategize?
23. What to Measure
What will our demographics look like a
year from now based on current trends?
Do we have an equitably workplace?
Do we have a diverse workforce?
Are there biases in our hiring process?
Are people engaged?
Are employees using vacation time?
How much is turnover costing us?
Do we have the right skills?
Are my teams upskilling?
Do we have pipelines for career growth?
What will the cost of my workforce be?
Are workers returning to the of
Are we meeting our turnover goals?
Are we recognizing top performers?
Does the labor market view us as a great
place to work?
What chokepoints in the employee
experience is impacting our growth?
Business Impact Strategic Goals
How long does it take to back
ll my roles? Where are the greatest risks to the
24. How to Measure
Representation forecasting (predictive
Pay Equity Analysis along gender lines
Representation by Demographics
Funnel conversion rates for candidates
Engagement Survey scores
Vacation utilization rates at the end of
Total Cost of Attrition by Month
Employee Skills vs. Role Required Skills
Learning and Development time per FTE
Promotion Rates and Likeliest Next
Payroll cost forecasting (predictive
Badge Entry Trends
Rolling 12-month Turnover Rates against
Recognition Sentiment among Top
Glassdoor Score - 3 year trend
Employee Journey mapping - Candidate
experience; Onboarding; etc.
Business Impact Strategic Goals
Requisition Average Time to Fill
Impact and outlier analysis on tracked
26. Driving Engagement
Tie to Strategy
Answer strategic questions and tie insights
to goals to add context of the big picture.
Tie to Operations
Ensure users connect that insights about
their teams will enable them to run their
teams more ef
Tie to Action
Let users know that these aren’t just pretty
numbers on a screen - they can have
impact if you use them.
Hit the ground running with trainings,
communications, usage guides / job aides,
Foster a culture of learning, reasonable
assistance, and self-discovery through
Continually enhance your products post-
launch taking in feedback from users and
keeping them at the center of all you do.
27. Requisitions, Candidates, Applications, Interview Feedback, Assessment Tools, Labor Market Benchmarks, and Skills.
One HR’s more complicated, but insightful, datasets.
28. Candidates aren’t Employees
We are often working with personal emails,
personal addresses, and con
about people who are not workers of our
organizations, care needs to be taken to protect
Many Talent Acquisition tools do not store
requisition, candidate, and application information
as you’d expect - extra caution needs to be put
towards ensuring you are capturing the full picture.
TA, at the end of the day, is sales. Sales people
need easily accessible, easy to understand
information - fast.
Applicants put in the info they choose to disclose.
At best recruiters may put data in after the fact, at
worst they may skip steps in the hiring process.
We need to tell a story through incomplete data.
Pitfalls of Talent Acquisition Insights
30. Whats easy to measure, on just about
Open Req Count
Average Time to Fill
Candidate Gender Representation
Candidate URM Representation
Open Reqs per Recruiter
Reqs Filled in Previous Period
ciency by Demographics
Applicants per Open Role
Time in Stage / Stage Drop Off
Offer Acceptance / Decline Rate