Big data is hype. AI is for the movies. A computer might be able to beat the chess or Go world champion—but HR requires a subtler touch, right? Well, hang on. Even simple off-the-shelf analytics can generate real insights into people, productivity, and profits—insights that cement HR as a strategic advisor.
Implementing these systems (and understanding their limitations) should already be on the to-do list. Smart HR leaders also need to keep an eye on the next leap forward. Machine learning, natural language processing, sentiment analysis, and behavioral economics are going to become essential parts of the HR systems vocabulary.
This digibook is a roadmap to insight from analytics and essential advice for working with IT, management, and employees to lay the foundations for the future of HR.
In this guide, you’ll find…
If we seem to be missing three big-ticket items, it’s because they’re covered elsewhere in this series. Check out our three digibooks on Career Management and Development, Managing Organizational Culture, and HR Transformation.
Who will find this digibook useful?
HR leaders. The technology to understand data is developing fast. As sources of data proliferate—from enterprise systems to social media—our ability to understand people is moving into new dimensions. Analytics can put you ahead of this wave—riding it, not being submerged.
C-level execs. The new generation of database tools, data-gathering systems, and analytics, underpinned by emerging artificial intelligence technology, is going to give you more visibility of, and more insight into, your people than ever before. If you need to act fast, HR analytics will tell you where, how, and what the effect will be.
Line management. No technology is going to replace the judgment of a good manager. But the more you know about your teams, how they rate against benchmarks, how they feel, and what they care about, the better you can guide them to new territory. HR analytics can be your compass: Who’s doing brilliantly, how can you optimize their performance, and what’s the action plan for the laggards?
How far is too far?
Avoiding “the creepy line” with employee data.
The percentage of companies fully capable of developing predictive models in HR rose from 4 percent to 8 percent in 2016. Last year, only a quarter of companies felt ready for analytics. That proportion is now one-third.
But get this: Three-quarters of executives surveyed now rate people analytics as a key priority.1
The message? The race to better analytics is on!
We are all now “quantified beings”: constantly monitored, our preferences stored, our decisions mapped out. Algorithms—complex equations—crunch our data and tell businesses what to sell us, how to talk to us, what we think, and what we’re going to do next.
And thanks to enterprisewide systems and HR analytics, that’s now true inside our organizations, too.
The workflow for analytics is simple (we’ll come to the potential bumps in the road shortly):
What sort of questions, then, might HR analytics answer?
|Efficiency||HR Strategy||Business Planning|
|What are the turnover trends across the business?||Are we rewarding our high performers properly?||What would expansion into new markets mean?|
|What’s the return on our wellness program?||Is there a viable pipeline for leadership positions?||How do we design teams for better productivity?|
|How should we be allocating L&D spend in five years?||How should we redesign our recruitment processes?||How do we raise levels of customer experience?|
The creepy line.
An all-seeing HR function with the analytical capability to create insights from every aspect of an employee’s life risks crossing “the creepy line.” That’s what ex-Google CEO Eric Schmidt2 called the point at which people feel stalked by their technology.
Consider these five areas—all perfectly natural tools for HR and the business to deploy today:
- Biometric tracking of employees
- On-premises video surveillance
- Worker-productivity tracking
- Social-media monitoring
- Employee-engagement programs (with opt-in models)
The problem is that they might also be perceived as intrusive or even oppressive by employees.
In fact, by 2018, half of business ethics violations will occur through improper use of big data analytics, according to a report from Gartner.3 Understanding your employees without scaring them requires constant self-awareness from the HR team.
Disentangling the Jargon
The eight new buzzwords you need to understand.
There’s so much jargon and new technology floating around the world of data and analytics—and not all of it is useful. Let’s review…
1. Analytical maturity
A basic first question: how well does your organization analyze and derive insight from the data it already collects? A high level of “analytical maturity” implies it does complex analysis of individual variables and uses those resulting insights to predict outcomes of decisions.
2. Big data
Big data allows the whole workforce to be properly understood in every dimension. Focus on the four Vs: volume (there’s a lot of data), velocity (it grows and changes fast), variety (from raw pay-and-presence-type data to completely unstructured reports), and veracity (how accurate it is). Actually, make that five: Don’t forget value (how it might be exploited).
3. Data maturity
How good is your organization at collecting, sorting, storing, and exploiting its data? This is a key precursor to analytical maturity. If you’re going to apply powerful tools to your people data, you need to know it’s in good shape first. You need data of high integrity from multiple data sources, with the potential for different definitions and time horizons.
4. Enterprise reporting
Some businesses still live by their financials. But many have gone beyond pure “market data.” So-called integrated reporting recognizes key drivers of value—in particular, people and their relationships, skills, intellectual capital, and networks. It means HR data, analysis, and insight are combined with financial and other operational data to shape the business.
5. Machine learning
The most significant development in computing for a generation, this type of artificial intelligence provides computers with the ability to learn without being explicitly programmed. It means they can adapt to new data and start to uncover and analyze patterns without being told what to look for. That opens up powerful applications in sentiment and predictive analysis (see below).
6. Natural language processing
Computers used to be dumb. Make a small error in the language you’d use to query data, and you’d get nothing back. Now computers can understand the way real people speak, and respond with the answer you need.
7. Predictive analytics
A simple spreadsheet contains descriptive analytics—it’s numbers in a graph, like performance against benchmarks or team scorecards. But predictive analytics looks for correlations, identifies causes of outcomes, forecasts them, and helps you optimize the business to deliver whichever goals you want. It’s a crystal ball and the Holy Grail.
8. Social sentiment analysis
Knowing when people punch in or how much they’re paid is important, but dull. Knowing how they feel—about work, life, even their sports team—gives more insight into how they might be managed. If passive listening on company intranets and emails feels creepy, take a leaf out of GE’s book. It’s delivering “Fast Feedback” using PD@GE, an app that allows employees to post notes of encouragement, advice, or criticism under categories like Insight, Consider, and Continue.
“Castles made of sand.”
The credibility of HR’s insight program rests on getting the basics right first. If you can’t get correct headcounts or staff turnover rates, nobody is going to believe in your brilliant analysis of productivity data.
Here’s a warning from a CIPD Outlook report4 based on a survey of executives:
“While 37 percent of HR professionals believe their functional colleagues are satisfied with the people analytics they receive, only 14 percent of those ‘customers’ agreed.”
Reach too far, too fast, and leadership colleagues, line managers, and employees will become skeptical.
Baby steps for big questions.
Don’t spend your time gathering all the data you can, in the interests of understanding everything, all at once. Instead, follow these seven simple steps:
- 1.What’s the business question you’re answering? Be specific!
- 2.Have a go at a hypothesis. Don’t be afraid to have a best guess.
- 3.Gather accurate data. If garbage goes in, garbage will come out.
- 4.Analyze! But don’t fudge it. You need experts who understand stats.
- 5.Create insight. This is the “so what?” moment, the advice to other people.
- 6.Sell the story. An insight is no use unless HR makes the case compelling.
So remember, focus first on gathering:
- Ultrareliable data…
- …on a small number of…
- …very important areas…
- …that the board can use to make decisions.
Moneyball for Business
The golden age of advanced analytics.
When HR analytics gets more advanced, it’s capable of elevating the HR function to a new level. Insights will be delivered as decision support for front-line managers, or as clear-eyed strategic advice for leaders. The potential for delving into behavioral economics and proactive people management is immense.
If HR can frame that in clear business terms—analytics, ultimately, is about creating useful, easy-to-understand information—it’s the perfect way to gain the ear of board colleagues eager to keep the organization agile and efficient.
Better still, analytics can deliver unexpected solutions to common problems. Michael Lewis’s book Moneyball is a great example. It’s the story of how Billy Beane managed the Oakland A’s baseball team to success using statistical analysis rather than relying on the wisdom of the sport’s so-called experts (including coaches, players, and scouts), and spending on prestige players. HR can do the same for business.
The analytics advantage.
According to a 2015 report on C-level attitudes toward workforce analytics by IBM and Oracle,5 organizations that are good at using HR analytics perform better on every metric. “Companies need to place the same emphasis on knowing and engaging the workforce as they do on knowing and engaging their customers,” says the report. “Gone are the days of managing talent based on hunches or loose predictions.”
Nine Compelling Applications of HR Analytics
The use cases that put your team in the driving seat.
Analytics can reveal which criteria in your screening processes actually deliver high performers. Ditch the psychometric testing against notional ideals for certain jobs—and create your own tailored approach to parsing CVs and asking the right questions. Analytics can also reveal the best sources for new hires and the best channels through which to approach them.
Crunching data on leavers is a no-brainer. Google’s HR team noticed that there was a jump in departures after employees returned from maternity leave. Simply raising the amount of leave cut staff turnover by 50 percent.6 Predictive analytics can attach “flight risk” scores to individual employees, providing a new tool to HR and line management.
If the HR function is transactional, compliance-led, or focused solely on driving down costs, it’s going to become a commodity. Analytics opens HR to giving insights on the big challenges, such as boosting sales or accelerating R&D. Measure, monitor, and predict the effect of risk factors over time; devise contingency plans based on insight.
Analytics allows you to look at market rates, benchmark against peer organizations—and produce evidence that demonstrates an employee is on track. In agile businesses that need to upskill in new areas quickly, this could be invaluable.
5. Reputation management
Social-media analytics helps HR refine employee engagement or internal marketing campaigns. Employer brand management sits squarely with HR, not just for optimized recruitment, but also to maintain vibrant networks outside the organization—crucial to referrals.
6. Skills and talent optimization
Correlations between productivity, engagement, behaviors and assignments means getting the right people into the right jobs becomes easier. Analytics turns learning and development7 from quota-filling into targeted skills and behavior change. Machine-learning technologies will automatically understand an employee’s needs and adapt their work context and tools to their job profiles.
7. Setting expectations
If you know how long it takes for employees in any given role to work through promotions, for example, you can focus those people on realistic progression timeframes and skills requirements.
8. Managing poor performers
HR analytics can spot behaviors that flag toxic activities. But why not refine further: One Silicon Valley company used analytics to spot specific interview questions that “toxic” people answered in an exaggerated way, making them high-probability unethical employees.
This is both the biggest, and the easiest, role analytics can play. Deliberate or unconscious bias in the hiring, remuneration, promotion, and retention of staff from a range of backgrounds can be eliminated by four simple steps:
Gather raw data on ethnicity, sexual orientation, and social background to complement basic data like age and gender. Do you know your workforce?
Analyze the data. Benchmark against local populace, customer makeup, and peer organizations. Check career progression and retention numbers for target groups. Where is there under-representation?
Work out where and why unconscious bias might be creeping into hiring and progression. Where are your black, Asian, ethnic-minority, LGBT, and female employees stalling?
Point this out. Data is powerful: It’s not an opinion about specific racist or sexist behaviors. It’s not criticism. It’s just the facts.
This is not just about HR crunching the algorithms and reporting back. It must also use analytics to suggest solutions where it’s clear there are issues. This is very much a strategic play for HR.
Some data-focused companies, such as advertising agency JWT,8 are sufficiently far advanced in this process—and determined enough to drive better diversity—that they use “blind hiring” for junior roles, shortlisting candidates algorithmically without looking at their CVs.9
The Challenges of HR Analytics
Keeping control, making people understand.
HR analytics and data-driven insights hold great promise. But unless HR, IT, and finance—as well as employees and line managers using systems and applying insights—are working in close collaboration, your project is going nowhere. This is our guide to ensuring that relationship works.
The first rule of HR analytics.
People analytics must report into the chief HR officer. Let’s repeat: PEOPLE ANALYTICS MUST REPORT INTO THE CHIEF HR OFFICER. Analytics blurs lots of lines. The systems fall under IT. The inputs are often departmental. The outputs ought to drive finance, marketing, and operations. Everyone wants to know what drives people and how they’re performing.
Fight. Them. All. Off.
But do it nicely. Because the kicker is that you’re going to need all of those people to help make HR analytics work. It’s about offering analysis to managers, then leaving the door open to discussions about drilling into the issue. And when managers are equipped with the analysis, often they do want to drill down—with HR there to help.
And if you’re not a chief HR officer with a seat on the board? It might be a good idea to swallow your pride and stay under the CFO’s wing (still the most common place to find HR when it’s not on the board).
The three people you need to hire now.
An HR analytics lead. You need someone to own this—someone who can draw together the different strands and has a clear idea of what’s possible now, and a vision for where you want to be in five years’ time.
“Quants.” Statistics and modeling are highly technical areas that contain bear traps galore. Unless you really know your sampling distribution from your scatter plot, get help. Ambitious quants from the finance (or even marketing) function are a good bet. They also know cross-departmental experience will accelerate their career. Think about hiring a data scientist, too (see box). The roles are similar, but not identical.
Storytellers. Data, well analyzed, becomes insight—but needs a compelling story to make it resonate with business decision-makers. They need to get why your analysis matters, fast.
Appliance of science: The traditional “quant” is evolving—moving beyond sifting and recutting data to become a “data scientist.” This implies a deeper understanding of database management and designing algorithms. As big data technologies mature, the focus will shift again, away from these skills and toward predictive modeling, and away from programming or pattern analysis.10
Winning over finance and the C-level.
According to a survey of finance leaders by the Canadian Financial Executives Research Foundation (CFERF), two-thirds of executives rated visually appealing HR analytics as important.
More than half the finance execs stated that they would focus HR on a top-line metric such as revenue per FTE (full-time equivalent) or labor costs as a percentage of revenue; 47 percent of public companies indicated a preference for so-called bottom-line measures: profit margin per FTE or human capital ROI.
The CFO is watching…and could be an important ally.
Four things to ask for.
Enterprise systems. HR data in a silo means analytics is always going to be limited. You need to have access to all data. IT strategy should be looking at smarter data-warehouse solutions.
Accurate data. Some systems are great at managing unstructured data; most aren’t (yet). If the data isn’t timely and comparable, it’s garbage.
Easy-to-use systems. That IT staffer with a passion for algorithms and black-box mathematical models has her uses. But designing your analytics engine isn’t one of them. Simple, actionable analytical models are key. HR people need to get it; and communicate it.
Operational efficiency. The facts change in HR all the time. In big organizations, hundreds of people come and go every month. So predictive analysis models should also be run at least monthly—and dashboarding systems should be in real time.
And a final word of warning.
Even with a great predictive model, personal factors will define an individual’s actions. So analytics should not be a crutch—it’s a guide, not a prescription.
You still need management judgment.
Both HR and operational managers need to know that this isn’t about rendering them obsolete; it’s about supporting better decisions and opening up areas for discussion that might have been missed.
Use this checklist to build your action plan. As you select each item, they will build into a comprehensive set of next steps for you.
Which of the following do you need to do?
Check those that apply:
Your action points
Get the basics right
- Evaluate your current data generation, collection and usage.
- Audit for accuracy.
- Aim for pinpoint accuracy on the basic numbers—in large companies, headcount, remuneration and productivity are often inaccurate.
- Set some ethical red lines about how data might be gathered or deployed.
Talk to IT, Finance and the business
- How integrated are your enterprise systems? If you’re siloed, pause.
- Work out what analysis might drive change—Finance is desperate to know more about people, for example.
- What do line managers, the board and employees themselves care about? Culture? Productivity? Efficiency of HR? Talent management? These can set your objectives.
Frame the questions analytics can answer
- Resist the temptation to oversell analytics.
- Decide on simple questions, answerable by smart analysis of existing data.
- Have a clear idea how leaders will react when you provide those answers.
- Consider high-visibility areas for new data analysis—like diversity.
Get your team right
- HR skills will need to be augmented. Target people with skills from Finance (especially data scientists), Marketing (framing stories) and IT (you need a nerd).
- If you’re a CHRO, be clear in your ownership of people analytics on behalf of the business. If you’re not, find a sponsor (like the CFO or COO).
- Make maximum use of HR business partners to sell—but not oversell—the output.
Prove, enhance, expand
- Deliver quick wins—insights that change HR policies and help managers do relatively straightforward people tasks better with less effort.
- Apply experience to more datasets over time. Ask tougher questions. Invite requests for deeper insights into the people dimension.
- Refine systems and capabilities to develop a long-term analytics vision.
- Target the cutting edge: sentiment analysis, behavioral economics, and more.