Workforce Analytics: Using Data to Drive Talent Decisions

A head of HR at a 600-person manufacturer in Riverside called us last summer. She wanted to hire “a people analytics person.” Halfway through the intake call she paused and asked if we could help her figure out what she actually meant by that. Did she want a SQL analyst? A statistician? A Workday admin who could build a dashboard? She wasn’t sure. Her CEO had read a McKinsey article on a flight and told her to “get the data thing going.” That was the brief.

Workforce analytics is the practice of pulling employee, hiring, and operational data into one place so that talent decisions stop being made from gut feel. It covers turnover, time-to-fill, internal mobility, compensation, engagement, and the dozen other metrics that tell you whether your people strategy is actually working. Done well, it answers questions a CFO will respect. Done poorly, it produces dashboards nobody opens.

I’m Devin Hornick. I co-founded KORE1 about two decades ago and I’ve watched the analytics conversation inside HR mature from “can someone export this from PeopleSoft” to “we need a director of workforce analytics with a doctorate.” The market is in a strange spot right now. Boards want the AI version of this work. The actual companies trying to build it are mostly still cleaning up their exit interview data. We place the people who do both ends of that, through KORE1’s IT staffing practice and our HR vertical, and what I want to do here is give you the version of this conversation we have on intake calls. Without the vendor pitch.

One thing up front. We make money when companies hire analytics talent through us. Read everything below knowing that. I’ll still tell you when we’re not the right fit, because that part matters too.

What Workforce Analytics Actually Is (Without the Vendor Pitch)

The plain version, the one I’d give a CEO who has thirty seconds, is that workforce analytics means pulling employee, hiring, and labor-cost data into one place so you can answer questions about your people the same way finance answers questions about money. The inputs come from the HRIS, the ATS, payroll, performance reviews, and whatever engagement survey tool the previous head of HR signed a three-year contract for. The output is usually a dashboard. Sometimes a model. Occasionally a single number that changes how the leadership team thinks about a problem they’ve been arguing about in meetings for six months.

You’ll see three other terms thrown around constantly. HR analytics. People analytics. Talent analytics. Walk into ten HR conferences and you’ll get ten different opinions on what each one means. The honest answer? In the work itself, they’re about 80% the same thing. The differences are mostly marketing, with a few academics and a few platform vendors trying to enforce distinctions that nobody else respects. If you hear someone draw a sharp line between “people analytics” and “workforce analytics,” they’re either selling you something or defending a job title.

Here’s a real example of what good workforce analytics looks like in practice. A client we worked with last year, mid-market software company, was losing engineers at a 22% annual rate. Descriptive analytics told them what they already knew. People are quitting. Diagnostic analytics told them something they didn’t know. The turnover wasn’t distributed evenly. It clustered under three specific managers, all promoted from senior engineer roles in the same eighteen-month window. Comp wasn’t the issue. Tenure of the manager was.

That insight cost them about four hours of an analyst’s time. The downstream value was a manager training program for the three folks in question, two quiet reassignments back to senior IC roles, and an engineering organization that finally stopped bleeding mid-level talent at the rate it had been bleeding for the previous eighteen months. Their turnover dropped to 14% the following year. None of that required predictive AI. It required someone willing to slice the data more than one way.

The Four Types — and Which One You Actually Need

Workforce analytics gets sorted into four buckets. The buckets matter because companies tend to want the most advanced one before they have the basics working.

1. Descriptive. What happened. Headcount last quarter. Turnover by department. Time-to-fill for the open reqs. This is where 70% of mid-market companies live, and most of them think they’re behind because of it. They’re not behind. They’re at the starting line.
2. Diagnostic — why did it happen? Same data, but you slice it differently. Why is sales turnover three times higher in the Texas region? Why are women leaving engineering after eighteen months specifically? This is the layer where most of the actual value gets created, and it’s also the layer that requires almost no new tooling. A spreadsheet and a curious analyst will get you there.
3. Predictive analytics. Will it happen. Forecasting which employees are at flight risk, which open reqs will take longer than six weeks, which compensation bands will fall behind market by Q3. Real, useful, hard. Requires clean historical data going back at least three years. Most companies don’t have it.
4. Prescriptive. The vendor demo layer. The dashboard tells you what to do about it. We’ve sat through a lot of these demos. They look great in the controlled environment. They look different when they’re connected to your actual messy HRIS data.

Most companies want to skip from descriptive straight to predictive. That’s where it falls apart. The diagnostic layer is the one that pays for itself, and almost nobody talks about it because there’s no software to sell against it.

The Metrics That Actually Drive Talent Decisions

Here are the metrics we see on real workforce analytics dashboards. Some of them are useful. Some of them lie to you in specific, predictable ways. Knowing the difference is most of the job.

The single most useful metric I’ve seen, and almost no company tracks it, is what I call the “would you re-hire” rating. Six months after a new hire starts, you walk over to their manager and ask one question. If you could rewind to the day this person accepted the offer, with everything you know now about how they’re actually performing in the role, would you make the same hire again? Yes or no. No scale, no rating, no comments box that turns into a HR-speak essay. The answer correlates with first-year retention better than any structured interview score, any personality assessment, and any reference-check rubric we’ve ever benchmarked across our placements. It also takes about ninety seconds per hire to collect, which is exactly why nobody on the talent acquisition side has an excuse not to be doing it already.

What This Looks Like at a 200-Person Company vs a Fortune 500

Most workforce analytics content on the internet is written for the wrong company. It assumes you already have a Workday implementation, a Visier seat, a People Analytics team of twelve, and a CHRO who reports to the CEO with a quarterly seat at the board meeting. That describes maybe 5% of the actual employer market in the United States, and the rest of you, the 95% running smaller and messier operations, get treated like an afterthought in every published guide on this subject.

At 200 to 2,000 employees

You probably have BambooHR, Rippling, or Paylocity. One HR ops person who knows just enough Excel to be dangerous. A CSV export button on every screen. No data warehouse. No analyst on staff.

The realistic version of workforce analytics for you looks like this. One person, ten hours a week, building dashboards in Looker Studio or Power BI off CSV exports refreshed weekly. Total tooling cost under $200 a month if you’re already on Google Workspace or Microsoft 365. The “analyst” is usually an HR coordinator with strong spreadsheet instincts who you train up, or a part-time hire from a finance background who already speaks the language. Either works. A six-figure people analytics manager is overkill at this scale and they will be bored within four months.

The trap at this size is buying the platform before you have the person. Visier, Tableau Cloud, Crunchr — they all sell to mid-market now. The pitch sounds great. Then the platform sits half-implemented because nobody on staff has the time or skill to model the data, and twelve months later the contract renewal email lands and someone realizes the dashboards are still empty.

At 5,000 employees and up

Different planet. Workday or SAP SuccessFactors. A real data warehouse, usually Snowflake or BigQuery, with HRIS data flowing in nightly. A People Analytics team of four to twelve, reporting either to the CHRO directly or through a head of HR strategy. Vendor stack might include Visier, ChartHop, One Model, or in-house Tableau builds. Annual spend on the function alone, including salaries and tools, runs $1.5M to $4M.

At this size the conversation isn’t whether to do workforce analytics. It’s whether your function is mature enough to influence executive decisions or whether it’s still a backwater that produces quarterly reports nobody reads. The good ones are sitting in headcount-planning meetings with the CFO. The mediocre ones are still answering ad hoc requests from HRBPs.

Hiring the Person Who Runs This

This is the part the other guides skip. They tell you what workforce analytics is in the abstract sense, complete with maturity models and four-stage frameworks, but they almost never tell you what it actually costs to hire a real human being who can do the work, which is the only number that matters when you’re trying to get a budget approved. We will, because that’s literally our day job, and the compensation data we give clients comes from data analytics placements we close every single quarter, not from a self-reported salary aggregator that pools fifteen years of bad data into a misleading average.

Three roles. Three very different price tags. Three different things they should actually be doing.

HR Data Analyst. Base comp $75K to $115K depending on market and experience. Glassdoor reports an average around $87K nationally as of early 2026, with the senior end pushing into the low six figures in major metros. PayScale lands a bit lower because their dataset skews toward smaller employers. The job is SQL, Excel, Tableau or Looker, occasional Python for the brave. Lives inside the HRIS. Builds the recurring reports, answers ad hoc questions from HR leadership, and slowly turns the function from reactive into something useful. Best fit for companies in the 500 to 3,000 employee range that have outgrown the spreadsheet phase.

People Analytics Manager. Base comp $130K to $180K, with total comp pushing $200K-plus in tech-heavy markets. This is the inflection point hire. They own the function. They build the dashboards and the models. They sit in meetings with VPs and translate “I think we have a retention problem in the Midwest” into a data question that can actually be answered. The skills the job description usually asks for are Tableau, Workday, SQL, R or Python, and a master’s in I/O psychology or HR. The skills that actually predict success are different. Curiosity. The ability to push back on a senior leader who is wrong about what the data says. Comfort with ambiguity. We’ve placed analysts with no master’s degree who outperformed PhDs because they could ask better questions.

Director or Head of Workforce Analytics. Base comp $180K to $260K, with total comp reaching $350K at large enterprises. Strategy role. Picks the vendors. Builds the team. Reports to the CHRO or sometimes directly to the COO. Presents to the board. The hardest of the three to find because you need someone with both the technical credibility to run the function and the executive presence to defend it in budget conversations. Most clients we run this search for take three to five months to fill it.

One specific pattern we’ve seen go wrong, repeatedly. Company hires a data scientist out of a FAANG into a 400-person company that has no data infrastructure. The new hire shows up expecting to build predictive models. What they find is dirty CSV exports and a manager who wants a weekly turnover report. They quit within nine to twelve months. The hire looked great on paper and was a perfect technical fit for a company they didn’t actually work at. We’ve watched this exact pattern play out four times in the last three years and the salary the company “saved” by not hiring through a search firm gets erased the second they have to backfill the role.

If you’re hiring for a permanent analytics lead, we tend to recommend going through a direct hire search rather than starting on contract. Analytics functions take six to nine months to build trust with executives. A contract-to-hire arrangement signals to the candidate that you’re not committed to the function, and the strongest candidates pass.

Where It Goes Wrong

I want to walk through three failure modes specifically because we see them constantly and they cost real money. First one. Company buys Visier or one of the other workforce analytics platforms before they have anyone qualified to implement it, on the theory that the platform will somehow enforce discipline on the underlying data. It doesn’t. Visier is a terrific tool when it’s connected to clean, structured HRIS data with consistent job codes, standardized exit reason categories, and a complete history. Most mid-market companies don’t have any of that. They have job titles that changed three times in five years, exit reasons that are 40% blank because the offboarding form was optional, and a Workday implementation that nobody fully signed off on. Plug that into Visier and you get expensive dashboards that show garbage. The platform doesn’t fix the data problem. The platform exposes it. Second failure mode. Letting IT own the workforce analytics platform instead of HR. This sounds like the right call because IT has the engineers and the data warehouse expertise. In practice, the dashboards get built by people who don’t talk to HR every week, the metrics drift away from what HR leaders actually need, and the function becomes a tools project instead of a talent project. The HRBPs stop opening the dashboards. The CHRO stops asking for them. Six months later the whole effort is dead and nobody is sure exactly when it died. Third one is the data scientist hire I mentioned earlier. A FAANG-trained ML person walks into a 400-person company expecting clean data and a real engineering platform, finds neither, and is gone before they ship anything. The fix isn’t to hire less senior people. The fix is to hire someone whose actual experience matches your actual environment. A senior analyst from a mid-market HR tech company will outperform a junior person from Google in your specific situation, ten times out of ten.

AI and Workforce Analytics: The 2026 Reality Check

You can’t write about workforce analytics in 2026 without addressing the AI question, partly because every vendor pitch starts with it now and partly because the actual answer is more interesting than the hype suggests. Here’s how I think about it after sitting in too many demos.

According to the SHRM 2026 State of AI in HR report, 26% of HR professionals at AI-adopting organizations now use AI tools weekly, 20% daily, and 9% several times a day. The number is real. The shift is real. The interpretation is where people get it wrong.

What’s actually working right now, in production, at companies we’ve placed people into in the last six months, includes natural language queries against HRIS data where you ask “show me the regretted attrition rate for engineering in Q1 by tenure band” and the tool writes the SQL for you in two seconds. AI-generated first drafts of job descriptions. Auto-summarized candidate screening notes. Predictive matching of internal candidates against open requisitions. These are real productivity gains and they are saving real analyst hours every single week, which is the only honest measure of whether a tool is doing what its marketing says it does.

What isn’t working yet, no matter what the keynote demos showed. Autonomous AI workforce planners that supposedly take messy enterprise data and produce executive-ready insights without human cleanup. We’ve sat through demos of three of these in the last year. They’re impressive on the rehearsed dataset. They fall apart on real client data within ten minutes. The data quality problem at most companies is the rate-limiting step, not the AI capability, and no amount of clever modeling can rescue garbage inputs.

“The AI is ready. The data isn’t. Most workforce analytics failures in 2026 will look like AI failures, but they’ll actually be data hygiene failures wearing an AI costume.”

If your HRIS data is clean, AI will accelerate your analytics function dramatically. If it isn’t, AI will accelerate the production of confidently wrong answers. Pick which problem to solve first.

Questions Ryan and I Get on These Searches

These come up on almost every intake call. Short version of each, the way I’d answer over the phone.

So what’s the actual difference between workforce analytics and people analytics?

For 80% of the work, none. The vendors and the academics will tell you otherwise. The recruiters who hire for these roles every week will tell you that the same candidate gets considered for both job titles, and the salary range is identical.

Do we need to buy software before we hire someone?

Wrong order. Hire the person first. A good analyst can build something useful out of CSV exports and a free Looker Studio license while you figure out what platform you actually need. We’ve watched a lot of companies do it the other way around and the pattern is consistent. They buy Visier or Tableau Cloud first because the vendor sales cycle is fast and the hiring cycle is slow, then the contract sits half-used while they search for someone who can actually implement it. A year goes by. The CFO asks why the line item exists. Awkward conversation. Don’t start there. Hire the analyst, give them three months to figure out where the gaps are, then let them tell you which platform to buy. They’ll pick a better one than your vendor shortlist would have.

What does a workforce analytics lead actually do day-to-day?

Mornings tend to be dashboard maintenance, answering urgent questions from HR or finance, and the kind of triage work that doesn’t show up on any job description but eats four hours a day if you let it. Afternoons are deeper work. Building a model, running an analysis, prepping for an exec meeting. About a third of the week goes to stakeholder management, which sounds like overhead but is actually the part that keeps the function alive. A people analytics manager who never leaves their desk gets defunded in the next budget cycle.

Realistic salary range for a senior people analytics manager?

$150K to $185K base in most US markets, with bonus and equity layered on top. NYC, SF, and Seattle push higher. Smaller markets like Phoenix or Charlotte come in around $130K to $160K. Use our salary benchmark assistant if you need something more specific to your geography.

Can we just train our HRIS admin to do this?

Sometimes. The HRIS admin who already wonders why the exit data is messy and has been quietly building spreadsheets in their spare time? Yes, train them. The HRIS admin who treats the system as a closed box and only opens it when something breaks? No. Different mindset, different person. You can’t train curiosity.

How long until we see ROI?

Wrong question, slightly. The ROI isn’t a single number. Diagnostic insights pay back in weeks, sometimes the first month, because they redirect decisions you were already going to make. The full predictive layer takes twelve to eighteen months because that’s how long it takes to build clean enough data to train models on. If your CFO is demanding a one-line ROI commitment before approving the function, give them the diagnostic horizon, not the predictive one. Set the bar where you can clear it.

The Bottom Line

Workforce analytics isn’t a software problem. It’s a hiring problem. The companies that get value from this work are the ones who hire the right person before they sign a platform contract, not the other way around. Order matters more than budget.

If you run a 200 to 2,000 employee company, you almost certainly need an HR data analyst and a free dashboard tool first, not a six-figure platform with a multi-year contract. If you’re at 5,000-plus, you need a people analytics manager and a real conversation about data infrastructure with whoever currently owns the warehouse. If you’re somewhere in the awkward middle and genuinely unsure which version of this story applies to your company, that’s the call to make first, before anything else gets ordered or signed.

KORE1 isn’t the right partner for everyone. If you’re a Fortune 100 with an existing People Analytics team and you need a niche specialist in workforce planning modeling, you’re probably better served by a boutique that does only that. If you’re under 200 employees, you don’t need us yet. Build the function with one curious internal hire and come back when you’re ready to scale. For everything in between, especially the messy 500 to 5,000 employee range where most of these searches actually happen and where the wrong hire is most expensive, that is exactly where we work. When you’re ready to talk through what your workforce analytics function should look like, who should lead it, and what the realistic comp range is in your specific market, you can talk to our team and we’ll set up a call.