The Insight-Driven Leader: How High-Performing Companies are Using Analytics to Unlock Business Value

By Jenny Dearborn & Kelly Rider

10 min read

Why does it matter? Because the metrics HR is most proud of are the ones executives have learned to ignore.

Your HR department is probably hitting every target it was given — and that's the problem. Time-to-fill: green. Cost-per-hire: green. Candidate ratios: green. Meanwhile the business is bleeding revenue, top performers are walking out the door, and the CEO is quietly wondering whether to hand your function to someone from finance who actually speaks the language of risk and growth. The gap — between internal scorecards that look fine and business outcomes that don't — isn't a resourcing failure or a technology gap. It's a thinking failure. HR has been measuring the wrong things with enormous precision. What follows is a ladder out of that trap: a four-stage framework that moves people leaders from reporting what already happened toward predicting and shaping what happens next. The seat at the strategy table isn't given. It's earned, one translated data point at a time.

Every HR Dashboard Is Green. The Business Is Bleeding.

Elke Andersen had every reason to feel good about her numbers. The head of Talent Acquisition at Exalted Industries — a struggling market leader hemorrhaging sales reps and missing revenue targets — could point to a five-day time-to-post, a candidate-to-opening ratio of 15-to-1 (three times the industry average), and a cost-per-hire that made competitors look reckless. By every standard she'd been measured against, her team was doing its job.

Then Pam Sharp, the company's new chief HR officer, asked a different question: do any of those numbers connect to whether the hire actually succeeds?

Elke paused.

The Same Data That Looks Fine at the Top Reveals a Crisis Underneath

The attrition problem at Exalted looked, from the outside, like a recruiting failure. Positions were going unfilled. Top producers were walking out. The obvious fix: hire faster, hire better. What the data actually showed was something nobody in the building was ready to hear — the company's best salespeople were being financially punished for succeeding, and the incentive structure doing the punishing had gone completely unexamined.

When Chloe, the analyst brought in to build Exalted's first real people-data capability, mapped commission rates against actual sales performance, she found a result that stopped the room cold. The highest-producing reps — the ones generating around $16 million in revenue — were earning commissions at a rate nearly 25 percent lower than reps producing a fraction of that. The instinct is to dismiss the finding as a data error. But pressing on the why revealed something more damaging: those top reps had been assigned quotas so impossibly high that even their outsized results left them technically short of target. Meanwhile, newer reps with small territories and modest numbers were clearing their modest quotas easily and pocketing better commission rates. The people the company most needed to keep were being told, in the language of their paychecks, that their performance wasn't good enough. And because quota-setting sat with Sales Operations — separate from the HR team tracking compensation — nobody had ever put those two data streams in the same room.

The compensation team touched base pay. Sales Ops controlled quotas.

Nobody owned the gap between them.

The result was a slow hemorrhage of exactly the talent the company couldn't afford to lose, with leadership reading their individual departmental reports and seeing nothing alarming.

Attrition data, read at the surface, will always point you toward the symptom: we need more candidates, better recruiters, faster pipelines. Tracing a retention problem back through incentive design, quota mechanics, and cross-functional blind spots requires pulling data across silos that have never been asked to speak to each other. When they finally do, what looks like a pipeline problem turns out to be a leadership problem. And leadership, not HR, has to own the fix.

The Four-Stage Ladder That Turns People Data Into Executive Decisions

A hospital takes your temperature at the front desk — descriptive: here's what's happening. A specialist traces the fever back to an infection — diagnostic: here's why. An oncologist runs a risk model to forecast recurrence — predictive: here's what might happen next. The care team designs an intervention protocol — prescriptive: here's what you do about it. Most organizations never leave the waiting room. They take temperatures endlessly and call it insight.

Each rung of the ladder asks a harder question. Descriptive analytics asks what happened — headcount reports, time-to-fill, completion rates. Diagnostic analytics asks why, and this is where the first real work begins. Predictive analytics asks what's likely to happen next, using historical patterns to flag risks before they materialize. Prescriptive analytics asks what you should actually do, translating predictions into specific interventions.

The ladder earns its credibility through what Chloe did with Exalted's sales rep data. Her goal was to identify the 'DNA' of a top performer — not a hunch-based list of desirable traits, but a ranked map of which measurable behaviors actually drove success ranking. She assembled input variables for every rep: deal size, leads generated, number of training sessions completed, whether they collaborated across teams. Then she ran a machine learning model that treated sales ranking as the output and all those KPIs as competing inputs. The model's job was to learn which variables it kept reaching for when it correctly predicted who ranked where. The most frequently weighted variables became the DNA — not because someone in Sales declared them important, but because the data showed they were doing the predictive work. The intuition that a rep's eye contact in the interview room determines their ceiling gets replaced by something like: reps who complete product training within their first sixty days and carry average deal sizes above a certain threshold retain at twice the rate. That's a prescription you can act on by Monday.

The ladder teaches sequence above all else. Organizations that skip straight to prescriptive — buying a predictive tool before they can explain why their last hundred hires succeeded or failed — are trying to climb from rung one to rung four without touching two or three. You cannot prescribe the right intervention until you can predict which reps are flight risks. You cannot build that prediction without diagnosing what actually drives attrition. And you cannot run that diagnosis until your descriptive data is clean, consistent, and asking the right questions in the first place. The rungs are not optional stops. They are prerequisites.

The 'Refer' Trap: How Good Intentions Systematically Produce Bad Hires

Why do bad hires keep getting made even when the process looks like it's working? At Exalted, the analytics team found that hiring managers already knew the answer — they just never said it out loud.

When Chloe mapped candidate quality ratings against hiring outcomes, she found a structural trap built into the process itself. Managers were regularly rating candidates as low quality, but instead of rejecting them, they marked them 'refer' in the system. The logic was sympathetic: there's a vacant territory bleeding revenue, and passing a weak candidate sideways costs nothing. So the candidate moved to the next manager, who did the same calculation, who passed them again, until someone desperate enough to fill a seat extended an offer. The data showed exactly what this cost: a high-quality candidate received an offer within three weeks. A low-quality candidate, laundered through the refer chain, took nearly three months — and then failed in the role anyway.

The trap feeds itself. High attrition means Talent Acquisition is perpetually firefighting, which means there's never time to build a real candidate pipeline, which means every vacancy arrives with urgency attached, which means the pressure to refer rather than reject intensifies. Nobody in the system was making a cynical choice. Everyone was responding rationally to their local incentive, and the collective result was a machine that reliably produced hires who didn't last.

The fix wasn't widening the funnel. It was changing what the funnel screened for. Using the sales rep DNA profile — the quantitative model built from behaviors that actually predicted performance and retention — the team could identify candidates with a genuine propensity to succeed before the refer pressure ever kicked in. The target: double the rate at which quality candidates accepted offers, from 20 percent to 40 percent. That math only works if you stop recirculating the wrong candidates in the first place. More resumes was never the answer. Better criteria, applied earlier, was.

The Data Is Never the Hard Part

The phone call came right after business hours. Before Pam Sharp could set down her bag, the voice on the other end was already bellowing: 'And just who do you think you are?' Tom Ashcroft, a board member at Exalted Enterprises and the architect of its premium-brand identity, had heard that the company's new CHRO was circulating data suggesting his strategy was destroying the sales force. His message was blunt: she'd been hired to fix HR, not to meddle in departments above her weight class. Then he hung up.

Ashcroft wasn't reacting to a spreadsheet error. He was reacting to a threat. The data Pam's team had assembled — pulling from CRM records, exit interviews, pricing reports, and competitive intelligence — showed that the 'stay premium, sell value' strategy he'd championed was the reason 70 percent of deals were now being lost on price, up from 20 percent just three years earlier. The reps weren't making excuses. They were accurately describing the market. But Ashcroft's identity, and the identity he'd cultivated in CMO Anne Rodriguez over a year of mentorship, was inseparable from the Maserati positioning he'd sold to the board. Data that contradicted it wasn't a finding; it was a political attack.

Almost nobody talks about this when HR analytics gets sold as a capability investment. Good analysis produces findings that redistribute accountability. When Chloe's work showed that attrition was driven partly by unrealistic quotas, someone in Sales Operations had to own that. When it showed that 90 percent of departing reps cited their manager, suddenly Bobby Cash's promotion practices were on the table. Every insight has a constituency it threatens.

Pam got the data anyway — from Anne, from Bobby, from functions that had never shared raw files with anyone outside their teams — not by wielding authority but by treating every resistance conversation as a relationship to build rather than an obstacle to bulldoze. The squash games with Anne were the least surprising part; what actually moved things was what she told Bobby Cash: that she wouldn't go to the CEO without him, and that any finding pointing at his department would reach him first. That promise cost her nothing and bought her everything, because it reframed the analytics from a surveillance operation into something he had a stake in. None of that is analytics work. All of it determined whether the analytics work survived.

The hard part is never the data. It's the ecosystem of ego and turf that data, when it's working, inevitably disturbs.

Analytics Becomes Real When the Skeptic Becomes Its Loudest Advocate

Thomas Ashcroft had dismissed the entire effort as an HR experiment, and when he sat down at the January board table, he was still planning to say so. Nine months of analytics work, a CFO-validated 25 percent jump in new product revenue bookings, sales rep attrition brought down from 60 percent to something approaching industry norms — and Ashcroft called it a rising tide. A market trend. A one-time win that would evaporate once the board came to its senses.

What finished him wasn't Pam's data. It was Bobby Cash.

The same CSO who had started the year insisting he needed to look candidates in the eyes — who had been openly skeptical that HR had anything to teach Sales about building a team — raised his hand at the board meeting and told Ashcroft he was wrong. Not diplomatically wrong. Actually wrong. Bobby described what had changed on the ground: managers coaching differently, reps working as a unit, attrition numbers that had gone from catastrophic to unremarkable. He said he'd been just as suspicious as Ashcroft once. The results changed his mind. Then Anne Rodriguez, Ashcroft's own former protégée, looked directly at her mentor and told him she couldn't ignore the numbers — that the market had simply moved, and that the fresh perspective had taken Exalted somewhere it couldn't have reached on its own. Ashcroft leaned forward to respond, thought better of it, and gave a single nod. He resigned before the day was out.

The data hadn't converted Bobby. Watching his team stabilize, seeing his managers develop, tracking the revenue curve — that was the conversion. What the analytics did was give him a language precise enough to defend what he'd lived through. And that's the distinction worth holding: an analytics initiative completes itself not when the models are built, but when the people whose departments the data touched are willing to stand up in a room and call it true. That's when the methodology crosses from HR experiment to corporate strategy.

The Question the Data Can't Answer for You

Pam didn't build a better HR function by buying better software or hiring a sharper analyst. She got there because she made a decision, before the first dataset was pulled, that she would follow the evidence even when it pointed somewhere inconvenient — at a board member's legacy, at a colleague's blind spot, at her own organization's flattering self-image. That's the thing the tools can't give you. The frameworks are learnable in an afternoon. The data your company needs almost certainly already exists, scattered across systems that have simply never been asked to talk to each other. What's genuinely scarce is the willingness to surface a finding that implicates someone with more tenure than you, and to build the alliances — before the analysis is done — that keep that finding alive long enough to matter. Analytics isn't a mirror that flatters. The only real question is whether you're ready to look.